Research Methodology

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/369479003
Research Methodology
Book · December 2018
CITATIONS READS
0 19
2 authors:
A. Sivakumar GOVINDARAJAN SINGARAVELU

K.S.R. College of Engineering Bharathiar University
79 PUBLICATIONS 266 CITATIONS 204 PUBLICATIONS 182 CITATIONS
SEE PROFILE SEE PROFILE
All content following this page was uploaded by A. Sivakumar on 24 March 2023.
The user has requested enhancement of the downloaded file.

Research Methodology
Dr. G. Singaravelu
and
Dr. A. Sivakumar
A.P.H. PUBLISHING CORPORATION

4435–36/7, ANSARI ROAD, DARYA GANJ,
NEW DELHI-110002
Published by
S.B. Nangia
A.P.H. Publishing Corporation
4435–36/7, Ansari Road, Darya Ganj,
New Delhi-110002
Phone: 011–23274050
e-mail: [email protected]
PREFACE
Educational research endeavors to eliminate the problems of scholars,

teachers, students, administrators and policy makers. A knowledge of
2018 research methodology and statistics are indispensable for all those who
either taken an active role in the conduct of research or desire to keep
themselves abreast of the new development in education. Enhancing
the knowledge of research methodology is inevitable to scholars for
prolonging and completing their studies. Learning statistics encourage
the scholars in acquiring competency in analysis of data. In the present
© Reserved scenario, interdisciplinary research should be fostered for understanding
different disciplines. Research can help scholars for learning, finding
out facts, analyzing information, adapting their behavior according to
information received, looking to improve and adapting modern demand.
Researchers should acquire the knowledge on research problems,
selection of research methodology, sampling techniques, selection of
tools, preparation of tools, procedure of conducting research, collection
Typeset by
of data, applying appropriate statistical techniques for analyzing the
Ideal Publishing Solutions
C-90, J.D. Cambridge School, collected data , reaching the accurate result of the study and writing
West Vinod Nagar, Delhi-110092 research proposal. Learning advanced statistics is needed in educational
research for acquiring accuracy in the findings. Research in Educational
technology is indispensable for eliminating the problems of the studentsin
learning. Problems of the students in learning can be solved by using
technology based classroom transactions. Innovative research and
advanced statistics assure the quality education.
This book is specially prepared for the student-teachers of Master
Printed at of Philosophy in Educational Technology & Education and Doctor
BALAJI OFFSET of Philosophy in Educational Technology & Education. It consists of
Navin Shahdara, Delhi-110032
five units which deal with Research methods, Sampling techniques
iii
iv Research Methodology
♦
and measurement, Introduction to Statistics, Computer Application
and Report preparation.
The present book is dedicated to our parents. We extend our thanks
to all the members of faculty of Education, Educational Technology and
Department of Education (SDE) ,Bharathiar University, Coimbatore.
Dr. G. Singaravelu CONTENTS

M.A, M.Ed, M.Phil, Ph.D , D.Litt
Professor & Head, Department of Education,
Bharathiar University, Coimbatore.Tamilnadu.
and
Dr. A. Sivakumar Chapter-1: Research Methods 1
M.Sc, M.Ed, M.Phil, Ph.D 1.1. Meaning of Research 1
Guest Lecturer, Department of Educational Technology, 1.2. Nature of Research 2
Bharathiar University 1.3. Types of Research 2
1.4. Scope of Research 5
1.5. Trends in Educational Research 6
1.6. Qualitative Research Techniques 8
1.6.1. Historical Methods 9
1.6.2. Ethnographic Study 13
1.7. Quantitative Research Techniques 17
1.7.1. Survey 19
1.7.2. Experimental Research 21
1.7.3. Action Research 27
1.7.4. Observation 31
1.7.5. Interview 36
1.7.6. Questionnaire 38
1.1.7. Opinionnaire 42
1.8. Meaning and Definition of Rating Scale 44
1.9. Construction of Rating Scale 47
1.10. Scale Construction Techniques 48
1.10.1. Inventories 48
1.10.2. Check List 49
1.10.3. Achievement Test 51
v
vi Research Methodology Contents vii
♦ ♦
1.11. Norm-referenced tests (NRTs) 53 3.4. Importance of Statistics 99
1.12. Criterion-referenced tests (CRTs) 54 3.4. Limitations of Statistics 99
3.5. Source of Data 101
Chapter-2: Sampling Techniques and Measurement 59
3.6. Populations Parameters, Samples, And Statistics 103
2.1. Meaning and Characteristics of Sampling 59
3.7. Samples and Statistics 103
2.2. Types of Sampling 60
3.8. Descriptive Statistics 104
2.2.1. Simple Random Sampling 61
3.9. Inferential Statistics 107
2.2.2. Stratified Random Sampling 63 3.10. Parametric and Non Parametric Statistics 110
2.2.3. Systematic Random Sampling 64 3.11. Variable 111
2.2.4. Cluster (Area) Random Sampling 65 3.12. Types of the Variables 111
2.2.5. Multi-Stage Sampling 66 3.13. Types of Tabulation 116
2.3. Non –Probability Sampling 67 3.14. Construction of Frequency Distribution 118
2.3.1. Accidental, Haphazard or Convenience Sampling 67 3.15. Frequency Distribution of Discrete Data 119
2.3.2. Purposive Sampling 68 3.16. Diagrams and Graphs of Statistical Data 120
2.4. Theories of Probability 71
2.5. Levels of Confidence 72
2.6. Degrees of Freedom 75
2.7. Hypothesis 81
2.8. One Tailed Test and Two Tailed Tests 83
2.9. Two Tailed Test 83
2.10. One Tailed Test 84
2.11. Formulation of Hypothesis 85
2.12. Testing of Hypothesis 87
2.13. Measurement Scales 87
2.14. Nominal Scale of Measurement 88
2.15. Ordinal Scale of Measurement 88
2.16. Interval Scale of Measurement 89
2.17. Ratio Scale of Measurement 91
Chapter-3: Introduction to Statistics 92

3.1. Introduction to Statistics 92
3.2. Meaning of Statistics 93
3.3. Scope of Statistics 96
ABOUT THE AUTHOR
Dr. G. Singaravelu holds the Degrees M.A(English).,M.Ed.,M.

Phil(English).,Ph.D (Education) and D.Litt.(Doctor of Letters in
Education). He is working as Professor& Head in Department of
Education and In charge of Head of the Department of Educational
Technology, Member of Syndicate and Senate, Bharathiar University,
Coimbatore, South India. He served as Coordinators for B.Edprogramme
and CRPF in the same university. He served as in charge of Head of the
Department of Education, School of Distance Education, Bharathiar
University, He served as an Associate Professor in UGC Academic
Staff College of the same university. He has been serving as a Lecturer
in English in Arts and Science College and Lecturer in English,
District Institute of Education. He was serving as a Resource person
for DIET faculties, Block Resource Teachers and Primary Teachers,
Secondary Teachers and College Teachers. He is specialized in Primary
Education, Secondary Education, Teacher Education, Higher Education
and English Education. He published twelve dozens of research articles
and participated in 60 conferences held in various parts of India. He
chaired many seminar sessions in various universities. He is an editorial
committee member in I manager’s Journal on School Educational
Technology and Journal of English Language Teaching, Research
Demagogue etc. NCERT’s meritorious National Award crowned him
for introducing innovative gadget in the field of Mathematics. Abdul
Kalam Award was crowned him for Best teacher. He has published twenty
one books entitled English Education, Primary Education ,Methods
of Teaching English. Micro Teaching in English , Challenges in
Indian Education, Improve your English and Electronic Materials
in Teaching of English, Philosophical and Sociological perspectives
ix
x Research Methodology
♦
of Education, Educational Psychology, Problems of Education,
Research Methodology , Web Based Learning etc. He is guiding 8
Ph.D scholars in Bharathiar University. He has produced 64 M.Phil in
Education. He has produced 4 Ph.D in Educational Technology and 16
Ph.D in Education. He has completed a Major project entitled e-TLM Chapter-1
Package. He participated and presented papers in conferences held in
University of Cologne, Germany and visited Dubai as well as Sharja. RESEARCH METHODS
He participated and presented papers in conferences held in University
of Pennsylvania, Philadelphia, USA, University of Malaya, Malaysia,
Infrastructure University, Kualalumper and Nanyang University,
Singapore. He is a visiting Team member of NCTE , selection committee
member of Principals and Faculties and Inspection committee member 1.1. MEANING OF RESEARCH
of TNTEU. He has been acting members in INFITT, ISFTE, Research Research comprises “creative work undertaken on a systematic
Advisory Board and Board of study members in various universities basis in order to increase the stock of knowledge, including knowledge
of humans, culture and society, and the use of this stock of knowledge
Dr. A. Sivakumar holds the Degrees of M.Sc,,M.Ed,M.Phil(education), to devise new applications.” It is used to establish or to confirm facts, to
Ph.D(education). He has been serving as Guest Lecturer, Department reaffirm the results of previous work, to solve new or existing problems,
of Educational Technology, Bharathiar University, Coimbatore, Tamil to support theorems, or to develop new theories. A research project may
Nadu. He is specialized in Computer Education and SPSS Package. also be an expansion on past work in the field. To test the validity of
He published 10 papers and participated10 conferences in Tamil Nadu. instruments, procedures, or experiments, research may replicate elements
He published 2 books in the field of education. of prior projects, or the project as a whole. The primary purposes of
basic research (as opposed to applied research) are documentation,
discovery, interpretation, or the research (R&D) of methods and systems
for the advancement of human knowledge. Approaches to research
depend on epistemologies which vary considerably both within and
between humanities and sciences.
Educational research refers to a variety of methods, in which
individuals evaluate different aspects of education including: “student
learning, teaching methods, teacher training, and classroom dynamics”.
Educational researchers have come to the consensus that educational
research must be conducted in a rigorous and systematic way, although
what this implies is often debated. There are a variety of disciplines
which are in educational research. These include psychology, sociology,
anthropology and philosophy. The overlap in disciplines creates a broad
range from which methodology can be drawn. The findings of educational
research also need to be interpreted within the context in which they
are discovered as they may not be applicable in every time or place.
1
2 Research Methodology Research Methods 3
♦ ♦
Gary Anderson outlined ten aspects of educational research: Basic Research
• Educational research attempts to solve a problem. The research is done for knowledge enhancement, the research
• Research involves gathering new data from primary or first-hand which does not have immediate commercial potential. The research is
sources or using existing data for a new purpose. done for human welfare, animal welfare and plant kingdom welfare. It
is called basic, pure, fundamental research. The main motivation is to
• Research is based upon observable experience or empirical evidence.
expand man’s knowledge, not to create or invent something. According
• Research demands accurate observation and description. to Travers, “Basic Research is designed to add to an organized body
• Research generally employs carefully designed procedures and of scientific knowledge and does not necessarily produce results of
rigorous analysis. immediate practical value.” Such a research is time and cost intensive.
• Research emphasizes the development of generalizations, principles
Applied Research
or theories that will help in understanding, prediction and/or control.
Applied research is designed to solve practical problem of the
• Research requires expertise—familiarity with the field; competence
modern world, rather than to acquire knowledge for knowledge’s
in methodology; technical skill in collecting and analyzing the data.
sake. The goal of applied research is to improve the human condition.
• Research attempts to find an objective, unbiased solution to the It focus on analysis and solving social and real life problems. This
problem and takes great pains to validate the procedures employed. research is generally conducted on large scale basis, it is expensive.
• Research is a deliberate and unhurried activity which is directional As such, it often conducted with the support of some financing agency
but often refines the problem or questions as the research progresses. like government, public corporation, world bank, UNICEF, UGC,
• Research is carefully recorded and reported to other persons Etc,. According to hunt, “applied research is an investigation for
interested in the problem. ways of using scientific knowledge to solve practical problems”
for example:- improve agriculture crop production, treat or cure a
1.2. NATURE OF RESEARCH specific disease, improve the energy efficiency homes, offices, how
• Research strives to be objective and logical. can communication among workers in large companies be improved?
Applied research can be further classified as problem oriented and
• Research is based on observable experience and empirical evidence.
problem solving research.
• Research is characterized by patient and unhurried activity.
• Research demands accurate observation, reservation and description. Problem oriented research
• Research is directed towards the solution of the problem. Research is done by industry apex body for sorting out problems
faced by all the companies. Eg:- WTO does problem oriented research
• Research is carefully recorded and reported.
for developing countries, in India agriculture and processed food export
• Research acquires expertise. development authority (APEDA) conduct regular research for the benefit
• Research involves gathering new data from primary data or first- of agri-industry.
hand sources or using existing data for new purpose.
Problem solving
1.3. TYPES OF RESEARCH This type of research is done by an individual company for the
Research can be classified in many different ways on the basis of problem faced by it. Marketing research and market research are the
the methodology of research, the knowledge it creates, the user group, applied research. For eg:-video conference, international conducts
the research problem it investigates etc, research to study customer satisfaction level, it will be problem solving
♦ ♦
research. In short, the main aim of applied research is to discover some • Ethnography: - This type of research focuses on describing the
solution for some pressing practical problem. culture of a group of people. A culture is the shared attributes,
values, norms, practices, language, and material things of a group
Quantitative Research of people.
This research is based on numeric figures or numbers. Quantitative • Case study:-It is a form of qualitative research that is focused on
research aims to measure the quantity or amount and compares it providing a detailed account of one or more cases.
with past records and tries to project for future period. In social
• Grounded theory:- It is an inductive type of research, based or
sciences, “quantitative research refers to the systematic empirical
grounded in the observations of data from which it is developed. It
investigation of quantitative properties and phenomena and their
uses a variety of data sources, including quantitative data, review
relationships”. The objective of quantitative research is to develop
of records, interviews, observation and surveys
and to employ mathematical models, theories or hypothesis pertaining
to phenomena. • Historical research:-It allows one to discuss the past and the
The process of measurement is central to quantitative research present events in the context of the present condition and allows
because it provides fundamental connection between empirical one to reflect and to provide possible answers to current issues
observation and mathematical expression of quantitative relationships. and problems.
Statistics is the most widely used branch of mathematics in quantitative In addition to the above, the researcher also have the descriptive
research. Statistical methods are used extensively with in fields such research. Fundamental research, of which this is based on establishing
as economics and commerce. various theories, also the research is classified in to
Quantitative research involving the use of structured questions, • Descriptive research
where the response options have been pre-determined and large number
• Analytical research
of respondents is involved. Eg:-total sales of soap industry in terms of
rupees cores and quantity in terms of lakhs tones for particular year. • Fundamental research
• Conceptual research
Qualitative Research
• Empirical research
Qualitative research presents non-quantitative type of analysis.
• One time research or longitudinal research
Qualitative research is collecting, analyzing and interpreting data
by observing what people do and say. Qualitative research refers to • Field-setting research or laboratory research or simulation research
the meanings, definitions, characteristics, symbols, metaphors and • Clinical or diagnostic research
description of things. Qualitative research is much more subjective • Exploratory research
and uses very different methods of collecting information, mainly
• Historical research
individual, in-depth interviews and focus groups.
The nature of this type of research is exploratory and open ended. • Conclusion oriented research
Small numbers of people are interviewed in depth and or a relatively • Case study research
small number of focus groups are conducted. Qualitative research can • Short term research
be further classified in the following type.
• Phenomenology:-A form of research in which the researcher 1.4. SCOPE OF RESEARCH
attempts to understand how one or more individuals experience All research is goal oriented. The goal of research is three fold in
a phenomenon. nature, discovery of new knowledge, refinement of existing knowledge
♦ ♦
and finding the solution to problem. A piece of research may have one or better grasp core subjects. This type of instruction moves away from
more of these goals. Research gives birth to new knowledge or develops, the traditional lecture model of teaching and offers students a way to
elaborates and refines the existing knowledge. The new knowledge may learn at their own pace. Slower learners can take the time they need
help to improve the prevailing conditions and also correct errors and to develop a solid understanding of material while more advanced
misconceptions. In the words of J. Francis Rummel, “Research is an students don’t have to wait to move on to the next concept. This better
endeavor to discover, develop and verify knowledge. It is an intellectual addresses individual learning styles and allows teachers to utilize diverse
process that has developed over hundreds of years, every changing in platforms to provide instruction.
purpose and form and always searching for truth”. According to George
J. Mouly, “Research is a most important tool for advancing knowledge, Mixed-Age Teaching
for promoting progress and enabling man to relate more effectively to his The potential benefits of putting kids of different ages together
environment, to accomplish his purposes and to resolve his conflicts”. in the same classroom are still being researched, but many schools
Research clarifies certain problems that exist and also assist to find the have been using this teaching method for years with good results.
solution to them. To quote Mouly, “Actually research is simply the process In a mixed-age classroom, the young students learn how to interact
of arriving at dependable solution to problems through the planned and with older ones and benefit from observing their academic abilities.
systematic collection, analysis and interpretation of data”. Collaboration between age groups facilitates further development.
Achieving advanced competencies at the young age increases confidence.
1.5. TRENDS IN EDUCATIONAL RESEARCH Teachers also benefit from mixed-age classrooms in that they’re able
Looking at current trends in educational research provides an eye- to track their students through more than one grade to get a clearer
opening view of the modern classroom. If the researcher are looking picture of academic growth.
into a career in education, keeping up with changing instructional styles
prepares the researcher to enter the field with the tools the researcher Brain-Based Teaching
need to best serve today’s students. Based on neuroscience research that shows consistent practice of a
concept creates strong pathways between neurons, brain-based teaching
Team Teaching takes advantage of the power of the developing mind to create the
Putting two instructors in one classroom is known as team teaching. greatest amount of positive changes during the short period that kids
It can otherwise be called collaborative team teaching or co-teaching. are in school. Kids work in groups, participate in games that address
This model provides benefits and presents challenges to both teachers specific concepts and engage in lessons built around central themes, all
and students. Kids in co-teaching classrooms observe teamwork in with the goal of improving their understanding of important academic
action and learn to see concepts from more than one perspective. They skills and ideas. Physical education is included to facilitate the growth
have the chance to learn from teachers with diverse backgrounds and and development of new neuronal pathways.
can delve deeper into subjects thanks to the opportunity for more
one-on-one instruction time. However, conflicting personalities and Critical Thinking Skills
the tendency of students to favor one teach over the other can threaten Although memorization of fundamental facts is critical to building
the stability of this model. a foundation of knowledge in the early grades, kids also need to be
taught how to apply those facts in a multitude of situations. Research
Individualized Instruction into critical thinking skills is driving new ways of teaching that include
A great deal of research is being done on how diversification of encouraging students to ask questions, develop problem-solving skills by
instruction to cater to the unique needs of students may help them to working in groups and discuss the outcomes of their experiments. These
♦ ♦
methods of discovery give kids a more active role in their education 1.6.1. Historical Methods
and teach them how to make decisions by applying the facts they’ve Historical method comprises of the techniques and guidelines by
learned to the problem at hand. which historians use primary sources and other evidence, including the
evidence of archaeology, to research and then to write histories in the
1.6. QUALITATIVE RESEARCH TECHNIQUES
form of accounts of the past. The question of the nature, and even the
Qualitative research is a type of social science research that collects possibility, of a sound historical method is raised in the philosophy of
and works with non-numerical data and that seeks to interpret meaning history as a question of epistemology. The study of historical method
from these data that help us understand social life through the study and of different ways of writing history is known as historiography.
of targeted populations or places. People often frame it in opposition The task of the historian is to locate and interrogate appropriate
to quantitative research, which uses numerical data to identify large- sources for the reconstruction of an era. Therefore, it is important to
scale trends and employs statistical operations to determine causal and recognize the different sources used by historians and the value which
correlative relationships between variables. Within sociology, qualitative can be placed upon them. But, before discussing particular types of
research is typically focused on the micro-level of social interaction that source material it must be 0very clearly recognized that all sources
composes everyday life, whereas quantitative research typically focuses are biased. Therefore, the historian must recognize the outlook of the
on macro-level trends and phenomena. Methods of qualitative research source and take it into account when using it as evidence. The major
include observation and immersion, interviews, open-ended surveys, focus types of sources recognized by historians can be classified as follows: -
groups, content analysis of visual and textual materials, and oral history.
Primary sources
The Purpose of Qualitative Research These are the earliest available accounts of an event which are used by
Qualitative research has a long history in sociology and has been later writers to interpret that event. They are the raw material used by other
used within it for as long as the field itself has existed. This type writers to provide them with information and data. In using primary sources
of research has long appealed to social scientists because it allows it is important to recognize the value judgments made by the people who
the research to investigate the meanings that people attribute to their wrote them and the intent of the author in writing the original document.
behavior, actions, and interactions with others. While quantitative
Types of primary sources
research is useful for identifying relationships between variables,
like, for example, the connection between poverty and racial hate, it 1. Original handwritten documents, early copies of original documents,
is qualitative research that can illuminate why this connection exists letters, diaries and book manuscripts;
by going directly to the source -- the people themselves. Qualitative 2. Printed documents and published books;
research is designed to reveal the meaning that informs the action or
3. Personal documents, private documents, government documents
outcomes that are typically measured by quantitative research. So,
and public documents;
qualitative researchers investigate meanings, interpretations, symbols,
and the processes and relations of social life. What this type of research 4. Pictures, photographs and film;
produces is descriptive data that the researcher must then interpret 5. Archaeological evidence and statistical data derived from documents;
using rigorous and systematic methods of transcribing, coding, and 6. Oral evidence.
analysis of trends and themes. Because its focus is everyday life and
people’s experiences, qualitative research lends itself well to creating Secondary sources
new theories using the inductive method, which can then be tested Secondary source discuss a subject either from a great distance or after
with further research. the event on the basis of second-hand or even more remote information.
♦ ♦
Secondary sources provide interpretations and make judgments about • Data collection or literature review.
primary sources. When using secondary sources it is important to realize • Evaluation of materials.
that the account they give, even when it involves lengthy quotations from
• Data synthesis.
primary sources, may not be accurate. Secondary sources can be used to
understand primary sources but must not be confused with them. Several • Report preparation or preparation of the narrative exposition.
cautions need to be observed when working with secondary sources. Each of these steps is discussed briefly below.
Historical research is the process of systematically examining past Identification of the Research Topic and Formulation of the Research
events to give an account of what has happened in the past. Problem or Question
• It is not a mere accumulation of facts and dates or even a description This is the first step in any type of educational research including
of past events. historical research.
Ideas for historical research topics can come from many different
• It is a flowing, dynamic account of the past events which involves
sources such as current issues in education, the accomplishments
an interpretation of the these events in an attempt to recapture the
of an individual, an educational policy, or the relationship between
nuances, personalities, and ideas that influenced these events.
events.
• One of the goals of historical research is to communicate an
understanding of past events. Data Collection or Literature Review
Significance of Historical Research This step involves identifying, locating, and collecting information
pertaining to the research topic.
The following gives five important reasons for conducting historical The information sources are often contained in documents such
research (based on Berg, 1998): as diaries or newspapers, records, photographs, relics, and interviews
• To uncover the unknown (i.e., some historical events are not recorded). with individuals who have had experience with or have knowledge
• To answer questions (i.e., there are many questions about our past that of the research topic.
the researcher not only want to know but can profit from knowing). • Interviews with individuals who have knowledge of the research
• To identify the relationship that the past has to the present (i.e., topic are called oral histories.
knowing about the past can frequently give a better perspective • The documents, records, oral histories, and other information
of current events). sources can be primary or secondary sources.
• To record and evaluate the accomplishments of individuals, agencies, • A primary source is a source that has a direct involvement with
or institutions. the event being investigated like a diary, an original map, or an
• To assist in understanding the culture in which the researcher live interview with a person that experienced the event.
(e.g., education is a part of our history and our culture). • A secondary source is a source that was created from a primary
source such as books written about the event. Secondary sources
Historical Research Methodology are considered less useful than primary sources.
There is no one approach that is used in conducting historical
research although there is a general set of steps that are typically Evaluation of Materials
followed. It includes the following steps. Every information source must be evaluated for its authenticity
• Identification of the research topic and formulation of the research and accuracy because any source can be affected by a variety of factors
problem or question. such as prejudice, economic conditions, and political climate.
♦ ♦
There are two types of evaluations as follow.s Data Synthesis and Report Preparation
1. External Criticism–involves the process of determining the validity, It refers to synthesizing, or putting the material collected into a
trustworthiness, or authenticity of the source. Sometimes this is narrative account of the topic selected.
difficult to do but other times it can easily be done by handwriting • Synthesis refers to selecting, organizing, and analyzing the materials
analysis or determining the age of the paper on which something collected into topical themes and central ideas or concepts. These
was written. themes are contiguous and meaningful.
2. Internal Criticism–involves the process of determining the • It is prerequisite to watch out for these four problems that might be
reliability or accuracy of the information contained in the sources encountered when the researcher attempt to synthesize the material
collected. This is done by positive and negative criticism. collected and prepare the narrative account.
• Positive criticism refers to assuring that the statements made • Trying to infer causation from correlated events is the first problem.
or the meaning conveyed in the sources is understood. This Just because two events occurred together does not necessarily
is frequently difficult because of the problems of vagueness mean that one event was the cause of the other.
and presentism.
• The second problem is defining and interpreting key words so as to
• Vagueness refers to uncertainty in the meaning of the words avoid ambiguity and to insure that they have the correct connotation.
and phrases used in the source.
• The third problem is differentiating between evidence indicating
• Presentism refers to the assumption that the present-day how people should behave and how they in fact did behave.
connotations of terms also existed in the past.
• The fourth problem is maintaining a distinction between intent
• Negative criticism refers to establishing the reliability or and consequences. In other words, educational historians must
authenticity and accuracy of the content of the sources used. make sure that the consequences that were observed from some
This is the more difficult part because it requires a judgment activity or policy were the intended consequences.
about the accuracy and authenticity of what is contained in
the source. 1.6.2. Ethnographic Study
• Firsthand accounts by witnesses to an event are typically • Ethnographic study Involves investigation of very few cases, maybe
assumed to be reliable and accurate. just one case, in detail.
• Historians often use three heuristics in handling evidence. • Often it involves working with primarily unconstructed data. This
These are corroboration, sourcing, and contextualization. data had not been coded at the point of data collection in terms of
• Corroboration, or comparing documents to each other to a closed set of analytic categories.
determine if they provide the same information, is often used • It mphasizes on exploring social phenomena rather than testing
to obtain accurate and authentic information. hypotheses.
• Sourcing, or identifying the author, date of creation of a • Data analysis involves interpretation of the functions and meanings
document, and the place it was created is another technique that of human actions. The product of this is mainly verbal explanations,
is used to establish the authenticity or accuracy of information. where statistical analysis and quantification play a subordinate role.
• Contextualization, or identifying when and where an event • Methodological discussions focus more on questions about how
took place, is another technique used to establish authenticity to report findings in the field than on methods of data collection
and accuracy of information. and interpretation.
♦ ♦
• It focuses on describing the culture of a group in very detailed and • Collect information in the context or setting where the group works
complex manner. The ethnography can be of the entire group or or lives. This is called fieldwork. Types of information typically
a subpart of it. needed in ethnography are collected by going to the research site,
• It involves engaging in extensive field work where data collection respecting the daily lives of individuals at the site and collecting
is mainly by interviews, symbols, artifacts, observations, and many a wide variety of materials. Field issues of respect, reciprocity,
other sources of data. deciding who owns the data and others are central to ethnography
(Qualitative Inquiry and Research Design, 95)
• The researcher in ethnography type of research looks for patterns of
the groups mental activities, that is their ideas and beliefs expressed • From the many sources collected, the ethnographer analyzes the
through language or other activities, and how they behave in their data for a description of the culture-sharing group, themes that
groups as expressed through their actions that the researcher observed. emerge from the group and an overall interpretation (Wolcott,
1994b). The researcher begins to compile a detailed description
Procedures For Conducting Ethnography of the culture-sharing group, by focusing on a single event, on
several activities, or on the group over a prolonged period of
• Determining if ethnography is the most appropriate design for using time.
to study the research problem. Ethnography is suitable if the needs
• Forge a working set of rules or generalizations as to how the
are to describe how a cultural group works and to explore their
culture sharing group works as the final product of this analysis.
beliefs, language, behaviours and also issues faced by the group,
The final product is a holistic cultural portrait of the group that
such as power, resistance and dominance. (Qualitative Inquiry
incorporates the views of the participants (emic) as well as the
and Research Design, 94)
views of the researcher (etic). It might also advocate for the needs
• Then identify and locate a culture sharing group to study. This of the group or suggest changes in society. (Qualitative Inquiry
group is one whose members have been together for an extended and Research Design, 96)
period of time, so their shared language, patterns of behaviour and
The ethnographic method is different from other ways of conducting
attitudes have merged into discernible patterns. This group can
social science approach due to the following reasons:
also be a group that has been marginalized by society. (Qualitative
Inquiry and Research Design, 94) • It is field-based. It is conducted in the settings in which real people
actually live, rather than in laboratories where the researcher controls
• Select cultural themes, issues or theories to study about the
the elements of the behaviours to be observed or measured.
group. These themes, issues and theories provide an orienting
framework for the study of the culture-sharing group. As discussed • It is personalized. It is conducted by researchers who are in day-to
by Hammersley and Atkinson (2007), Wolcott (1987, 1994b, 2008- day, face-to-face contact with the people they are studying and who
1), and Fetterman (2009). The ethnographer begins the study by are both participants in and observers of the lives under study.
examining people in interaction in ordinary settings and discerns • It is multifactorial. It is conducted through the use of two or more
pervasive patterns such as life cycles, events and cultural themes. data collection techniques - which may be qualitative or quantitative
(Qualitative Inquiry and Research Design, 94-95) in nature - in order to get a conclusion.
• For studying cultural concepts, determine which type of ethnography • It requires a long term commitment i.e. it is conducted by researcher
to use. Perhaps how the group works need to be described, or a critical who intends to interact with people they are studying for an extended
ethnography can expose issues such as power, hegemony and advocacy period of time. The exact time frame can vary from several weeks
for certain groups (Qualitative Inquiry and Research Design, 95) to a year or more.
♦ ♦
• It is inductive. It is conducted in such a way to use an accumulation • Duration and cost: Research can involve prolonged time in the
of descriptive detail to build toward general patterns or explanatory field, particularly because building trust with participants is usually
theories rather than structured to test hypotheses derived from necessary for obtaining rich data.
existing theories or models. • Bias: Ethnographers bring their own experience to bear in pursuing
• It is dialogic and conducted by a researcher whose interpretations questions to ask and reviewing data, which can lead to biases in
and findings may be expounded on by the study’s participants directions of inquiry and analysis.
while conclusions are still in the process of formulation. • Descriptive approach: Ethnography relies heavily on storytelling and
• It is holistic and conducted so as to yield the fullest possible portrait the presentation of critical incidents, which is inevitably selective
of the group under study. and viewed as a weakness by those used to the scientific approaches
Ethnography, which is a method dedicated entirely to field work, of hypothesis testing, quantification and replication.
is aimed at gaining a deeper insight of a certain people’s knowledge
1.7. QUANTITATIVE RESEARCH TECHNIQUES
and social culture.
Quantitative Research
Advantages of Ethnography are
Quantitative methods emphasize objective measurements and the
• It can open up certain experiences during group research that other statistical, mathematical, or numerical analysis of data collected through
research methods fail to cover. polls, questionnaires, and surveys, or by manipulating pre-existing
• Notions that are taken for granted can be highlighted and confronted. statistical data using computational techniques. Quantitative research
• It can tap into intuitive and deep human understanding of and focuses on gathering numerical data and generalizing it across groups
interpretations of (by the ethnographer) the accounts of informants of people or to explain a particular phenomenon.
(those who are being studied), which goes far beyond what Characteristics of Quantitative Research
quantitative research can do in terms of extracting meanings.
Researchers goal in conducting quantitative research study is to
• Ethnography allows people outside of a culture (whether of a determine the relationship between one thing [an independent variable]
primitive tribe or of a corporation’s employees) to learn about its and another [a dependent or outcome variable] within a population.
members’ practices, motives, understandings and values. Quantitative research designs are either descriptive [subjects usually
However, there are certain challenges or limitations for the measured once] or experimental [subjects measured before and after
ethnographic method: a treatment]. A descriptive study establishes only associations between
• Deep expertise is required: Ethnographers must accumulate variables; an experimental study establishes causality. Quantitative
knowledge about the methods and domains of interest, which research deals in numbers, logic, and an objective stance. Quantitative
can take considerable training and time. research focuses on numeric and unchanging data and detailed, convergent
reasoning rather than divergent reasoning [i.e., the generation of a variety
• Sensitivity: The ethnographer is an outsider and must exercise
of ideas about a research problem in a spontaneous, free-flowing manner].
discretion and caution to avoid offending, alienating or harming
those being observed. Its main characteristics are
• Access: Negotiating access to field sites and participants can be
• The data is usually gathered using structured research instruments.
time-consuming and difficult. Secretive or guarded organizations
may require different approaches in order for researchers to • The results are based on larger sample sizes that are representative
succeed. of the population.
♦ ♦
• The research study can usually be replicated or repeated, given 8. Use tables to provide exact values; use figures to convey global
its high reliability. effects. Keep figures small in size; include graphic representations
• Researcher has a clearly defined research question to which objective of confidence intervals whenever possible.
answers are sought. 9. Always tell the reader what to look for in tables and figures.
• All aspects of the study are carefully designed before data is collected.
1.7.1. Survey
• Data are in the form of numbers and statistics, often arranged in
tables, charts, figures, or other non-textual forms. In research of human subjects, a survey is a list of questions aimed
at extracting specific data from a particular group of people. Surveys
• Project can be used to generalize concepts more widely, predict
may be conducted by phone, mail, via the internet, and sometimes
future results, or investigate causal relationships.
face-to-face on busy street corners or in malls. Surveys are used to
• Researcher uses tools, such as questionnaires or computer software, increase knowledge in fields such as social research and demography.
to collect numerical data. Survey research is often used to assess thoughts, opinions and
The overarching aim of a quantitative research study is to classify feelings. Surveys can be specific and limited, or they can have more
features, count them, and construct statistical models in an attempt to global, widespread goals. Psychologists and sociologists often use
explain what is observed. surveys to analyze behaviour, while it is also used to meet the more
Things to keep in mind when reporting the results of a study pragmatic needs of media, such as, in evaluating political candidates,
using quantitative methods public health officials, professional organizations, and advertising and
1. Explain the data collected and their statistical treatment as well marketing directors. A survey consists of a predetermined set of questions
as all relevant results in relation to the research problem you are that is given to a sample.[1] With a representative sample, that is, one that
investigating. Interpretation of results is not appropriate in this section. is representative of the larger population of interest, one can describe
the attitudes of the population from which the sample was drawn.
2. Report unanticipated events that occurred during your data
Further, one can compare the attitudes of different populations as well
collection. Explain how the actual analysis differs from the planned
as look for changes in attitudes over time. A good sample selection is
analysis. Explain your handling of missing data and why any
key as it allows one to generalize the findings from the sample to the
missing data does not undermine the validity of your analysis.
population, which is the whole purpose of survey research.
3. Explain the techniques you used to «clean» your data set.
4. Choose a minimally sufficient statistical procedure; provide a Types
rationale for its use and a reference for it. Specify any computer
Census
programs used.
A census is the procedure of systematically acquiring and recording
5. Describe the assumptions for each procedure and the steps you
information about the members of a given population. It is a regularly
took to ensure that they were not violated.
occurring and official count of a particular population.The term is
6. When using inferential statistics, provide the descriptive statistics, used mostly in connection with national population and housing
confidence intervals, and sample sizes for each variable as well as censuses; other common censuses include agriculture, business, and
the value of the test statistic, its direction, the degrees of freedom, traffic censuses. The United Nations defines the essential features of
and the significance level [report the actual p value]. population and housing censuses as “individual enumeration, universality
7. Avoid inferring causality, particularly in nonrandomized designs within a defined territory, simultaneity and defined periodicity”, and
or without further experimentation. recommends that population censuses be taken at least every 10 years.
♦ ♦
Other household surveys Opinion poll
Other surveys than the census may explore characteristics in An opinion poll is a survey of public opinion from a particular
households, such as fertility, family structure, and demographics. sample. Opinion polls are usually designed to represent the opinions of
Major household surveys include: a population by conducting a series of questions and then extrapolating
• General Household Survey, conducted in private households in Great generalities in ratio or within confidence intervals.
Britain. It is a repeated cross-sectional study, conducted annually,
Methodology
which uses a sample of 9,731 households in the 2006 survey.
A single survey is made of at least a sample (or full population in
• Generations and Gender Survey, conducted in several countries in
the case of a census), a method of data collection (e.g., a questionnaire)
Europe as well as Australia and Japan.Theprogramme has collected
and individual questions or items that become data that can be analyzed
least one wave of surveys in 19 countries, with an average of 9,000
statistically. A single survey may focus on different types of topics such
respondents per country.
as preferences (e.g., for a presidential candidate), opinions (e.g., should
• Household, Income and Labour Dynamics in Australia Survey, abortion be legal?), behaviour (smoking and alcohol use), or factual
where the wave 1 panel consisted of 7,682 households and 19,914 information (e.g., income), depending on its purpose. Since survey
individuals research is almost always based on a sample of the population, the success
• Integrated Household Survey, a survey made up of multiple of the research is dependent on the representativeness of the sample with
other surveys in the UK. It includes about 340,000 respondents, respect to a target population of interest to the researcher. That target
making it the largest collection of social data in the UK after the population can range from the general population of a given country to
census. specific groups of people within that country, to a membership list of a
• National Survey of Family Growth, conducted in the United professional organization, or list of students enrolled in a school system.
States by the National Center for Health Statistics division of
1.7.2. Experimental Research
the Centers for Disease Control and Prevention to understand
trends related to fertility, family structure, and demographics Experimental Research is commonly used in sciences such as
in the United States. The 2006-2010 NSFG surveyed 22,682 sociology, psychology, physics, chemistry, biology and medicine etc.
interviews. It is a collection of research designs which use manipulation and
controlled testing to understand causal processes. Generally, one or more
• Panel Study of Income Dynamics in the United States, wherein data
variables are manipulated to determine their effect on a dependent variable.
have been collected from the same families and their descendants
The experimental method is a systematic and scientific approach
since 1968. The study involved more than 9,000 individuals as of
to research in which the researcher manipulates one or more variables,
2009.
and controls and measures any change in other variables.
• Socio-Economic Panel, a longitudinal panel dataset of the Experimental Research is often used where:
population in Germany. It is a household based study which started
1. There is time priority in a causal relationship (cause precedes effect)
in 1984 and which reinterviews adult household members annually.
In 2007, the study involved about 12,000 households, with more 2. There is consistency in a causal relationship (a cause will always
than 20,000 adult persons sampled. lead to the same effect)
• UK households: a longitudinal study, now known as Understanding 3. The magnitude of the correlation is great.
Society.Its sample size is 40,000 households from the United The word experimental research has a range of definitions. In the
Kingdom or approx. 100,000 individuals. strict sense, experimental research is what the researcher call a true
♦ ♦
experiment. This is an experiment where the researcher manipulates Sampling Groups to Study
one variable, and control/randomizes the rest of the variables. It has Sampling groups correctly is especially important when researchers
a control group, the subjects have been randomly assigned between have more than one condition in the experiment. One sample group
the groups, and the researcher only tests one effect at a time. It is also often serves as a control group, whilst others are tested under the
important to know what variable(s) want to test and measure. experimental conditions.
A very wide definition of experimental research, or a quasi Deciding the sample groups can be done in using many different
experiment, is research where the scientist actively influences something sampling techniques. Population sampling may be chosen by a number
to observe the consequences. Most experiments tend to fall in between the of methods, such as randomization, “quasi-randomization” and pairing.
strict and the wide definition. A rule of thumb is that physical sciences, Reducing sampling errors is vital for getting valid results from
such as physics, chemistry and geology tend to define experiments experiments. Researchers often adjust the sample size to minimize
more narrowly than social sciences, such as sociology and psychology, chances of random errors.
which conduct experiments closer to the wider definition. Here are some common sampling techniques:
Aims of Experimental Research • probability sampling
Experiments are conducted to be able to predict phenomenons. • non-probability sampling
Typically, an experiment is constructed to be able to explain some kind • simple random sampling
of causation. Experimental research is important to society - it helps • convenience sampling
us to improve our everyday lives.
• stratified sampling
Identifying the Research Problem • systematic sampling
After deciding the topic of interest, the researcher tries to define the • cluster sampling
research problem. This helps the researcher to focus on a more narrow
• sequential sampling
research area to be able to study it appropriately. Defining the research
problem helps the researcher to formulate a research hypothesis, which • disproportional sampling
is tested against the null hypothesis. • judgmental sampling
The research problem is often operationalization, to define how • snowball sampling
to measure the research problem. The results will depend on the exact
• quota sampling
measurements that the researcher chooses and may be operationalized
differently in another study to test the main conclusions of the study. Creating the Design
An ad hoc analysis is a hypothesis invented after testing is done, to The research design is chosen based on a range of factors. Important
try to explain why the contrary evidence. A poor ad hoc analysis may factors when choosing the design are feasibility, time, cost, ethics,
be seen as the researcher’s inability to accept that his/her hypothesis measurement problems and what researcher would like to test. The
is wrong, while a great ad hoc analysis may lead to more testing and design of the experiment is critical for the validity of the results.
possibly a significant discovery.
Typical Designs and Features in Experimental Design
Constructing the Experiment
There are various aspects to remember when constructing an • Pretest-Posttest Design –It is different before the manipulation starts
experiment. Planning ahead ensures that the experiment is carried out and the effect of the manipulation. Pretests sometimes influence
properly and that the results reflect the real world, in the best possible way. the effect.
♦ ♦
• Control Group: Control groups are designed to measure research If the experiments involve human, a common strategy is to first
bias and measurement effects, such as the Hawthorne Effect or the have a pilot study with someone involved in the research, but not too
Placebo Effect. A control group is a group not receiving the same closely, and then arrange a pilot with a person who resembles the
manipulation as the experimental group. Experiments frequently have subject(s). Those two different pilots are likely to give the researcher
2 conditions, but rarely more than 3 conditions at the same time. good information about any problems in the experiment.
• Randomized Controlled Trials: Randomized Sampling, comparison
Conducting the Experiment
between an Experimental Group and a Control Group and strict
control/randomization of all other variables An experiment is typically carried out by manipulating a variable,
• Solomon Four-Group Design: With two control groups and two called the independent variable, affecting the experimental group. The
experimental groups. Half the groups have a pretest and half do effect that the researcher is interested in, the dependent variable(s),
not have a pretest. This to test both the effect itself and the effect is measured.
of the pretest. Identifying and controlling non-experimental factors which
the researcher does not want to influence the effects, is crucial
• Between Subjects Design: Grouping Participants to Different to drawing a valid conclusion. This is often done by controlling
Conditions variables, if possible, or randomizing variables to minimize effects
• Within Subject Design: Participants Take Part in the Different that can be traced back to third variables. Researchers only want to
Conditions - See also: Repeated Measures Design measure the effect of the independent variable(s) when conducting
• Counter balanced Measures Design: Testing the effect of the an experiment, allowing them to conclude that this was the reason
order of treatments when no control group is available/ethical for the effect.
• Matched Subjects Design: Matching Participants to Create Similar
Analysis and Conclusions
Experimental- and Control-Groups
• Double-Blind Experiment: Neither the researcher, nor the In quantitative research, the amount of data measured can be
participants, know which is the control group. The results can be enormous. Data not prepared to be analyzed is called “raw data”. The
affected if the researcher or participants know this. raw data is often summarized as something called “output data”, which
typically consists of one line per subject (or item). A cell of the output
• Bayesian Probability: Using Bayesian probability to “interact” data is, for example, an average of an effect in many trials for a subject.
with participants is a more “advanced” experimental design. It The output data is used for statistical analysis, e.g. significance tests,
can be used for settings were there are many variables which are to see if there really is an effect.
hard to isolate. The researcher starts with a set of initial beliefs, The aim of an analysis is to draw a conclusion, together with
and tries to adjust them to how participants have responded other observations. The researcher might generalize the results to a
Pilot Study wider phenomenon, if there is no indication of confounding variables
It may be wise to first conduct a pilot-study or do before the “polluting” the results.
researcher do the real experiment. This ensures that the experiment If the researcher suspects that the effect stems from a different
measures what it should, and that everything is set up right. Minor variable than the independent variable, further investigation is needed
errors, which could potentially destroy the experiment, are often found to gauge the validity of the results. An experiment is often conducted
during this process. With a pilot study, the researcher can get information because the scientist wants to know if the independent variable is
about errors and problems, and improve the design, before putting a having any effect upon the dependent variable. Variables correlating
lot of effort into the real experiment. are not proof that there is causation.
♦ ♦
Experiments are more often of quantitative nature than qualitative sociometric assessment often focuses on a child’s relationships with
nature, although it happens. regard to social popularity, peer acceptance, peer rejection, and
reputation. Some sociometric assessment methods derive information
Sociometry on social relationships by assessing children’s positive and negative
Sociometry is a quantitative method for measuring social relationships. social perceptions of one another, whereas other methods involve adult
It was developed by psychotherapist Jacob L. Moreno in his studies of (teacher, parent) and self- perceptions of children’s social competencies
the relationship between social structures and psychological well-being. or standing. Sociometric assessment methods were introduced in the
Sociometry, measurement techniques are used in social psychology, 1930s and advanced in the journal Sociometry. In the 1950s, several
in sociology, and sometimes in social anthropology and psychiatry based books were published on the topic and sociometric measurements
on the assessment of social choice and interpersonal attractiveness. often were part of research and school-based assessments of social
The term is closely associated with the work of the Austrian-born relationships. The use of classic sociometric procedures declined in
psychiatrist J.L. Moreno, who developed the method as a research and the following decades, due to the advancement of social behaviour
therapeutic technique. Sociometry has come to have several meanings; rating scales and ethical concerns regarding the use of peer nomination
it is most commonly applied to the quantitative treatment of preferential methods with children.
interpersonal relations, but it is also used to mean the quantitative
treatment of all kinds of interpersonal relations. The emphasis may 1.7.3. Action Research
be psychological or sociological. A succinct definition of action research appears in the workshop
A sociometric measure assesses the attractions (or repulsions) within materials the researcher use at the Institute for the Study of Inquiry in
a given group. The basic technique involves asking all group members Education. That definition states that action researchis a disciplined
to identify specific persons within the group they would prefer (or would process of inquiry conducted by and for those taking the action. The
not prefer) to have as partners in a given activity. Many variations on primary reason for engaging in action research is to assist the “actor”
this technique exist for studying different aspects of social preference. in improving and/or refining his or her actions.
For example, a group’s networking structure can be exposed through Practitioners who engage in action research inevitably find it to be
the sociometric technique of recording all interactions between group an empowering experience. Action research has this positive effect for
members. The technique can also be applied on larger scale to reveal many reasons. Obviously, the most important is that action research is
inter-organizational networks by treating organizations as individual units. always relevant to the participants. Relevance is guaranteed because
Much work has focused on the concept of sociometric status. the focus of each research project is determined by the researchers,
This includes studies of leadership; of social adjustment, ranging from who are also the primary consumers of the findings.
the social isolate (or unchosen individual) to the sociometric star (or Perhaps even more important is the fact that action research helps
highly chosen); of the relationship between sociometric status and educators be more effective at what they care most about—their teaching
other personality variables, demographic variables, and intelligence; and the development of their students. Seeing students grow is probably
and of minority-group prejudice the greatest joy educators can experience. When teachers have convincing
evidence that their work has made a real difference in their students’ lives,
Implications of socio-metric assessment for educational research the countless hours and endless efforts of teaching seem worthwhile.
Sociometric assessment can be defined as the measurement of
interpersonal relationships in a social group. Sociometric measurement The Action Research Process
or assessment methods provide information about an individual’s Educational action research can be engaged in by a single teacher or
social competence and standing within a peer group. School-based by a group of colleagues who share an interest in a common problem, or
♦ ♦
by the entire faculty of a school. Whatever the scenario, action research Collecting Data
always involves the same seven-step process. These seven steps which Professional educators always want their instructional decisions to
become an endless cycle for the inquiring teacher are the following: be based on the best possible data. Action researchers can accomplish
• Selecting a focus this by making sure that the data used to justify their actions are valid
• Clarifying theories (meaning the information represents what the researchers say it does)
and reliable (meaning the researchers are confident about the accuracy
• Identifying research questions
of their data). Lastly, before data are used to make teaching decisions,
• Collecting data teachers must be confident that the lessons drawn from the data align
• Analyzing data with any unique characteristics of their classroom or school.
• Reporting results To ensure reasonable validity and reliability, action researchers
• Taking informed action should avoid relying on any single source of data. Most teacher
researchers use a process called triangulation to enhance the validity and
Selecting a Focus reliability of their findings. Basically, triangulation means using multiple
independent sources of data to answer one’s questions. Triangulation
The action research process begins with serious reflection directed
is like studying an object located inside a box by viewing it through
toward identifying a topic or topics worthy of a busy teacher’s time.
various windows cut into the sides of the box. Observing a phenomenon
Considering the incredible demands on today’s classroom teachers, no
through multiple “windows” can help a single researcher compare and
activity is worth doing unless it promises to make the central part of a
contrast what is being seen through a variety of lenses.
teacher’s work more successful and satisfying. Thus, selecting a focus,
When planning instruction, teachers want the techniques to choose
the first step in the process, is vitally important. Selecting a focus begins
the suitable for the unique qualities of the students. All teachers have
with the teacher researcher or the team of action researchers asking:
the experience of implementing a “research-proven” strategy only
What element(s) of our practice or what aspect of student learning
to have it fail with their students. The desire of teachers is to use
do the researcher wish to investigate?
approaches that “fit” their particular students is not dissimilar to a
Clarifying Theories doctor’s concern that the specific medicine being prescribed be the
correct one for the individual patient. The ability of the action research
The second step involves identifying the values, beliefs, and process is to satisfy an educator’s need for “fit” which may be its most
theoretical perspectives the researchers hold relating to their focus. powerful attribute. Because the data being collected come from the
For example, if teachers are concerned about increasing responsible students and teachers who are engaged with the treatment, the relevance
classroom behaviour, it will be helpful for them to begin by clarifying of the findings is assured.
which approach—using punishments and rewards, allowing students For the harried and overworked teacher, “data collection” may
to experience the natural consequences of their behaviours, or some appear to be the most intimidating aspect of the entire seven-step action
other strategy—they feel will work best in helping students acquire research process. The question I am repeatedly asked, “Where will I
responsible classroom behaviour habits. find the time and expertise to develop valid and reliable instruments
for data collection?”, gives voice to a realistic fear regarding time
Identifying Research Questions
management. Fortunately, classrooms and schools are, by their nature,
Once a focus area has been selected and the researcher’s perspectives data-rich environments. Each day a child is in class, the researcher
and beliefs about that focus have been clarified, the next step is to generate is producing or not producing work, is interacting productively with
a set of personally meaningful research questions to guide the inquiry. classmates or experiencing difficulties in social situations, and is
♦ ♦
completing assignments proficiently or poorly. Teachers not only see Regardless of which venue or technique educators select for reporting
these events transpiring before their eyes, they generally record these on research, the simple knowledge that they are making a contribution to
events in their grade books. The key to managing triangulated data a collective knowledge base regarding teaching and learning frequently
collection is, first, to be effective and efficient in collecting the material proves to be among the most rewarding aspects of this work.
that is already swirling around the classroom, and, second, to identify
other sources of data that might be effectively surfaced with tests, Taking Informed Action
classroom discussions, or questionnaires. Taking informed action, or “action planning,” the last step in
the action research process, is very familiar to most teachers. When
Analyzing Data teachers write lesson plans or develop academic programs, they are
Although data analysis often brings to mind the use of complex engaged in the action planning process. What makes action planning
statistical calculations, this is rarely the case for the action researcher. particularly satisfying for the teacher researcher is that with each piece
A number of relatively user-friendly procedures can help a practitioner of data uncovered (about teaching or student learning) the educator will
identify the trends and patterns in action research data. During this feel greater confidence in the wisdom of the next steps. Although all
portion of the seven-step process, teacher researchers will methodically teaching can be classified as trial and error, action researchers find that
sort, sift, rank, and examine their data to answer two generic questions: the research process liberates them from continuously repeating their
What is the story told by these data? past mistakes. More important, with each refinement of practice, action
Why did the story play itself out this way? researchers gain valid and reliable data on their developing virtuosity.
By answering these two questions, the teacher researcher can acquire
a better understanding of the phenomenon under investigation and as 1.7.4. Observation
a result can end up producing grounded theory regarding what might Observation would seem to be an obvious method of carrying
be done to improve the situation. out research in psychology. However, there are different types of
observational methods and distinctions need to be made between:
Reporting Results
1. Controlled Observations
It is often said that teaching is a lonely endeavor. It is doubly sad
2. Natural Observations
that so many teachers are left alone in their classrooms to reinvent the
wheel on a daily basis. The loneliness of teaching is unfortunate not 3. Participant Observations
only because of its inefficiency, but also because when dealing with In addition to the above categories observations can also be either
complex problems the wisdom of several minds is inevitably better overt/disclosed (the participants know they are being studied) or covert/
than one. undisclosed (the research keeps their real identity a secret from the
The sad history of teacher isolation may explain why the very act research subjects, acting as a genuine member of the group).
of reporting on their action research has proven so powerful for both the Observations are relatively cheap to carry out and few resources are
researchers and their colleagues. The reporting of action research most needed by the researcher. However, it can often be very time consuming
often occurs in informal settings that are far less intimidating than the and longitudinal.
venues where scholarly research has traditionally been shared. Faculty
meetings, brown bag lunch seminars, and teacher conferences are among Controlled Observation
the most common venues for sharing action research with peers. However, Controlled observations are likely to be carried out in a psychology
each year more and more teacher researchers are writing up their work laboratory. The researcher decides where the observation will take
for publication or to help fulfill requirements in graduate programs. place, at what time, with which participants, in what circumstances
♦ ♦
and uses a standardized procedure. Participants are randomly allocated Sometimes the behaviour of participants is observed through a
to each independent variable group. two-way mirror or they are secretly filmed. This method was used
Rather than writing a detailed description of all behaviour observed, by Albert Bandura to study aggression in children (the Bobo doll
it is often easier to code behaviour according to a previously agreed scale studies).
using a behaviour schedule (i.e. conducting a structured observation). A lot of research has been carried out in sleep laboratories as
The researcher systematically classifies the behaviour and observe well. Here electrodes are attached to the scalp of participants and what
the distinct categories. Coding might involve numbers or letters to observed are the changes in electrical activity in the brain during sleep
describe a characteristics, or use of a scale to measure behaviour intensity. (the machine is called an electroencephalogram – an EEG).
The categories on the schedule are coded so that the data collected can Controlled observations are usually overt as the researcher explains
be easily counted and turned into statistics. the research aim to the group, so the participants know they are being
For example, Mary Ainsworth used a behaviour schedule to study observed. Controlled observations are also usually non-participant as
how infants respond to brief periods of separation from their mothers. the researcher avoids any direct contact with the group, keeping a
During the Strange Situation procedure infant’s interaction behaviours distance (e.g. observing behind a two-way mirror).
directed toward the mother were measured, e.g.
Strengths
1. Proximity and contacting seeking
2. Contact maintaining 1. Controlled observations can be easily replicated by other researchers
3. Avoidance of proximity and contact by using the same observation schedule. This means it is easy to
3. to
4. Resistance Avoidance
contact andofcomforting
proximity and contact test for reliability.
4. Resistance to contact and comforting 2. The data obtained from structured observations is easier and quicker
The observer
Thenoted downnoted
observer the behaviour displayed during
down the behaviour 15 second
displayed intervals
during 15 secondand scored the to analyze as it is quantitative (i.e. numerical) - making this a less
intervalsonand
behaviour for intensity scored
a scale of 1 the
to 7.behaviour for intensity on a scale of 1 to 7. time consuming method compared to naturalistic observations.
3. Controlled observations are fairly quick to conduct which means
that many observations can take place within a short amount of
time. This means a large sample can be obtained resulting in the
findings being representative and having the ability to be generalized
to a large population..
Limitations
1. Controlled observations can lack validity due to the Hawthorne
effect/demand characteristics. When participants know they are
being watched they may act differently.
Naturalistic Observation
Naturalistic observation (i.e. unstructured observation) involves
studying the spontaneous behaviour of participants in natural
surroundings. The researcher simply records what they see in whatever
way they can.
Sometimes the behaviour of participants is observed through a two-way mirror or they are secretly filmed.
This method was used by Albert Bandura to study aggression in children (the Bobo doll studies).
♦ ♦
Compared with controlled/structured methods it is like the difference This approach was used by Leon Festinger in a famous study
between studying wild animals in a zoo and studying them in their into a religious cult who believed that the end of the world was about
natural habitat. to occur. He joined the cult and studied how they reacted when the
With regard to human subjects Margaret Mead used this method prophecy did not come true.
to research the way of life of different tribes living on islands in the Participant observations can be either cover or overt. Covert is
South Pacific. Kathy Sylva used it to study children at play by observing where the study is carried out ‘under cover’. The researcher’s real
their behaviour in a playgroup in Oxfordshire. identity and purpose are kept concealed from the group being studied.
The researcher takes a false identity and role, usually posing as a
Strengths genuine member of the group. On the other hand, overt is where the
1 By being able to observe the flow of behaviour in its own setting researcher reveals his or her true identity and purpose to the group and
studies have greater ecological validity. asks permission to observe.
2. Like case studies naturalistic observation is often used to generate
Limitations
new ideas. Because it gives the researcher the opportunity to study
the total situation it often suggests avenues of enquiry not thought 1. It can be difficult to get time / privacy for recording. For example,
of before. with covert observations researchers can’t take notes openly as
this would blow their cover. This means they have to wait until
Limitations they are alone and reply on their memory. This is a problem
1. These observations are often conducted on a micro (small) scale as they may forget details and are unlikely to remember direct
and may lack a representative sample (biased in relation to age, quotations.
gender, social class or ethnicity). This may result in the findings 2. If the researcher becomes too involved, he may lose objectivity
lacking the ability to be generalized to wider society. and become bias. Thus reducing the validity of their data.
2. Natural observations are less reliable as other variables cannot be
Recording of Data
controlled. This makes it difficult for another researcher to repeat
the study in exactly the same way. With all observation studies an important decision the researcher
has to make is how to classify and record the data. Usually this will
3. A further disadvantage is that the researcher needs to be trained to
involve a method of sampling. The three main sampling methods are:
be able to recognise aspects of a situation that are psychologically
significant and worth further attention. 1. Event sampling. The observer decides in advance what types of
behaviour (events) she/he is interested in and records all occurrences.
4. With observations the researcher do not have manipulations of
All other types of behaviour are ignored.
variables (or control over extraneous variables) which means cause
and effect relationships cannot be established. 2. Time sampling. The observer decides in advance that observation
will take place only during specified time periods (e.g. 10 minutes
Participant Observation every hour, 1 hour per day) and records the occurrence of the
Participant observation is a variant of the above (natural observations) specified behaviour during that period only.
but here the researcher joins in and becomes part of the group they are 3. Instantaneous (target time) sampling. The observer decides in
studying to get a deeper insight into their lives. If it were research on advance the pre-selected moments when observation will take
animals the researcher would now not only be studying them in their place and records what is happening at that instant. Everything
natural habitat but be living alongside them as well! happening before or after is ignored.
♦ ♦
1.7.5. Interview 3. Phone interview—With this type of interview, interviewee speak
Interviewing is a valuable primary research method, and choosing directly to the interviewer over the phone but are not in the same
from the different interviewing techniques is the first step in the process physical location. There are still benefits of hearing the tone of
of undertaking this type of research. Interviews allow to learn more someone’s voice, adapting questions based on responses and recording
detailed information. Whether one is interviewing by an expert or the interview. This is a good choice if the researcher prefer speaking
gathering information from a small group of individuals, the researcher directly with an interview subject but constraints prevent the researcher
must decide how to conduct the interview. from meeting face-to-face or if the interviewee does not want a face-
The five most common interview techniques include the following to-face interview or one requiring the use of the internet.
interview types: 4. Email interviews—With this type of interview, one does not have any
1. Face-to-face interviews face-to-face or direct interaction with the interviewee. Interviewer
2. Webcam interviews can email the interview questions, and receive the responses via
3. Telephonic interviews email. The benefit of this type of interview is that a transcript in
digital form is already made once the responses are returned to
4. Email interviews
the interviewer. However, one cannot watch for social cues or ask
5. Instant message/chat interviews additional questions based off responses in real time. Sometimes
With each interviewing technique, there are advantages and these results in receiving less information than the researcher could
disadvantages, so choosing the right one for primary research method is via a telephone, webcam or face-to-face interview.
done by considering the value of each type, time or location constraints 5. Chat/instant message interviews—With this type of interview,
and the personal preferences of researcher.. the researcher conduct the researcher interview over an instant
Interviewing techniques messaging service, such as Yahoo! Messenger, Google Talk or
There are five main interviewing techniques the researcher can MSN Messenger. Since this type of interviewing technique is
use as a primary research method. done in real time, the researcher have the benefit of adapting the
researcher questions or asking additional questions based on the
1. Face-to-face interview—With this type of interview, Interviewer
responses the researcher receive. The length of responses might
and interviewee are in the same location and are speaking directly to
be shorter with this type of interview, but one can save a copy of
one another. Benefits of this type of interviewing technique include
the chat session for an already-made digital copy of the interview.
being able to see and hear social cues, to adapt the researcher
questions based on how the person the researcher are interviewing Interviewing technique tips
responds and to record a copy of the interview that the researcher
When interviewing technique is used as a primary research method,
can later be transcribed to use as a reference. This is often the best
there are certain steps during the preparation and interviewing processes
option if it meets the needs and availability of both the researcher
that facilitate a more effective interview.
and the researcher interview subjects.
2. Webcam interviews—With this type of interview, Interviewer and Preparation tips
interviewee are face-to-face via webcam technology. It offers the same
1. The interviewer shall learn interviewee prior to the interview in
benefits as a traditional face-to-face interview, and recording either the
order to form questions that are specific to the person.
audio or the video output is a good idea in order to transcribe it and
refer back to the transcript as a reference. This type of interviewing 2. Plan ahead, and set aside an adequate amount of time that suits
technique is also useful when distance prohibits a face-to-face interview. the interviewing technique.
♦ ♦
3. Check the recording equipment in advance to make sure everything Advantages of the close form
is working correctly, and have a backup recorder ready to go in
1. It is easy to fill out.
case of technical difficulties.
2. It takes little time by respondents
4. Prepare the researcher questions carefully for any type of interview
technique by spending the time to write good interview questions. 3. It is relatively objective
4. Easy to tabulate and analyze
Interviewing tips
5. Answers are standardized.
1. Start the interview stating the interview subject at ease.
2. Respect the interviewee by following their wishes if the researcher Limitations of the close form
subject chooses to not answer a specific question. It fails to reveal the respondents’ motives and does not always get
3. Stay ready to adapt the questions if an answer to one of the questions information of sufficient scope and in depth and may not discriminate
triggers another question that pertains to the research. between the finer shades of meaning.
4. Keep the interview focused by redirecting the conversation with The open form
additional questions when the person the researcher are interviewing
The open form or unstructured type of questionnaire calls for a
strays too far from the focus of the researcher interview questions.
free response in respondents own words.
Interviews are a powerful primary research method, and the
information the researcher learn by completing any of the interviewing Advantages of the open form questionnaire
techniques adds value to the research.
1. Open end questions are flexible.
1.7.6. Questionnaire 2. They can be used when all possible answer categories are not known.
Questionnaire is a self- report data collection instrument that each 3. They are preferable to complex issues that cannot be condensed.
research participant fills out as part of a research study. Researchers 4. They allow more opportunity for creativity, thinking and
use questionnaire to obtain information about the thoughts, feelings, self-expression.
attitudes beliefs, values, perceptions, personality and behavioural
intentions of research participants. Limitation
According to John W. Best (1992) a questionnaire is used when
1. There is possibility of collection of worthless or irrelevant
factual information is desired, when opinion rather than facts are desired,
information.
an opinionnaire or Attitude scale is used.
2. Data collected through open end questionnaire are not often
Forms/Kinds Of Questionnaire standardized from person to person.
The researcher can construct questions in the form of a closed, 3. Coding is difficult and subjective.
open pictorial and scale items.
Pictorial form
1. Close Form Some questionnaires present respondents with drawings and
Questionnaire that calls for short check responses as the, restricted photographs rather than written statement from which to choose answers.
or close form type. They provide for marking a Yes or No a short This form of questionnaire is particularly suitable tool for collecting
response or checking an item from a list of suggested responses. data from children and adults who had not developed reading ability.
♦ ♦
Pictures often capture the attention of respondents more readily than • The sponsor
printed words, lessen subjects’ resistance and stimulate the interest in • The purpose
questions. “To get better answers, ask better questions”
• Return address and date
Characteristics of a Good Questionnaire • Confidentiality
• Questionnaire should deal with important or significant topic to • Voluntary responses and
create interest among respondents. • Thanks
• It should seek only that data which cannot be obtained from other Instructions to the Respondent
sources.
It is very important that respondents are instructed to go presented
• It should be as short as possible but should be comprehensive. at the start of the questionnaire which indicates what is expected from
• It should be attractive. the respondents. Specific instructions should be given for each question
• Directions should be clear and complete. where the style of questions varies throughout the questionnaire. For
• It should be represented in good Psychological order proceeding Example – Put a tick mark in the appropriate box and circle the relevant
from general to more specific responses. number etc.
• Double negatives in questions should be avoided. The Allocation of Serial Numbers

• Putting two questions in one question also should be avoided. Whether dealing with small or large numbers, a good researcher
• It should avoid annoying or embarrassing questions. needs to keep good records. Each questionnaire therefore should be
numbered.
• It should be designed to collect information which can be used
Coding Boxes:
subsequently as data for analysis.
When designing the questionnaire, it is necessary to prevent later
• It should consist of a written list of questions. complications which might arise at the coding stage. Therefore, the
• The questionnaire should also be used appropriately. researcher should note the following points:
Designs of Questionnaire: • Locate coding boxes neatly on the right hand side of the page.
After construction of questions on the basis of its characteristics • Allow one coding box for each answer.
it should be designed with some essential routines like:
• Identify each column in the complete data file underneath the
• Background information about the questionnaire. appropriate coding box in the questionnaire.
• Instructions to the respondent. Besides these, the researcher should also be very careful about the
• The allocation of serial numbers and length and appearance of the questionnaire, wording of the questions,
• Coding Boxes. order and types of questions while constructing a questionnaire
Advantages of Questionnaire
Background Information about
Questionnaire is economical. In terms of materials, money and
The Questionnaire Both from ethical and practical point of view,
time it can supply a considerable amount of research data.
the researcher needs to provide sufficient background information about
the research and the questionnaire. Each questionnaires should have a • It is easier to arrange.
cover page, on which some information appears about: • It supplies standardized answers
♦ ♦
• It encourages pre-coded answers. Characteristics
• It permits wide coverage. • The opinionnaire makes use of statements or questions on different
• It helps in conducting depth study. aspects of the problem under investigation.
Disadvantages • Responses are expected either on three point or five point scales.
• It uses favourable or unfavourable statements.
• It is reliable and valid, but slow.
• It may be sub-divided into sections.
• Pre-coding questions can deter them from answering.
• The gally poll ballots generally make use of questions instead of
• Pre-coded questions can bias the findings towards the researcher.
statements.
• Postal questionnaire offer little opportunities to check the truthfulness
• The public opinion polls generally rely on personal contacts rather
of the answers.
than mail ballots.
• It can not be used with illiterate and small children.
Schedules
1.1.7. Opinionnaire
“Opinion polling or opinion gauging represents a single question • A schedule is generally filled by the research worker or enumerator,
approach. The answers are usually in the form of ‘yes’ or ‘no’. An who can interpret the questions when necessary.
undecided category is often included. Sometimes large number of • Data collection is more expensive as money is spent on enumerators
response alternatives if provided.” - Anna Anastusi and in imparting trainings to them. Money is also spent in preparing
The terms opinion and attitude are not synonymous, through schedules.
sometimes the researcher used it synonymously. the researcher • Non response is very low because this is filled by enumerators
have till now discussed that attitudes scale. the researcher have also who are able to get answers to all questions. But even in this their
discussed that attitudes are impressed opinions. The researcher can remains the danger of interviewer bias and cheating.
now understand the difference between opinionnaire and attitude scale,
• Identity of respondent is not known.
when the researcher discuss of out opinionnaire, it is characteristics and
purposes. Opinion is what a person says on certain aspects of the issue • Information is collected well in time as they are filled by
under considerations. It is an outward expression of an attitude held by enumerators.
an individual. Attitudes of an individual can be inferred or estimated • Direct personal contact is established.
from his statements of opinions. An opinionnaire is defined as a special • The information can be gathered even when the respondents happen
form of inquiry. It is used by the researcher to collect the opinions of to be illiterate.
a sample of population on certain facts or factors the problem under • There remains the difficulty in sending enumerators over a relatively
investigation. These opinions on different facts of the problem under wider area.
study are further quantified, analysed and interpreted.
• The information collected is generally complete and accurate as
Purpose enumerators can remove difficulties if any faced by respondents in
Opinionnaire are usually used in researches of the descriptive type correctly understanding the questions. As a result the information
which demands survey of opinions of the concerned individuals. Public collected through schedule is relatively more accurate than that
opinion research is an example of opinion survey. Opinion polling enables obtained through questionnaires.
the researcher to forecast the coming happenings in successful manner. • It depends upon the honesty and competence of enumerators
♦ ♦
• This may not be the case as schedules are to be filled in by Response options
enumerators and not by respondents.
• Strongly disagree
• Along with schedule observation method can also be used.
• Disagree
1.8. MEANING AND DEFINITION OF RATING SCALE • Agree
A rating scale is a set of categories designed to elicit information • Strongly agree
about a quantitative or a qualitative attribute. In the social sciences, 2. Some data are measured at the interval level. Numbers indicate the
particularly psychology, common examples are the Likert response magnitude of difference between items, but there is no absolute
scale and 1-10 rating scales in which a person selects the number which zero point. A good example is a Fahrenheit/Celsius temperature
is considered to reflect the perceived quality of a product. scale where the differences between numbers matter, but placement
A rating scale is a method that requires the rater to assign a value, of zero does not.
sometimes numeric, to the rated object, as a measure of some rated attribute.
3. Some data are measured at the ratio level. Numbers indicate
Types of rating scales magnitude of difference and there is a fixed zero point. Ratios
Rating Scales allow teachers to indicate the degree or frequency can be calculated. Examples include age, income, price, costs,
of the behaviours, skills and strategies displayed by the learner. To sales revenue, sales volume and market share.
continue the light switch analogy, a rating scale is like a dimmer switch More than one rating scale question is required to measure an
that provides for a range of performance levels. Rating scales state the attitude or perception due to the requirement for statistical comparisons
criteria and provide three or four response selections to describe the between the categories in the polytomousRasch model for ordered
quality or frequency of student work. categories. In terms of Classical test theory, more than one question is
Teachers can use rating scales to record observations and students can required to obtain an index of internal reliability such as Cronbach’s
use them as self-assessment tools. Teaching students to use descriptive alpha, which is a basic criterion for assessing the effectiveness of a
words, such as always, usually, sometimes and never helps them pinpoint rating scale and, more generally, a psychometric instrument.
specific strengths and needs. Rating scales also give students information for
setting goals and improving performance. In a rating scale, the descriptive Scales
word is more important than the related number. The more precise and 1. Graphic Rating Scale
descriptive the words for each scale point, the more reliable the tool. A graphic rating scale, also known as a continuous rating scale
Effective rating scales use descriptors with clearly understood usually looks like the figure drawn above. The ends of the continuum
measures, such as frequency. Scales that rely on subjective descriptors are sometimes labeled with opposite values. Respondents are required
of quality, such as fair, good or excellent, are less effective because the to make a mark at any point on the scale that they find appropriate.
single adjective does not contain enough information on what criteria Sometimes, there are numbers along the markings of the line too. At
are indicated at each of these points on the scale. other times, there are no markings at all on the line.
All rating scales can be classified into one of three classifications:
1. Some data are measured at the ordinal level. Numbers indicate 2. Likert Scale
the relative position of items, but not the magnitude of difference. A Likert scale typically contains an odd number of options, usually
Attitude and opinion scales are usually ordinal; one example is a 5 to 7. One end is labeled as the most positive end while the other one
Likert response scale: is labeled as the most positive one with the label of ‘neutral’ in the
Statement: e.g. “I could not live without my computer”. middle of the scale.
♦ ♦
The phrases ‘purely negative’ and ‘mostly negative’ could also 1.9. CONSTRUCTION OF RATING SCALE
have been ‘extremely disagree’ and ‘slightly disagree’. Steps in constructing rating scale
3. Semantic Differential Scale (Max Diff) 1. Decide what areas are to be measured
A semantic scale is a combination of more than one continuum. It 2. For each area decide what characteristics are to be measured
usually contains an odd number of radio buttons with labels at opposite 3. Define a range for each characteristic
ends. Max Diff scales are often used in trade-off analysis such as conjoint. • Decide how many points on the scale
MaxDiff analysis can be used in new product features research or
• State extremes- very good and very bad
or even market segmentation research to get accurate orderings of the
most important product features. Discriminate among feature strengths • State points between these extremes
more effectively than derived importance methodologies. Like other 4. Arrange items to form the scale
trade-off analyses, the analysis derives utilities for each of the most 5. Design directions
important product features which can be used to derive optimal products, 6. Pilot test scale
using market segmentation to put respondents into groups with similar
7. Make needed revisions, based on pilot test
preference structures, or to prioritize strategic product goals.
The researcher can have the researcher respondents perform Forced- Basic Rules
choice nature of the tasks, and disentangle the relative feature importance in • Include basic information
cases where average Likert-style ratings might all have very similar ratings.
• Student’s name
4. Side-by-Side Matrix • Rater’s name
Another very commonly used scale in questionnaires is the side- • Rater’s position
by-side matrix. A common and powerful application of the side-by-side • Setting in which student was observed
matrix is the importance/satisfaction type of question. • Rating period (From ___ to ___)
First, ask the respondent how important an attribute is, then ask
• Date scale was completed
them how satisfied they are with the researcher performance in this
area. QuestionPro’s logic and loop functions also allow the researcher • Other information important to the researcher
to run through this question multiple times with other alternatives that • Decide on odd or even number of responses
the respondent might consider. This yields benchmark data that will • Decide whether or not to group items with same content together
allow the researcher to compare the researcher performance against • Allow space for comments after each item
other competing alternatives.
• Allow space for comments at end of scale
Here is an example of data from an importance/satisfaction question.
The importance rating is the line and the performance ratings are the • Write specific directions, including the purpose of scale and how
bars. With this type of data, the researcher can actually see where the to complete
researcher company needs to increase its efforts to more closely meet • Put labels at tope of response choices (on every page)
the needs of the customer.
Principles For Preparation Rating Scale
While there are many online survey tools and online survey software
to choose from, will find that not all of them have these different types • Use action oriented precise verbs.
of scales available to them. • Each item should deal with important content area
♦ ♦
• Question can be as long as possible, but answer should be short. are sorted into various interest areas and percentile scores calculated
• Use precise, simple and accurate language relation to the subject for each area. The area where a person’s percentile scores are relatively
matter area. higher is considered to be the area of his greatest interests, the area in
• Provide the necessary space for answers below each question asked. which the researcher would be the happiest and the most successful.
As a part of educational surveys of many kinds, children’s interest in
1.10. SCALE CONSTRUCTION TECHNIQUES reading, in games, in dramatics, in other extracurricular activities and in
curricular work etc. is studied. One kind of instrument, most commonly
• Arbitrory approach- scales on ad hoc basis. used in interest measurement is known as Strong’s Vocational Interest
• Consensus approach- panel of judges evaluate Inventory. It compares the subject’s pattern of interest to the interest
• Item analysis approach-individual items into test. patterns of successful individuals in a number of vocational fields.
• Cumulative scales- ranking of items. This inventory consists of the 400 different items. The subject has
to tick mark one of the alternatives i. e. L(for like), I(indifference)
• Factors scales-inter correlation of items.
or D(Dislike) provided against each item. When the inventory is
1.10.1. Inventories standardized, the scoring keys and percentile norms are prepared
on the basis of the responses of a fairly large number of successful
Inventory is a list, record or catalog containing list of traits,
individuals of a particular vocation. A separate scoring key is therefore
preferences, attitudes, interests or abilities used to evaluate personal
prepared for each separate vocation or subject area. The subject’s
characteristics or skills. The purpose of inventory is to make a list about
responses are scored with the scoring key of a particular vocation
a specific trait, activity or programme and to check to what extent the
in order to know his interest or lack of interest or lack of interest in
presence of that ability types of Inventories like
the vocation concerned. Similarly his responses can be scored with
• Personality Inventory scoring keys standardized for other vocational areas. In this way the
• Interest Inventory researcher can determine one’s areas of vocational interest. Another
Persons differ in their interests, likes and dislikes. Interests are well-known interest inventories, there are also personality inventories
significant element in the personality pattern of individuals and play to measure the personality.
an important role in their educational and professional careers. The
1.10.2. Check List
tools used for describing and measuring interests of individuals are the
internet inventories or interest blanks. They are self report instruments A checklist is a type of informational job aid used to reduce failure
in which the individuals note their own likes and dislikes. They are by compensating for potential limits of human memory and attention.
of the nature of standardized interviews in which the subject gives It helps to ensure consistency and completeness in carrying out a task.
an introspective report of his feelings about certain situations and A basic example is the “to do list.” A more advanced checklist would
phenomena which is then interpreted in terms of internets. be a schedule, which lays out tasks to be done according to time of
The use of interest inventories is most frequent in the areas of day or other factors.
educational and vocational guidance and case studies. Distinctive Applications
patterns of interest that go with success have been discovered through
research in a number of educational and vocational fields. Mechanical, 1. pre-flight checklists aid in aviation safety to ensure that critical
computational, scientific, artifice, literary, musical, social service, items are not forgotten
clerical and many other areas of interest have been analyzed informs 2. Use in medical practice to ensure that clinical practice guidelines
of activities. In terms of specific activities, a person’s likes and dislikes are followed. An example is the Surgical Safety Checklist developed
♦ ♦
for the World Health Organization by Dr. AtulGawande. Evidence Actions from the checklist include
to support surgical checklists is tentative but limited.
1. ensuring absolute clarity about the expected standard of pupils’
3. Used in quality assurance of software engineering, to check process behaviour
compliance, code standardization and error prevention, and others.
2. Displaying school rules clearly in classes and around the building.
4. Often used in industry in operations procedures. Staff and pupils should know what they are
5. Used in civil litigation to deal with the complexity of discovery and 3. ensuring that children actually receive rewards every time they
motions practice. An example is the open-source litigation checklist. have earned them and receive a sanction every time they behave
6. Used by some investors as a critical part of their investment process badly
7. Can aid in mitigating claims of negligence in public liability claims 4. taking action to deal with poor teaching or staff who fail to follow
by providing evidence of a risk management system being in place. the behaviour policy
8. An ornithological checklist, a list of birds with standardized names 5. ensuring pupils come in from the playground and move around
that helps ornithologists communicate with the public without the the school in an orderly manner
use of scientific names in Latin. 6. ensuring that the senior leadership team like the head and assistant
9. A popular tool for tracking sports card collections. Randomly head are a visible presence around the school during the day, including
inserted in packs, checklist cards provide information on the contents in the lunch hall and playground, and are not confined to offices
of sports card set.
1.10.3. Achievement Test
Format
Achievement test is an important tool in school evaluation and has
Checklists are often presented as lists with small checkboxes down great significance in measuring instructional progress and progress of
the left hand side of the page. A small tick or checkmarks drawn in the the students in the subject area.
box after the item has been completed. Achievement means one’s learning attainments, accomplishments,
Other formats are also sometimes used. Aviation checklists generally proficiencies, etc. It is directly related to the pupil’s growth and
consist of a system and an action divided by a dashed line, and lacks a development in educational situations.
checkbox as they are often read aloud and are usually intended to be reused. Tests should give an accurate picture of students’ knowledge and skills
Check List In Education in the subject area or domain being tested. Accurate achievement data are
A simple checklist of what schools can do to instill good behaviour very important for planning curriculum and instruction and for program
in the classroom has been developed and published today by Charlie evaluation. Test scores that overestimate or underestimate students’ actual
Taylor - the head teacher of a special school with some of the toughest knowledge and skills cannot serve these important purposes.
behaviour issues and the government’s expert adviser on behaviour. Definition
The behaviour checklist - entitled ‘Getting the simple things right’
- follows Charlie Taylor’s recent behaviour summit, where outstanding “Any test that measures the attainments and accomplishments of
head teachers from schools in areas of high deprivation gathered to an individual after a period of training or learning”.
discuss the key principles for improving behaviour. What soon became -NM Downie
clear was how much similarity there was between the approaches that ‘The type of ability test that describes what a person has learned
the head teachers followed. Many of them emphasised the simplicity of to do’
their approach but they agreed that most important of all was consistency. –Throndike and Hagen
♦ ♦
“A systematic procedure for determining the amount a student has • It should be divided into different knowledge and skills according
learned through instructions” to behaviour to be measured.
–Groulund • Standardized the items and made instructions clear so that different
Objectives users can utilize it.
• Rules and norms have to be developed so that various age groups
• Identify and explain reasons for performing tests. can use at various levels.
• Understand testing terminology to communicate clearly with • It provides equivalent and comparable forms of the test.
students and colleagues.
• A test manual has to be prepared, which can act as a guide for
• Evaluate a test’s validity and reliability. administering and scoring.
• Select appropriate tests.
• Administer test protocols properly and safely. 1.11. NORM-REFERENCED TESTS (NRTS)
A person’s score on a norm-referenced test (NRT) shows where the
Functions of test person stands relative to other people who have taken the test (Seels&
• It provides basis for promotion to the next grade. Glasgow, 1990). The Scholastic Aptitude Test (SAT) is an example of
• To find out where each student stands in various academic areas. a NRT because it shows how a person stands relative to other potential
college students. Success is measured by how far ahead a person is
• It helps in determination about the placement of the students in a
of the other test takers. NRTs are usually used for selection purposes.
particular section.
NRTs are designed to measure relative standing in a group rather than
• To motivate the students before a new assignment has taken up. mastery of a specific skill.
• To know effectively the student is performing in theory as well The norm referenced test is easiest to use when comparing student
as in clinical areas. progress and performance. These tests provide a quick view of the
• To expose pupil’s difficulties which the teacher can help them to solve. things that students are expected to learn and are cheap and easy to
administer.
Characteristics of a good test Norm referenced testing assumes that human traits are distributed
Test preparation activities which promote quality, long-term learning along a normal probability or normal curve. This curve represents the
are appropriate, even essential. Good test-taking skills and appropriate average, high and low test scores of a population. A norm referenced
content learning can reduce the likelihood that extraneous factors will test compares test performance amongst students of the same age.
influence students’ test scores. The various characteristics of a good test are: While there are advantages to norm referenced testing, the
• It can be tried out and selected on the basis of its difficulty level disadvantages outweigh them. These tests can only measure a small
and discriminating power. part of a subject and a minor section of human abilities. Most of these
types of tests focus too much on memorization and routine procedures,
• Directly related to the educational objectives.
leaving out critical thinking and reasoning. Whether a student can write
• It should possess description of measure behaviour in realistic and a quality paper, understand the impact of science and current events or
practical terms. has the ability to debate can not be tested with norm referenced tests
• Contains a sufficient number of test items for each measured due to their multiple-choice and short answer questions. These tests
behaviour; concerned with important and useful matter; can cause academic expectations to be lower for those students who
comprehensive, brief, precise and clear. have low scores
♦ ♦
1.12. CRITERION-REFERENCED TESTS (CRTS) special needs to average students will not achieve much in measuring
Criterion-referenced tests (CRTs) are also referred to as content- the student’s current level of performance.
referenced or objective-referenced tests. “A test is criterion-referenced
Intended Purposes
when its score can be translated into a statement about what a person
has learned relative to a standard; a CRT score provides information The major reason for using norm-referenced tests (NRT) is to classify
about a person’s mastery of a behaviour relative to the objective and students. NRTs are designed to highlight achievement differences between
reflects that person’s mastery of one specific skill”. A state’s automobile and among students to produce a dependable rank order of students
driving test is an example of a CRT.Typically a cutoff score is set and across a continuum of achievement from high achievers to low achievers
those who meet or exceed to score pass the test. Any number of test (Stiggins, 1994). School systems might want to classify students in this
takers can pass a CRT. way so that they can be properly placed in remedial or gifted programs.
These types of tests are also used to help teachers select students for
different ability level reading or mathematics instructional groups.
Disadvantages
With norm-referenced tests, a representative group of students
Although these assessments are becoming more popular in the is given the test prior to its availability to the public. The scores of
special education field they do have some drawbacks. These include: the students who take the test after publication are then compared to
• It does not allow for comparing the performance of students in a those of the norm group. Tests such as the California Achievement Test
particular location with national norms. For example, a school would (CTB/McGraw-Hill), the Iowa Test of Basic Skills (Riverside), and the
be unable to compare 5th grade achievement levels in a district, Metropolitan Achievement Test (Psychological Corporation) are normed
and therefore be unable to measure how a school is performing using a national sample of students. Because norming a test is such an
against other schools. elaborate and expensive process, the norms are typically used by test
• It is time-consuming and complex to develop. Teachers will be required publishers for 7 years. All students who take the test during that seven
to find time to write a curriculum and assessments with an already year period have their scores compared to the original norm group.
full work-load. It might require more staff to come in and help. While norm-referenced tests ascertains the rank of students,
• It costs a lot of money, time and effort. Creating a specific curriculum criterion-referenced tests (CRTs) determine “...what test takers can
takes time and money to hire more staff; and most likely the staff do and what they know, not how they compare to others (Anastasi,
will have to be professionals who have experience. 1988, p. 102). CRTs report how well students are doing relative to a
pre-determined performance level on a specified set of educational
• It needs efficient leadership and collaboration, and lack of leadership
goals or outcomes included in the school, district, or state curriculum.
can cause problems - for instance, if a school is creating assessments
Educators or policy makers may choose to use a CRT when they
for special education students with no well-trained professionals,
wish to see how well students have learned the knowledge and skills
they might not be able to create assessments that are learner-centered.
which they are expected to have mastered. This information may be
• It may slow the process of curriculum change if tests are constantly used as one piece of information to determine how well the student is
changed. It is diifficult for curriculum developers to know what is learning the desired curriculum and how well the school is teaching
working and what is not working because tests tend to be different that curriculum.
from one school to another. It would require years of collecting Both NRTs and CRTs can be standardized. The U.S. Congress,
data to know what is working and what is not. Office of Technology Assessment (1992) defines a standardized test
Despite it’s flaws, criterion-referenced assessments will still be as one that uses uniform procedures for administration and scoring in
important in special education because comparing scores of students with order to assure that the results from different people are comparable.
♦ ♦
Any kind of test--from multiple choice to essays to oral examinations- information which is found on the standardized tests. This is particularly
-can be standardized if uniform scoring and administration are used true if the standardized tests are also used to measure an educator’s
(p. 165). This means that the comparison of student scores is possible. teaching ability. The result of this pressure placed upon teachers for
Thus, it can be assumed that two students who receive the identical their students to perform well on these tests has resulted in an emphasis
scores on the same standardized test demonstrate corresponding levels on low level skills in the classroom (Corbett & Wilson, 1991). With
of performance. Most national, state and district tests are standardized curriculum specialists and educational policy makers alike calling for
so that every score can be interpreted in a uniform manner for all more attention to higher level skills, these tests may be driving classroom
students and schools. practice in the opposite direction of educational reform.
The major difference between criterion-referenced and norm-
Selection Of Test Content referenced assessments has to do with how a student’s performance is
Test content is an important factor choosing between an NRT test measured. In norm-referenced assessments, the students are compared
and a CRT test. The content of an NRT test is selected according to to other students using a percentile system. In criterion-referenced
how well it ranks students from high achievers to low. The content of assessments, the students level of performance is measured according
a CRT test is determined by how well it matches the learning outcomes to a standard that the student has to meet.
deemed most important. Although no test can measure everything of For example, if a swimmer’s speed on the 100 meters lap is
importance, the content selected for the CRT is selected on the basis compared to other swimmers using the norm-referenced concept, the
of its significance in the curriculum while that of the NRT is chosen examiner would report that the swimmer who swam the 100 meters in
by how well it discriminates among students. one minute was in the 70thpercentile. While using the criterion assessment
Any national, state or district test communicates to the public the the examiner would report that the standard for swimming 100 meters
skills that students should have acquired as well as the levels of student is 50 seconds and the swimmer did it in one minute, 10 seconds short
performance that are considered satisfactory. Therefore, education of the criterion.
officials at any level should carefully consider content of the test which
is selected or developed. Because of the importance placed upon high Research Design
scores, the content of a standardized test can be very influential in the Different textbooks place different meanings on research design.
development of a school’s curriculum and standards of excellence. Some authors consider research design as the choice between qualitative
NRTs have come under attack recently because they traditionally and quantitative research methods. Others argue that research design
have purportedly focused on low level, basic skills. This emphasis is in refers to the choice of specific methods ofdata collection and analysis.
direct contrast to the recommendations made by the latest research on In the researcher dissertation the researcher can define research design
teaching and learning which calls for educators to stress the acquisition as a general plan about what the researcher will do to answer the
of conceptual understanding as well as the application of skills. The research question. Important elements of research design include
National Council of Teachers of Mathematics (NCTM) has been research strategies and methods related to data collection and analysis.
particularly vocal about this concern. In an NCTM publication (1991), Research design can be divided into two groups: exploratory and
Romberg (1989) cited that “a recent study of the six most commonly conclusive. Exploratory research, according to its name merely aims to
used commercial achievement tests found that at grade 8, on average, explore specific aspects of the research area. Exploratory research does
only 1 percent of the items were problem solving while 77 percent not aim to provide final and conclusive answers to research questions.
were computation or estimation” (p. 8). The researcher may even change the direction of the study to a certain
In order to best prepare their students for the standardized extent, however not fundamentally, according to new evidences gained
achievement tests, teachers usually devote much time to teaching the during the research process.
58 Research Methodology
♦
The following table illustrates the main differences between exploratory
and conclusive research in relation to important components of a dissertation:
Major Differences Between Exploratory And Conclusive Research Design
Research project Exploratory research Conclusive research Chapter-2
components
Research purpose General: to generate Specific: to verify insights SAMPLING TECHNIQUES AND MEASUREMENT
insights about a situation and aid in selecting a course
of action
Data needs Vague Clear
Data sources Ill defined Well defined
2.1. MEANING AND CHARACTERISTICS OF SAMPLING
Data collection Open-ended, rough Usually structured
In the Research Methodology, practical formulation of the research
form
is very much important and so it should be done very carefully with
Sample Relatively small; Relatively large; objectively proper concentration and in the presence of a very good guidance. But
subjectively selected to selected to permit
maximize generalization generalization of findings
during the formulation of the research on the practical grounds, one
of insights tends to go through a large number of problems. These problems are
generally related to the knowing of the features of the universe or the
Data collection Flexible; no set Rigid; well-laid-out
procedure procedure population on the basis of studying the characteristics of the specific
part or some portion, generally called as the sample. So now sampling
Data analysis Informal; typically Formal; typically
can be defined as the method or the technique consisting of selection
non-quantitative quantitative
for the study of the so called part or the portion or the sample, with
Inferences/ More tentative than final More final than tentative a view to draw conclusions or the solutions about the universe or the
Recommendations
population.
The following can be mentioned as examples with exploratory design: According to Mildred Parton, “Sampling method is the process
or the method of drawing a definite number of the individuals, cases
1. A critical analysis of argument of mandatory CSR for UK private
or the observations from a particular universe, selecting part of a total
sector organisations
group for investigation.”
2. A study into contradictions between CSR program and initiatives
and business practices: a case study of Philip Morris USA Basic Principles of Sampling
3. An investigation into the ways of customer relationship management Theory of sampling is based on the following laws-
in mobile marketing environment • Law of Statistical Regularity – This law comes from the
4. Studies listed above do not aim to generate final and conclusive mathematical theory of probability. According to King,” Law of
evidences to research questions. These studies merely aim to explore Statistical Regularity says that a moderately large number of the
their respective research areas. items chosen at random from the large group are almost sure on
Conclusive research can be divided into two categories: descriptive the average to possess the features of the large group.” According
and causal. Descriptive research design, as the name suggests, describes to this law the units of the sample must be selected at random.
specific elements, causes, or phenomena in the research area.
59
60 Research Methodology Sampling Techniques and Measurement 61
♦ ♦
• Law of Inertia of Large Numbers – According to this law, the Probability Sampling
other things being equal – the larger the size of the sample; the A probability sampling method is any method of sampling that
more accurate the results are likely to be. utilizes some form of random selection. In order to have a random
Characteristics of the sampling technique selection method, the researcher must set up some process or procedure
that assures that the different units in the researcher population have
1. Much cheaper equal probabilities of being chosen. Humans have long practiced
2. Saves time various forms of random selection, such as picking a name out of a
3. Much reliable hat, or choosing the short straw. These days, the researcher tend to use
computers as the mechanism for generating random numbers as the
4. Very suitable for carrying out different surveys
basis for random selection.
5. Scientific in nature
Some Definitions
Advantages of sampling
Before one can explain the various probability methods the
1. Very accurate researcher have to define some basic terms. These are:
2. Economical in nature • N = the number of cases in the sampling frame
3. Very reliable • n = the number of cases in the sample
4. High suitability ratio towards the different surveys • NCn = the number of combinations (subsets) of n from N
5. Takes less time • f = n/N = the sampling fraction
6. In cases, when the universe is very large, then the sampling method That’s it. With those terms defined the researcher can begin to
is the only practical method for collecting the data define the different probability sampling methods.
Disadvantages of sampling 2.2.1. Simple Random Sampling
1. Inadequacy of the samples. The simplest form of random sampling is called simple random
2. Chances for bias. sampling. Pretty tricky, huh? Here’s the quick description of simple
random sampling:
3. Problems of accuracy.
• Objective: To select n units out of N such that each NCn has an
4. Difficulty of getting the representative sample
equal chance of being selected.
5. Untrained manpower.
• Procedure: Use a table of random numbers, a computer random
6. Absence of the informants. number generator, or a mechanical device to select the sample.
7. Chances of committing the errors in sampling A somewhat stilted, if accurate, definition. Let’s see if the researcher
2.2. TYPES OF SAMPLING can make it a little more real. How do the researcher select a simple
random sample? Let’s assume that the researcher are doing some research
Sampling Methods can be classified into two categories:
with a small service agency that wishes to assess clients’ views of
• Probability Sampling: Sample is a known probability of being selected quality of service over the past year. First, the researcher have to get
• Non-probability Sampling: Sample does not have known probability the sampling frame organized. To accomplish this, we’ll go through
of being selected as in convenience or voluntary response surveys agency records to identify every client over the past 12 months. If
♦ ♦
we’re lucky, the agency has good accurate computerized records and sample back to the population. Simple random sampling is not the
can quickly produce such a list. Then, the researcher have to actually most statistically efficient method of sampling and the researcher may,
draw the sample. Decide on the number of clients the researcher would just because of the luck of the draw, not get good representation of
like to have in the final sample. For the sake of the example, let’s say subgroups in a population. To deal with these issues, the researcher
the researcher want to select 100 clients to survey and that there were have to turn to other sampling methods.
1000 clients over the past 12 months. Then, the sampling fraction is
f = n/N = 100/1000 =. 10 or 10%. Now, to actually draw the sample, 2.2.2. Stratified Random Sampling
the researcher have several options. The researcher could print off the Stratified Random Sampling, also sometimes called proportional
list of 1000 clients, tear then into separate strips, put the strips in a hat, or quota random sampling, involves dividing the researcher population
mix them up real good, close the researcher eyes and pull out the first into homogeneous subgroups and then taking a simple random sample
100. But this mechanical procedure would be tedious and the quality in each subgroup. In more formal terms:
of the sample would depend on how thoroughly the researcher mixed Objective: Divide the population into non-overlapping groups (i.e.,
them up and how randomly the researcher reached in. Perhaps a better strata) N1, N2, N3,. .. Ni, such that N1 + N2 + N3 +. .. + Ni = N. Then do
procedure would be to use the kind of ball machine that is popular a simple random sample of f = n/N in each strata.
with many of the state lotteries. The researcher would need three sets There are several major reasons why the researcher might prefer
of balls numbered 0 to 9, one set for each of the digits from 000 to stratified sampling over simple random sampling. First, it assures that the
999 (if the researcher select 000 we’ll call that 1000). Number the list researcher will be able to represent not only the overall population, but
of names from 1 to 1000 and then use the ball machine to select the also key subgroups of the population, especially small minority groups.
three digits that selects each person. The obvious disadvantage here is If the researchers want to be able to talk about subgroups, this may be
that the researcher need to get the ball machines. (Where do they make the only way to effectively assure the researcher will be able to. If the
those things, anyway? Is there a ball machine industry?). subgroup is extremely small, the researcher can use different sampling
Neither of these mechanical procedures is very feasible and, with fractions (f) within the different strata to randomly over-sample the
the development of inexpensive computers there is a much easier way. small group (although the researcher will then have to weight the within-
Here’s a simple procedure that’s especially useful if the researcher group estimates using the sampling fraction whenever the researcher
have the names of the clients already on the computer. Many computer wants overall population estimates). When researcher use the same
programs can generate a series of random numbers. Let’s assume the sampling fraction within strata researcher is conducting proportionate
researcher can copy and paste the list of client names into a column stratified random sampling. When one use different sampling fractions
in an EXCEL spreadsheet. Then, in the column right next to it paste in the strata, an investigator call this disproportionate stratified random
the function =RAND () which is EXCEL’s way of putting a random sampling. Second, stratified random sampling will generally have more
number between 0 and 1 in the cells. Then, sort both columns -- the statistical precision than simple random sampling. This will only be true
list of names and the random number -- by the random numbers. This if the strata or groups are homogeneous. If they are, researcher expects
rearranges the list in random order from the lowest to the highest that the variability within-groups are lower than the variability for the
random number. Then, all the researcher has to do is take the first population as a whole. Stratified sampling capitalizes on that fact.
hundred names in this sorted list. Pretty simple. The researcher could For example, let’s say that the population of clients for our agency
probably accomplish the whole thing in under a minute. can be divided into three groups: Caucasian, African-American and
Simple random sampling is simple to accomplish and is easy to Hispanic-American. Furthermore, let’s assume that both the African-
explain to others. Because simple random sampling is a fair way to Americans and Hispanic-Americans are relatively small minorities of
select a sample, it is reasonable to generalize the results from the the clientele (10% and 5% respectively). If researcher just did a simple
♦ ♦
random sample of n=100 with a sampling fraction of 10%, researcher Although sometimes more convenient, systematic sampling provides
would expect by chance alone that researcher would only get 10 and 5 less protection against introducing biases in the sample compared to
persons from each of our two smaller groups. And, by chance, researcher random sampling. Estimators for systematic sampling and simple
could get fewer than that! If researcher stratifies, researcher can do better. random sampling are identical; only the method of sample selected
First, let’s determine how many people researcher want to have in each differs. Therefore, systematic sampling is used to simplify the process
group. Let’s say researcher still want to take a sample of 100 from the of selecting a sample or to ensure ideal dispersion of sample units
population of 1000 clients over the past year. But researcher think that throughout the population.
in order to say anything about subgroups researcher will need at least
25 cases in each group. So, let’s sample 50 Caucasians, 25 African- Advantages
Americans, and 25 Hispanic-Americans. Researcher know that 10% • Easy to implement
of the population, or 100 clients, are African-American. If researcher
• Maximum dispersion of sample units throughout the population
randomly sample 25 of these, researcher have a within-stratum sampling
fraction of 25/100 = 25%. Similarly, researcher know that 5% or 50 • Requires minimum knowledge of the population
clients are Hispanic-American. So within-stratum sampling fraction will
Disadvantages
be 25/50 = 50%. Finally, by subtraction researcher know that there are
850 Caucasian clients. Our within-stratum sampling fraction for them • Less protection from possible biases
is 50/850 = about 5.88%. Because the groups are more homogeneous • Can be imprecise and inefficient relative to other designs if the
within-group than across the population as a whole, researcher can population being sampled is heterogeneous
expect greater statistical precision (less variance).
2.2.4. Cluster (Area) Random Sampling
2.2.3. Systematic Random Sampling
The problem with random sampling methods when the researcher
Here are the steps the researcher need to follow in order to achieve have to sample a population that’s disbursed across a wide geographic
a systematic random sample: region is that the researcher will have to cover a lot of ground
• number the units in the population from 1 to N geographically in order to get to each of the units the researcher sampled.
• decide on the n (sample size) that the researcher want or need Imagine taking a simple random sample of all the residents of New
• k = N/n = the interval size York State in order to conduct personal interviews. By the luck of the
draw the researcher will wind up with respondents who come from
• randomly select an integer between 1 to k all over the state. The researcher interviewers are going to have a lot
• then take every kth unit of traveling to do. It is for precisely this problem thatcluster or area
Occasionally, selecting sample units at random can introduce random sampling was invented.
logistical challenges that preclude collecting data efficiently. If the In cluster sampling, the researcher follow these steps:
chance of introducing a bias is low or if ideal dispersion of sample • divide population into clusters (usually along geographic boundaries)
units in the population is a higher priority that a strictly random sample, • randomly sample clusters
then it might be appropriate to choose samples non‐randomly. Like
simple random sampling, systematic sampling is a type of probability • measure all units within sampled clusters
sampling where each element in the population has a known and For instance, in the figure the researcher see a map of the counties in
equal probability of being selected. The probabilistic framework is New York State. Let’s say that the researcher have to do a survey of town
maintained through selection of one or more random starting points. governments that will require us going to the towns personally. If the
♦ ♦
researcher do simple random samples state-wide we’ll have to cover the stages in the sampling process and the researcher use both stratified and
entire state geographically. Instead, the researcher decide to do a cluster simple random sampling. By combining different sampling methods the
sampling of five counties (marked in red in the figure). Once these are researcher are able to achieve a rich variety of probabilistic sampling
selected, the researcher go to every town government in the five areas. methods that can be used in a wide range of social research contexts.
Clearly this strategy will help us to economize on our mileage. Cluster or
area sampling, then, is useful in situations like this, and is done primarily for 2.3. NON –PROBABILITY SAMPLING
efficiency of administration. Note also, that the researcher probably don’t The difference between nonprobability and probability sampling
have to worry about using this approach if the researcher are conducting a is that nonprobability sampling does not involve random selection and
mail or telephone survey because it doesn’t matter as much (or cost more probability sampling does. Does that mean that nonprobability samples
or raise inefficiency) where the researcher call or send letters to. are not representative of the population? Not necessarily. But it does
mean that nonprobability samples cannot depend upon the rationale of
2.2.5. Multi-Stage Sampling probability theory. At least with a probabilistic sample, the researcher
The four methods we’ve covered so far -- simple, stratified, know the odds or probability that the researcher have represented the
systematic and cluster -- are the simplest random sampling strategies. population well. the researcher are able to estimate confidence intervals
In most real applied social research, the researcher would use sampling for the statistic. With nonprobability samples, the researcher may or may
methods that are considerably more complex than these simple variations. not represent the population well, and it will often be hard for us to know
The most important principle here is that the researcher can combine how well we’ve done so. In general, researchers prefer probabilistic or
the simple methods described earlier in a variety of useful ways that random sampling methods over non-probabilistic ones, and consider
help us address our sampling needs in the most efficient and effective them to be more accurate and rigorous. However, in applied social
manner possible. When the researcher combine sampling methods, research there may be circumstances where it is not feasible, practical
the researcher call this multi-stage sampling. or theoretically sensible to do random sampling. Here, the researcher
For example, consider the idea of sampling New York State residents consider a wide range of non-probabilistic alternatives.
for face-to-face interviews. Clearly the researcher would want to do some We can divide nonprobability sampling methods are divided into
type of cluster sampling as the first stage of the process. the researcher two broad types: accidental or purposive. Most sampling methods are
might sample townships or census tracts throughout the state. But in purposive in nature because usually the sampling problem is approached
cluster sampling the researcher would then go on to measure everyone with a specific plan in mind. The most important distinctions among
in the clusters the researcher select. Even if the researcher are sampling these types of sampling methods are the ones between the different
census tracts the researcher may not be able to measure everyone who is types of purposive sampling approaches.
in the census tract. So, the researcher might set up a stratified sampling
process within the clusters. In this case, the researcher would have 2.3.1. Accidental, Haphazard or Convenience Sampling
a two-stage sampling process with stratified samples within cluster One of the most common methods of sampling goes under the various
samples. Or, consider the problem of sampling students in grade schools. titles listed here. One would include in this category the traditional “man
the researcher might begin with a national sample of school districts on the street” (of course, now it’s probably the “person on the street”)
stratified by economics and educational level. Within selected districts, interviews conducted frequently by television news programs to get a
the researcher might do a simple random sample of schools. Within quick (although non-representative) reading of public opinion.onewould
schools, the researcher might do a simple random sample of classes also argue that the typical use of college students in much psychological
or grades. And, within classes, the researcher might even do a simple research is primarily a matter of convenience. (The researcher doesn’t
random sample of students. In this case, the researcher have three or four really believe that psychologists use college students because they believe
♦ ♦
they’re representative of the population at large, do the researcher?). In a lot of informal public opinion polls, for instance, they interview
In clinical practice, the researcher might use clients who are available a “typical” voter. There are a number of problems with this sampling
to us as our sample. In many research contexts, the researcher sample approach. First, how do the researcher know what the “typical” or “modal”
simply by asking for volunteers. Clearly, the problem with all of these case is? the researcher could say that the modal voter is a person who
types of samples is that the researcher have no evidence that they are is of average age, educational level, and income in the population. But,
representative of the populations we’re interested in generalizing to -- it’s not clear that using the averages of these is the fairest (consider the
and in many cases the researcher would clearly suspect that they are not. skewed distribution of income, for instance). And, how do the researcher
know that those three variables -- age, education, income -- are the
2.3.2. Purposive Sampling
only or even the most relevant for classifying the typical voter? What
In purposive sampling, the researcher select the sample with a purpose if religion or ethnicity is an important discriminator? Clearly, modal
in mind. the researcher usually would have one or more specific predefined instance sampling is only sensible for informal sampling contexts.
groups the researcher are seeking. For instance, have the researcher ever
run into people in a mall or on the street who are carrying a clipboard Expert Sampling
and who are stopping various people and asking if they could interview Expert sampling involves the assembling of a sample of persons
them? Most likely they are conducting a purposive sample (and most with known or demonstrable experience and expertise in some area.
likely they are engaged in market research). They might be looking for Often, the researcher convene such a sample under the auspices of a
Caucasian females between 30-40 years old. They size up the people “panel of experts.” There are actually two reasons the researcher might
passing by and anyone who looks to be in that category they stop to do expert sampling. First, because it would be the best way to elicit
ask if they will participate. One of the first things they’re likely to do is the views of persons who have specific expertise. In this case, expert
verify that the respondent does in fact meet the criteria for being in the sampling is essentially just a specific subcase of purposive sampling.
sample. Purposive sampling can be very useful for situations where the But the other reason the researcher might use expert sampling is to
researcher need to reach a targeted sample quickly and where sampling provide evidence for the validity of another sampling approach the
for proportionality is not the primary concern. With a purposive sample, researcher have chosen. For instance, let’s say the researcher do modal
the researcher are likely to get the opinions of the researcher target instance sampling and are concerned that the criteria the researcher used
population, but the researcher are also likely to overweight subgroups for defining the modal instance are subject to criticism. The researcher
in the researcher population that are more readily accessible. might convene an expert panel consisting of persons with acknowledged
All of the methods that follow can be considered subcategories of experience and insight into that field or topic and ask them to examine
purposive sampling methods. the researcher might sample for specific the researcher modal definitions and comment on their appropriateness
groups or types of people as in modal instance, expert, or quota sampling. and validity. The advantage of doing this is that the researcher aren’t out
the researcher might sample for diversity as in heterogeneity sampling. on the researcher own trying to defend the researcher decisions -- the
Or, the researcher might capitalize on informal social networks to researcher have some acknowledged experts to back the researcher.
identify specific respondents who are hard to locate otherwise, as in The disadvantage is that even the experts can be, and often are, wrong.
snowball sampling. In all of these methods the researcher know what
Quota Sampling
the researcher want -- the researcher are sampling with a purpose.
In quota sampling, the researcher select people non-randomly
Modal Instance Sampling according to some fixed quota. There are two types of quota sampling:
In statistics, the mode is the most frequently occurring value in a proportional and non proportional. In proportional quota sampling
distribution. In sampling, when the researcher do a modal instance sample, the researchers want to represent the major characteristics of the
the researcher are sampling the most frequent case, or the “typical” case. population by sampling a proportional amount of each. For instance,
♦ ♦
if the researcher know the population has 40% women and 60% Snowball Sampling
men, and that the researcher want a total sample size of 100, the In snowball sampling, the researcher begins by identifying someone
researcher will continue sampling until the researcher get those who meets the criteria for inclusion in the researcher study. The researchers
percentages and then the researcher will stop. So, if the researcher then ask them to recommend others who they may know who also meet
have already got the 40 women for the researcher sample, but not the criteria. Although this method would hardly lead to representative
the sixty men, the researcher will continue to sample men but even samples, there are times when it may be the best method available.
if legitimate women respondents come along, the researcher will not Snowball sampling is especially useful when the researcher are trying to
sample them because the researcher have already “met the researcher reach populations that are inaccessible or hard to find. For instance, if the
quota.” The problem here (as in much purposive sampling) is that the researcher are studying the homeless, the researcher are not likely to be
researcher have to decide the specific characteristics on which the able to find good lists of homeless people within a specific geographical
researcher will base the quota. Will it be by gender, age, education area. However, if the researcher go to that area and identify one or two,
race, religion, etc.? the researcher may find that they know very well who the other homeless
Non-proportional quota sampling is a bit less restrictive. In this people in their vicinity are and how the researcher can find them.
method, the researcher specifies the minimum number of sampled
units the researcher want in each category. Here, the researcher is 2.4. THEORIES OF PROBABILITY
not concerned with having numbers that match the proportions in the Probability theory is the branch of mathematics concerned
population. Instead, the researchers simply want to have enough to with probability. Although there are several different probability
assure that the researcher will be able to talk about even small groups interpretations, probability theory treats the concept in a rigorous
in the population. This method is the non-probabilistic analogue of mathematical manner by expressing it through a set of axioms. Typically
stratified random sampling in that it is typically used to assure that these axioms formalize probability in terms of a probability space,
smaller groups are adequately represented in the researcher sample. which assigns a measure taking values between 0 and 1, termed the
probability measure, to a set of outcomes called the sample space. Any
Heterogeneity Sampling
specified subset of these outcomes is called an event.
Sample for heterogeneity is when the researcher want to include all Central subjects in probability theory include discrete and continuous
opinions or views, and the researcher aren’t concerned about representing random variables, probability distributions, and stochastic processes,
these views proportionately. Another term for this is sampling for which provide mathematical abstractions of non-deterministic or
diversity. In many brainstorming or nominal group processes (including uncertain processes or measured quantities that may either be single
concept mapping), the researcher would use some form of heterogeneity occurrences or evolve over time in a random fashion.
sampling because our primary interest is in getting broad spectrum of Although it is not possible to perfectly predict random events, much
ideas, not identifying the “average” or “modal instance” ones. In effect, can be said about their behaviour. Two major results in probability
what the researcher would like to be sampling is not people, but ideas. the theory describing such behaviour are the law of large numbers and
researcher imagine that there is a universe of all possible ideas relevant the central limit theorem.
to some topic and that the researcher want to sample this population, As a mathematical foundation for statistics, probability theory is
not the population of people who have the ideas. Clearly, in order to essential to many human activities that involve quantitative analysis of data.
get all of the ideas, and especially the “outlier” or unusual ones, the Methods of probability theory also apply to descriptions of complex systems
researcher have to include a broad and diverse range of participants. given only partial knowledge of their state, as in statistical mechanics. A
Heterogeneity sampling is, in this sense, almost the opposite of modal great discovery of twentieth-century physics was the probabilistic nature
instance sampling. of physical phenomena at atomic scales, described in quantum mechanics.
♦ ♦
Probability theory, a branch of mathematics concerned with the can be 95% certain; the 99% confidence level means the researcher
analysis of random phenomena. The outcome of a random event cannot can be 99% certain. Most researchers use the 95% confidence level.
be determined before it occurs, but it may be any one of several possible When the researcher put the confidence level and the confidence
outcomes. The actual outcome is considered to be determined by chance. interval together, the researcher can say that the researcher are 95% sure
The word probability has several meanings in ordinary conversation. that the true percentage of the population is between 43% and 51%.The
Two of these are particularly important for the development and wider the confidence interval the researcher are willing to accept, the
applications of the mathematical theory of probability. One is the more certain the researcher can be that the whole population answers
interpretation of probabilities as relative frequencies, for which would be within that range. For example, if the researcher asked a
simple games involving coins, cards, dice, and roulette wheels sample of 1000 people in a city which brand of cola they preferred,
provide examples. The distinctive feature of games of chance is and 60% said Brand A, the researcher can be very certain that between
that the outcome of a given trial cannot be predicted with certainty, 40 and 80% of all the people in the city actually do prefer that brand,
although the collective results of a large number of trials display some but the researcher cannot be so sure that between 59 and 61% of the
regularity. For example, the statement that the probability of “heads” people in the city prefer the brand.
in tossing a coin equals one-half, according to the relative frequency
Factors that Affect Confidence Intervals
interpretation, implies that in a large number of tosses the relative
frequency with which “heads” actually occurs will be approximately There are three factors that determine the size of the confidence
one-half, although it contains no implication concerning the outcome interval for a given confidence level. These are: sample size, percentage
of any given toss. There are many similar examples involving groups and population size.
of people, molecules of a gas, genes, and so on. Actuarial statements Sample Size
about the life expectancy for persons of a certain age describe the The larger the researcher sample, the more sure the researcher can
collective experience of a large number of individuals but do not be that their answers truly reflect the population. This indicates that for a
purport to say what will happen to any particular person. Similarly, given confidence level, the larger the researcher sample size, the smaller
predictions about the chance of a genetic disease occurring in a child the researcher confidence interval. However, the relationship is not linear
of parents having a known genetic makeup are statements about (i.e., doubling the sample size does not halve the confidence interval).
relative frequencies of occurrence in a large number of cases but are
not predictions about a given individual. Percentage
The confidence interval is the plus-or-minus figure usually reported Accuracy also depends on the percentage of the sample that picks a
in newspaper or television opinion poll results. For example, if the particular answer. If 99% of the researcher sample said “Yes” and 1% said
researcher use a confidence interval of 4 and 47% percent of the “No” the chances of error are remote, irrespective of sample size. However,
researcher sample picks an answer the researcher can be “sure” that if the percentages are 51% and 49% the chances of error are much greater.
if the researcher had asked the question of the entire relevant population It is easier to be sure of extreme answers than of middle-of-the-road ones.
between 43% (47-4) and 51% (47+4) would have picked that answer. When determining the sample size needed for a given level of
accuracy the researcher must use the worst case percentage (50%).
2.5. LEVELS OF CONFIDENCE The researcher should also use this percentage if the researchers want
The confidence level tells the researcher how sure the researcher to determine a general level of accuracy for a sample the researcher
can be. It is expressed as a percentage and represents how often the true already has. To determine the confidence interval for a specific answer
percentage of the population who would pick an answer lies within the the researcher sample has given, the researcher can use the percentage
confidence interval. The 95% confidence level means the researcher picking that answer and get a smaller interval.
♦ ♦
How many people is there in the group the researcher sample 2.6. DEGREES OF FREEDOM
represents? This may be the number of people in a city the researcher are In statistics, degree of freedom (DF) is a difficult concept to explain.
studying, the number of people who buy new cars, etc. Often the researcher However, it is an important idea that appears in many different contexts
may not know the exact population size. This is not a problem. The throughout statistics including hypothesis tests, probability distributions,
mathematics of probability proves the size of the population is irrelevant, and regression analysis.
unless the size of the sample exceeds a few percent of the total population
the researcher are examining. This means that a sample of 500 people
Definition of Degrees of Freedom
is equally useful in examining the opinions of a state of 15,000,000 as it
would a city of 100,000. For this reason, the sample calculator ignores Degrees of freedom are the number of independent values that a
the population size when it is “large” or unknown. Population size is only statistical analysis can estimate. The researcher can also think of it as
likely to be a factor when the researcher work with a relatively small and the number of values that are free to vary as the researcher estimate
known group of people. parameters.
Degrees of freedom encompass the notion that the amount of
independent information the researcher have limits the number of
parameters that the researcher can estimate. Typically, the degrees
of freedom equal the researcher sample size minus the number of
parameters the researcher need to calculate during an analysis. It is
usually a positive whole number.
Degree of freedom is a combination of how much data the researcher
have and how many parameters the researcher need to estimate. It
indicates how much independent information goes into a parameter
estimate. In this vein, it’s easy to see that the researcher want a lot
of information to go into parameter estimates to obtain more precise
estimates and more powerful hypothesis tests. So, the researcher want
many degrees of freedom!
Independent Information and Restrictions on Values

The definitions talk about independent information. The researcher
might think this refers to the sample size, but it’s a little more
complicated than that. To understand why, the researcher need talk
about the freedom to vary. The best way to illustrate this concept is
with an example.
Suppose the researcher collect the random sample of observations
shown below. Now, imagine that the researcher know the mean, but
the researcher don’t know the value of an observation—the X in the
table below.
71
♦ ♦
to determine statistical significance. Each of these distributions is a
family of distributions where the degrees of freedom define the shape.
Hypothesis tests use these distributions to calculate p-values. So, the
DF are directly linked to p-values through these distributions
Related post: A Graphical Look at Significance Levels (Alpha)
and P values
Degrees of Freedom for t-Tests and the t-Distribution
t-tests are hypothesis tests for the mean and use the t-distribution
to determine statistical significance.
A 1-sample t-test determines whether the difference between the
sample mean and the null hypothesis value is statistically significant.
Let’s go back to our example of the mean above. the researcher know
that when the researcher have a sample and estimate the mean, the
researcher have n – 1 degrees of freedom, where n is the sample size.
Consequently, for a 1-sample t-test, the degrees of freedom is n – 1.
The mean is 6.9, and it is based on 10 values. So, the researcher
The mean is 6.9, and it is based on 10 values. So, the researcher know that the values must sum to The DF define the shape of the t-distribution that the researcher t-test
know that the values must sum to 69 based on the equation for the mean.
69 based on the equation for the mean. uses to calculate the p-value. The graph below shows the t-distribution
Using simple algebra (64 + X = 69), the researcher know that X
for several different degrees of freedom. Because the degrees of freedom
must equal 5.
Using simple algebra (64 + X = 69), the researcher know that X must equal 5. are so closely related to sample size, the researcher can see the effect of
sample size. As the degrees of freedom decreases, the t-distribution has
Estimating Parameters Imposes Constraints on the Data
thicker tails. This property allows for the greater uncertainty associated
Estimating Parameters
As theImposes Constraints
researcher can see,onthat
thelast
Datanumber has no freedom to vary.
with small sample sizes.
It is not an independent piece of information because it cannot be any
As theother value.
researcher canEstimating
see, that lastthe parameter,
number the mean
has no freedom in this
to vary. It iscase,
not animposes
independent piece of
a constraint
information because onbethe
it cannot anyfreedom to Estimating
other value. vary. The the
lastparameter,
value andthethe mean
mean in thisare
case, imposes
entirely dependent on each other. Consequently, after estimating the
a constraint on the freedom to vary. The last value and the mean are entirely dependent on each other.
mean, the researcher have only 9 independent pieces of information
Consequently, after estimating the mean, the researcher have only 9 independent pieces of information
even though our sample size is 10.
even though our sampleThat’ssizethe
is 10.
basic idea for degrees of freedom in statistics. In a
general sense, DF are the number of observations in a sample that are
That‘sfree to vary
the basic ideawhile estimating
for degrees statistical
of freedom parameters.
in statistics. The researcher
In a general sense, DF arecan
the number of
also think of it as the amount of independent data that the researcher
observations in a sample that are free to vary while estimating statistical parameters. The researcher can
also think of itcan useamount
as the to estimate a parameter.
of independent data that the researcher can use to estimate a parameter.
Degrees of Freedom and Probability Distributions

Degrees of freedom also define the distributions for testing
statistics of various hypothesis tests. For example, hypothesis tests
use the t-distribution, F-distribution,
73
and the chi-square distribution
To dig into t-tests, read my post about How t-Tests Work.one canshow how the different t-tests
calculate t-values and use t-distributions to calculate p-values.
♦ ♦
To dig into t-tests, read my post about How t-Tests Work.one Chi-Square 3 X 2 Table
canshow how the different t-tests calculate t-values and use t-distributions Now, let’s try a 3 X 2 table. The table below illustrates the example
to calculate p-values. that one use in my post about the chi-square test of independence. In that
The F-test in ANOVA also tests group means. It uses the post, one determine whether there is a statistically significant relationship
F-distribution, which is defined by the degrees of freedom. However, between uniform color and deaths on the original Star Trek TV series.
the researchers calculate the DF for an F-distribution differently. For
more information, read my post about how F-tests Work in ANOVA.
Degrees of Freedom for the Chi-Square Test of Independence
The chi-square test of independence determines whether there is a
statistically significant relationship between categorical variables. Just
like other hypothesis tests, this test incorporates degrees of freedom.
For a table with r rows and c columns, the general rule for calculating
degrees of freedom for a chi-square test is (r-1) (c-1).
However, the researcher can create tables to understand it more
intuitively. The degrees of freedom for a chi-square test of independence
are the number of cells in the table that can vary before the researcher
can calculate all the other cells. In a chi-square table, the cells represent In table,
In the the table, one categorical
one categorical variable
variable is shirt is shirt
color, which can becolor, which
blue, gold, canThebeother
or red.
the observed frequency for each combination of categorical variables. blue,variable
categorical gold,isorstatus,
red.which
Thecan
other categorical
be dead or alive. Aftervariable
one enteredis
thestatus, which
two bolded values,can
one can
The
Chi-Square 2 Xconstraints
2 Table are the totals in the margins. be dead
calculate or alive.
all the remaining After
cells. one entered
Consequently, this table the
has 2 two
DF. bolded values, one can
Chi-Square 2 X 2 Table calculate all the remaining cells. Consequently, this table has 2 DF.
For example, in a 2 X 2 table, after the researcher enter one value in the table, the researcher can calculate Like the t-distribution, the chi-square distribution is a family of distributions where the degrees of
For example, in a 2 X 2 table, after the researcher enter one value Like the t-distribution, the chi-square distribution is a family of
freedom define the shape. Chi-square tests use this distribution to calculate p-values. The graph below
the remaining cells.
in the table, the researcher can calculate the remaining cells. distributions where the degrees of freedom define the shape. Chi-
displays several chi-square distributions.
square tests use this distribution to calculate p-values. The graph below
displays several chi-square distributions.
In the table above, one entered the bold 15, and then one can calculate
thetable
In the remaining three
above, one values
entered in parentheses.
the bold Therefore,
15, and then one thisthe
can calculate table has 1 DF.
remaining three values in 77
parentheses. Therefore, this table has 1 DF.
Degrees of Freedom in Regression Analysis
Chi-Square 3 X 2 Table Degrees of freedom in regression are a bit more complicated, and it can keep it on the simple side.
♦ ♦
egrees of Freedom in Regression Analysis
Degrees of Freedom in Regression Analysis enough information for what the researcher want to do, the researcher
Degrees of freedom in regression are a bit more complicated, and will have imprecise estimates and low statistical power.
Degreesitofcan
freedom
keep itinon
regression are side.
the simple a bit In
more complicated,
a regression and iteach
model, can keep
termitison
anthe simple side. 2.7. HYPOTHESIS
a regressionestimated parameter
model, each term isthat
an uses one degree
estimated of freedom.
parameter that usesInone
thedegree
regression
of freedom. In the
A hypothesis (plural hypotheses) is a proposed explanation for
output
gression output below,
below, the researcher
the researcher can see can
howsee
eachhow
termeach terma requires
requires DF. Therea are
DF.28 observations
a phenomenon. For a hypothesis to be a scientific hypothesis, the
There are 28 observations and the two independent variables use a
d the two independent variables use a total of two degrees of freedom. The remaining 26 degrees of scientific method requires that one can test it. Scientists generally base
total of two degrees of freedom. The remaining 26 degrees of freedom
eedom are displayed in Error.in Error. scientific hypotheses on previous observations that cannot satisfactorily
are displayed
be explained with the available scientific theories. Even though the words
“hypothesis” and “theory” are often used synonymously, a scientific
hypothesis is not the same as a scientific theory. A working hypothesis
is a provisionally accepted hypothesis proposed for further research.
A different meaning of the term hypothesis is used in formal logic, to
denote the antecedent of a proposition; thus in the proposition “If P, then
Q”, P denotes the hypothesis (or antecedent); Q can be called a consequent.
P is the assumption in a (possibly counterfactual) What If question.
The adjective hypothetical, meaning “having the nature of a
hypothesis”, or “being assumed to exist as an immediate consequence of a
hypothesis”, can refer to any of these meanings of the term “hypothesis”.
The error
The error degrees degreesareofthefreedom
of freedom are the
independent independent
pieces pieces
of information that of
are available for Meaning and Importance of Hypothesis
information that are available for estimating the researcher coefficient
timating the researcher coefficient estimates. For precisecoefficient estimates and powerful hypothesis “Hypo” means below and “thesis” means idea or general opinion
estimates. For precisecoefficient estimates and powerful hypothesis
tests in regression, the researcher78must have many error degrees of to be defendedby a person and thus hypothesis means an idea form
freedom. This equates to having many observations for each model term. beforehand which has less value than thegenerally formed view. If
As the researcher add terms to the model, the error degrees of the researcher have to proceed towards some destination for which
freedom decreases. The researcher has fewer pieces of information the researcher do not knowthe way, the researcher try to form an idea
available to estimate the parameters. This situation reduces the precision about the direction to proceed and by trial and error, the researcher
of the parameters and the power of the tests. When the researcher has reach that goal. The primary idea is called a hypothesis.
too few remaining degrees of freedom, the researcher can’t trust the Characteristics of Good Hypothesis
regression results. If the researchers use all the researcher degrees of
Specific: The hypothesis should not be too vague or general.
freedom, the p-values can’t be calculated.
Conceptually clear: The hypothesis should be properly expressed.
For more information about the problems that occur when the researcher
It leads to discovery of additions to knowledge by helping to confirm
use too many degrees of freedom and how many observations the researcher
or disconfirm particular theories or propositions.
need, read my blog post about over fitting the researcher model.
Even though they might seem murky, degrees of freedom are Related to be available technique
essential to any statistical analysis! In a nutshell, DF define the amount The hypothesis should be capable of being verified. Itmust be
of information the researcher have relative to the number of properties simple and capable to empirical test. It provides the framework for
that the researcher want to estimate. If the researcher doesn’t have drawing theconclusion of a research endeavor. In fact, conclusions are
♦ ♦
direct response to the hypothesisformulated for the study as confirmed relation to theories in the field. Hypothesis testing isoften referred to as
or discontinued by data analysis. significance testing. A test of significance is conducted by comparing
thevalues of a statistics computed from a sample with values predicted
Types of Hypothesis by the samplingdistribution under the assumption that the null hypothesis
Hypotheses are mostly commonly classified according to the is true.
way they are stated in the research process. Based on this method of
classification, there are two types of hypothesis: Types of Errors
1. Null hypothesis: The null hypothesis may be crude or refined. There are few errors in the nature of problems and hypothesis
When a null hypothesis stated negatively, it is called a null. A 1. Scientific problem is not moral and ethical questions.
crude null hypothesis which is at the low levelof abstraction and 2. Value statements that indicate cultural or personal judgments or
it does not lead to higher theoretical research. The null hypothesis preferences should be avoided.
is aproposition that stipulates that there would be no relationship 3. Another common defect of problem statements often occurs in
or difference between thevariables being studied and that any such doctoral thesis: listing the methodological points or problems as
relationship or difference if found to exist does soaccidentally or as sub problems.
a result of chance it is more or less the negation of the directional
or alternative hypothesis. 2.8. ONE TAILED TEST AND TWO TAILED TESTS
2. Uniform: Uniform Hypothesis relate to the existence of empirical Conduct a test of statistical significance, whether it is from a
uniformities. correlation, an ANOVA, a regression or some other kind of test, the
Complex:The complex types are concerned with complex idea researcher are given a p-value somewhere in the output. If the researcher
type. They existence of logically desired relationship between empirical test statistic is symmetrically distributed, the researcher can select one
uniformities. of three alternative hypotheses. Two of these correspond to one-tailed
tests and one corresponds to a two-tailed test. However, the p-value
Analytic presented is (almost always) for a two-tailed test.
They are aimed at finding out the relationship between changes
in one property leading to change in another. And so many types are 2.9. TWO TAILED TEST
there which are relating to hypothesis. Using a significance level of 0.05, a two-tailed test allots half of the
i.e. Uni
variablbe and multi variable, universal and statistical, researcher’s alpha to testing the statistical significance in one direction
temporal and cross sectional, descriptive, relational, explanatory and half of the researcher alpha to testing statistical significance in the
etc. The problem and hypothesis direct investigation helps to deduce other direction. This means that. 025 is in each tail of the distribution of
specific empirical manifestation implied by them.It advances scientific the researcher test statistic. When using a two-tailed test, regardless of
knowledge by helping the investigator to confirm or reject theory and the direction of the relationship the researcher hypothesize, the researcher
hypothesis incorporates the theory or part of it, in testable or near are testing for the possibility of the relationship in both directions. For
testable form. example, the researcher may wish to compare the mean of a sample to
a given value x using a t-test. Our null hypothesis is that the mean is
Hypothesis Testing and Levels of Significance equal to x. A two-tailed test will test both if the mean is significantly
In qualitative studies, the main purpose of the analysis is to test greater than x and if the mean significantly less than x. The mean is
hypothesis, which form thebasis of the study and to discuss these in considered significantly different from x if the test statistic is in the
probability distribution or bottom 5% of its probability distribution, resulting in a p-value less th
en using a two-tailed test, regardless of the direction of the relationship the researcher hypothesize, the
0.05. The one-tailed test provides more power to detect an effect in one direction by not testing the eff
earcher are testing for the possibility of the relationship in both directions. For example, the researcher
in the other direction.
y wish to compare the mean of a sample to a given value x using a t-test. Our null hypothesis is that the
x. AResearch
an is equal to84 Methodology
two-tailed test will test both if the mean is significantly greater than x and if the mean Sampling Techniques and Measurement 85
♦ ♦
nificantly lesstop
than x. The
2.5% mean is 2.5%
or bottom considered
of its significantly x if the test
different fromresulting
probability distribution, in astatistic is in the
p-value
2.5% or bottom 2.5%less
of itsthan 0.05. distribution, resulting in a p-value less than 0.05.
probability
2.11. FORMULATION OF HYPOTHESIS

2.11. FORMULATION OF HYPOTHESIS
Once the researchers have identified the researcher research
82 Once the researchers have identified the researcher research question, it is time to formulate t
question, it is time to formulate the researcher hypothesis. While the
researcher hypothesis.
researchWhile the research
question is broadquestion is broad
and includes and variables
all the includes allthethe variables the research
researcher
2.10. ONE TAILED TEST
wantstudy
want the researcher the researcher
to consider,study to consider,
the hypothesis is athe hypothesis
statement is a statement
that specific that the research
relationship
Using a significance level of. 05, a one-tailed test allots the
specific relationship the researcher expect to find from the researcher
83
researchers’ entire alpha to testing the statistical significance in the
examination of these variables. When formulating the hypothesis(es)
one direction of interest. This means that. 05 is in one tail of the
for the researcher study, there are a few things the researcher need to
distribution of the researcher test statistic. When using a one-tailed test,
keep in mind. Good hypotheses meet the following criteria:
the researcher are testing for the possibility of the relationship in one
direction and completely disregarding the possibility of a relationship 1. Identify the independent and dependent variables to be studied.
in the other direction. Let’s return to our example comparing the mean 2. Specify the nature of the relationship that exists between these variables.
of a sample to a given value x using a t-test. Our null hypothesis is 3. Simple (often referred to as parsimonious). It is better to be concise
that the mean is equal to x. A one-tailed test will test either if the than to be long-winded. It is also better to have several simple
mean is significantly greater than x or if the mean is significantly less hypotheses than one complicated hypothesis.
than x, but not both. Then, depending on the chosen tail, the mean
4. Does not include reference to specific measures.
is significantly greater than or less than x if the test statistic is in the
top 5% of its probability distribution or bottom 5% of its probability 5. Does not refer to specific statistical procedures that will be used
distribution, resulting in a p-value less than 0.05. The one-tailed test in analysis.
provides more power to detect an effect in one direction by not testing 6. Implies the population that the researcher are going to study.
the effect in the other direction. 7. Is falsifiable and testable.
♦ ♦
As indicated above, it is better to have several simple hypotheses 2.12. TESTING OF HYPOTHESIS
than one complex one. However, it is also a good idea to limit the number Hypothesis test is a statistical test that is used to determine whether
of hypotheses the researcher use in a study to six or fewer. Studies that there is enough evidence in a sample of data to infer that a certain
address more hypotheses than six will often be too time consuming to condition is true for the entire population.
keep participants interested, and uninterested participants do not take A hypothesis test examines two opposing hypotheses about
the importance of their responses as seriously. Another advantage to a population: the null hypothesis and the alternative hypothesis.
limiting the number of formal hypotheses the researcher formulate is The null hypothesis is the statement being tested. Usually the null
that too many can make the discussion section of the researcher paper hypothesis is a statement of “no effect” or “no difference”. The
very hard to write. alternative hypothesis is the statement the researcher want to be
It is important to remember that the researcher do not have to have able to conclude is true.
a formal hypothesis to justify all comparisons and statistical procedures Based on the sample data, the test determines whether to reject the
the researcher might use. For instance, it is only when the researcher null hypothesis. The researcher use a p-value, to make the determination.
start doing exploratory analysis of the researcher data that the researcher If the p-value is less than or equal to the level of significance, which
realize that gender is an influencing factor. The researcher do not have is a cut-off point that the researcher define, then the researcher can
to back up and write a hypothesis that addresses this finding. In fact, reject the null hypothesis.
it is better in most cases to not do this. The researcher can report any A common misconception is that statistical hypothesis tests are
statistical findings the researcher feel are relevant, whether or not the designed to select the more likely of two hypotheses. Instead, a test
researcher have a hypothesis that addressed them. will remain with the null hypothesis until there is enough evidence
The final criterion listed above warrants additional mention. A good (data) to support the alternative hypothesis.
hypothesis is not only testable, that is, something the researcher can Examples of questions the researcher can answer with a hypothesis
actually test for in the researcher study, but is must also be falsifiable. It test include:
is tempting to ignore this requirement, especially as a new researcher.
• Does the mean height of undergraduate women differ from 66
the researcher want so badly to find great things, and for our study to
inches?
turn out exactly as the researcher expect it to, that the researcher tend to
ignore the possibility that the researcher don’t know everything and that • Is the standard deviation of their height equal less than 5 inches?
no prediction is failsafe when it comes to humans. Try to keep in mind • Do male and female undergraduates differ in height?
that all research is relevant. Whether or not the researcher findings are
2.13. MEASUREMENT SCALES
what the researcher expect, the researcher will find something. Believe
it or not, failing to find group differences can be just as important as Measurement scales are used to categorize and/or quantify variables.
finding expected group differences. In fact, studies that return results in The four scales of measurement that is commonly used in statistical
opposition to what the researcher were hoping for, or believed would analysis: nominal, ordinal, interval, and ratio scales.
logically occur, often lead to many more great studies than the researcher Properties of Measurement Scales
could have hoped for. After all, it could be great for the findings of Each scale of measurement satisfies one or more of the following
the researcher current research to act as a guiding principal to the properties of measurement.
researcher future research… it is likely that this would require less • Identity- Each value on the measurement scale has a unique meaning.
work in terms of literature review, as the researcher would always be • Magnitude- Values on the measurement scale have an ordered
familiar with at least a portion of the literature that is relevant to the relationship to one another. That is, some values are larger and
researcher latest study! some are smaller.
♦ ♦
• Equal intervals- Scale units along the scale are equal to one another. that a #4 is better than a #3 or #2, but the researcher don’t know–and
This means, for example, that the difference between 1 and 2 would cannot quantify–how much better it is. For example, is the difference
be equal to the difference between 19 and 20. between “OK” and “Unhappy” the same as the difference between
• A minimum value of zero. The scale has a true zero point, below “Very Happy” and “Happy?” the researcher can’t say.
which no values exist. Ordinal scales are typically measures of non-numeric concepts
like satisfaction, happiness, discomfort, etc.
2.14. NOMINAL SCALE OF MEASUREMENT “Ordinal” is easy to remember because it sounds like “order” and
The nominal scale of measurement only satisfies the identity that’s the key to remember with “ordinal scales”–it is the order that
property of measurement. Values assigned to variables represent a matters, but that’s all the researcher really get from these.
descriptive category, but have no inherent numerical value with respect Advanced note: The best way to determine central tendency on a
to magnitude. set of ordinal data is to use the mode or median; the mean cannot be
Gender is an example of a variable that is measured on a nominal defined from an ordinal set.
scale. Individuals may be classified as “male” or “female”, but neither
value represents more or less “gender” than the other. Religion and political
affiliation are other examples of variables that are normally measured on a
nominal scale. Nominal scales are used for labeling variables, without any
quantitative value. “Nominal” scales could simply be called “labels.” Here
are some examples, below. Notices that all of these scales are mutually
exclusive (no overlap) and none of them have any numerical significance.
A good way to remember all of this is that “nominal” sounds a lot like
“name” and nominal scales are kind of like “names” or labels.
Example
Example of Ordinal Scales of Ordinal Scales
The ordinal scale has the property of both identity and magnitude.
The ordinal scale has
Each value on the
theproperty
ordinalofscale
both identity and magnitude.
has a unique meaning, Eachand
valueit on
hastheanordinal scale
has a uniqueordered
meaning,relationship to every
and it has an ordered other value
relationship on other
to every the scale.
value on the scale.
An example of an ordinal scale in action would be the results of a
horse race,
An example of anreported as “win”,
ordinal scale “place”,
in action would and “show”.
be the results the
of aresearcher
horse race, know
reported as "win",
"place", andthe rankthe
"show". in researcher
which horses
knowfinished the race.
the rank order The horses
in which horse finished
that wonthefinished
race. The horse that
Examples of Nominal Scales won finished ahead of the horse that placed, and the horse that placed finished aheadahead
ahead of the horse that placed, and the horse that placed finished of the horse that
Examples of Nominal Scales of the horse that showed. However, the researcher cannot tell from
showed. However, the researcher cannot tell from this ordinal scale whether it was a close race or whether
Note: Note:
a sub-typeaofsub-type ofwith
nominal scale nominal scale with
only two categories only twois called
(e.g. male/female) categories (e.g.If
―dichotomous.‖ this ordinal scale whether it was a close race or whether the winning
male/female) is called
the researcher are a student, “dichotomous.”
the researcher If thetheresearcher
can use that to impress are a student,
researcher teacher. the winning horse won by a mile.
horse won by a mile.
the researcher can use that to impress the researcher teacher.
2.15. ORDINAL SCALE OF MEASUREMENT
2.16. SCALE
2.16. INTERVAL INTERVAL SCALE OF MEASUREMENT
OF MEASUREMENT
2.15.WithORDINAL SCALE OF MEASUREMENT
ordinal scales, it is the order of the values is what‘s important and significant, but the Interval scales are numeric scales in which the researcher know not
Withbetween
ordinal
each scales,
one is notit is the order
Takeof the atvalues is what’s
below. Inimportant
Interval scales are numeric scales in which the researcher know not only the order, but also the
differences really known. a look the example each case, the only the order, but also the exact differences between the values. The
and significant,
researcher know that abut
#4 is the
betterdifferences
than a #3 or #2,between eachdon‘t
but the researcher oneknow–and
is not really known.
cannot quantify–how exact differences
classicbetween
examplethe of
values. The classic
an interval scale example of temperature
is Celsius an interval scale is Celsius
because the temperature
Take a look
much better at the
it is. For example
example, below.
is the difference In each
between case,
―OK‖ and the researcher
―Unhappy‖ know
the same as the difference because the difference between each value is the same. For example, the difference 60 and 50
difference between each value is the same. For example, the difference between
between ―Very Happy‖ and ―Happy?‖ the researcher can‘t say.
degrees is a measurable 10 degrees, as is the difference between 80 and 70 degrees. Time is another good
Ordinal scales are typically measures of non-numeric concepts like satisfaction, happiness, example of an interval scale in which the increments are known, consistent, and measurable.
discomfort, etc.
♦ ♦
between 60 and 50 degrees is a measurable 10 degrees, as is the difference 50 degrees Fahrenheit is equal to the difference between 50 and 60
between 80 and 70 degrees. Time is another good example of an interval degrees Fahrenheit.
scale in which the increments are known, consistent, and measurable.
Interval scales are nice because the realm of statistical analysis on 2.17. RATIO SCALE OF MEASUREMENT
these data sets opens up. For example, central tendency can be measured Ratio scales are the ultimate nirvana when it comes to measurement
by mode, median, or mean; standard deviation can also be calculated. scales because they tell us about the order, they tell us the exact value
Like the others, one can remember the key points of an “interval between units, AND they also have an absolute zero–which allows for
scale” pretty easily. “Interval” itself means “space in between,” which a wide range of both descriptive and inferential statistics to be applied.
is the important thing to remember–interval scales not only tell us about At the risk of repeating myself, everything above about interval data
e‘s the problem with interval scales: they don‘t have a ―true zero.‖ For example, there is no
order, but also about the value between each item. applies to ratio scales + ratio scales have a clear definition of zero.
as ―no temperature.‖ Without
Here’s the a true
problem withzero, it isscales:
interval impossible to compute
they don’t ratios. With interval
have a “true Ratiodata,
scales Good
provide a wealth
examples of possibilities
of ratio when height
variables include it comesandtoweight.
statistical analysis. Thes
zero.” For example, there is no such thing as “no temperature.” Without Ratio scales provide a wealth of possibilities when it comes to
her can add and subtract, but cannot multiply or divide. Confused? Ok, consider can be meaningfully added, subtracted, multiplied, divided (ratios). Central tendency can be m
this: 10 statistical
a true zero, it is impossible to compute ratios. With interval data, the analysis. These variables can be meaningfully added, subtracted,
10 degrees researcher can add and
= 20 degrees. Nosubtract,
problem butthere.
cannot multiply or divide.
20 degrees Confused?
is not mode, median, or mean; measures
twice as hot as 10 degrees, multiplied, divided of (ratios).
dispersion, such
Central as standard
tendency can bedeviation
measured and coefficient of v
by mode,
Ok, consider this: 10 degrees + 10 degrees = 20 degrees. No problem also beone
calculated median,
ratioorscales.
mean; measures of dispersion, such as standard deviation
ecause therethere.
is no20such thing
degrees as ―no
is not temperature‖
twice when it however,
as hot as 10 degrees, comes tobecause
the Celsius scale. hope fromand coefficient of variation can also be calculated from ratio scales.
there is
sense. Bottom no such
line, thingscales
interval as “no temperature”
are great, but when
theitresearcher
comes to thecannot
Celsiuscalculate ratios, which
scale. one hope that makes sense. Bottom line, interval scales are great,
our last measurement scale…
but the researcher cannot calculate ratios, which brings us to our last
measurement scale…
This device provides two examples of Ratio Scales (height and

This device providesweight)
two examples of Ratio Scales (height and weight)
The ratio scale of measurement satisfies all four of the properties
of measurement:
The ratio scale of measurement identity,satisfies
magnitude,
all equal
four intervals, and a minimum
of the properties of measuremen
value of zero. The weight of an object would be an example of a ratio
Interval Scale Example of Interval Scale magnitude, equal intervals, andvalue
scale. Each a minimum value of
on the weight zero.
scale hasThe weight
a unique of an object
meaning, weightswould be an
The interval scale of measurement has the properties of identity, a ratio scale. Each can
valuebe on
rank
theordered,
weight units
scalealong
has athe weightmeaning,
unique scale areweights
equal tocan
onebe rank ord
magnitude, and equal intervals. A perfect example of an interval scale another, and the scale has a minimum value of zero. Weight scales have
interval scale of Fahrenheit
is the measurement scalehas
to the properties
measure of identity,
temperature. magnitude,
The scale is made and equal intervals.
along A scale
the weight are equalvalue
a minimum to one
of another, and the
zero because scaleat has
objects rest acan
minimum value of zero. W
be weightless,
up of equal
mple of an interval temperature
scale units, so that
is the Fahrenheit scalethetodifference
measure between 40 and
temperature. The scale is made
have but they
up of value
a minimum cannot
of zero have negative
because objects weight.
at rest can be weightless, but they cannot ha
weight.
erature units, so that the difference between 40 and 50 degrees Fahrenheit is equal to the
etween 50 and 60 degrees Fahrenheit.
Introduction to Statistics 93
♦
in a specific set of classrooms), the patterns revealed through analysis
of the data collected about that context can be generalized (or predicted
to occur in) similar contexts. The prediction of what may happen in a
similar context is probabilistic. That is, the researcher is not certain that
Chapter-3 the same things may happen in other contexts; instead, the researcher
can only reasonably expect that the same things may happen.
INTRODUCTION TO STATISTICS The words ‘Statistics’ and ‘Statistical’ are all derived from the
Latin word Status, meaning a political state. The theory of statistics as a
distinct branch of scientific method is of comparatively recent growth.
Research particularly into the mathematical theory of statistics is rapidly
proceeding and fresh discoveries are being made all over the world.
3.1. INTRODUCTION TO STATISTICS
Statistics is a set of tools used to organize and analyze data. Data 3.2. MEANING OF STATISTICS
must either be numeric in origin or transformed by researchers into Statistics is concerned with scientific methods for collecting,
numbers. In the modern world of computers and information technology, organizing, summarizing, presenting and analyzing data as well as
the importance of statistics is very well recognized by all the disciplines. deriving valid conclusions and making reasonable decisions on the basis
Statistics has originated as a science of statehood and found applications of this analysis. Statistics is concerned with the systematic collection
slowly and steadily in Agriculture, Economics, Commerce, Biology, of numerical data and its interpretation.
Medicine, Industry, planning, education and so on. As on date there The word ‘statistic’ is used to refer to
is no other human walk of life, where statistics cannot be applied. 1. Numerical facts, such as the number of people living in particular
Employing statistics serves two purposes, (1) description and (2) area.
prediction. Statistics are used to describe the characteristics of groups. 2. The study of ways of collecting, analyzing and interpreting the facts.
These characteristics are referred to as variables. Data is gathered and
recorded for each variable. Descriptive statistics can then be used to reveal Definitions
the distribution of the data in each variable. Descriptive statistics deals Statistics is defined differently by different authors over a period
with the processing of data without attempting to draw any inferences of time. In the olden days statistics was confined to only state affairs
from it. The data are presented in the form of tables and graphs. The but in modern days it embraces almost every sphere of human activity.
characteristics of the data are described in simple terms. Events that Therefore a number of old definitions, which was confined to narrow
are dealt with include everyday happenings such as accidents, prices field of enquiry, were replaced by more definitions, which are much
of goods, business, incomes, epidemics, sports data, population data. more comprehensive and exhaustive. Secondly, statistics has been
Inferential statistics is a scientific discipline that uses mathematical tools defined in two different ways – Statistical data and statistical methods.
to make forecasts and projections by analyzing the given data. This is The following are some of the definitions of statistics as numerical data.
of use to people employed in such fields as engineering, economics, 1. Statistics are the classified facts representing the conditions of people in
biology, the social sciences, business, agriculture and communications. a state. In particular they are the facts, which can be stated in numbers
Statistics is also frequently used for purposes of prediction. or in tables of numbers or in any tabular or classified arrangement.
Prediction is based on the concept of generalizability: if enough data 2. Statistics are measurements, enumerations or estimates of natural
is compiled about a particular context (e.g., students studying writing phenomenon usually systematically arranged, analysed and
presented as to exhibit important interrelationships among them.
92
94 Research Methodology Introduction to Statistics 95
♦ ♦
Definitions by A.L. Bowley Definition by Horace Secrist
Statistics are numerical statement of facts in any department of Statistics may be defined as the aggregate of facts affected to
enquiry placed in relation to each other. - A.L. Bowley a marked extent by multiplicity of causes, numerically expressed,
Statistics may be called the science of counting in one of the enumerated or estimated according to a reasonable standard of accuracy,
departments due to Bowley, obviously this is an incomplete definition collected in a systematic manner, for a predetermined purpose and
as it takes into account only the aspect of collection and ignores other placed in relation to each other. The above definition seems to be the
aspects such as analysis, presentation and interpretation. most comprehensive and exhaustive.
Bowley gives another definition for statistics, which states ‘statistics
may be rightly called the scheme of averages’. This definition is also Functions of Statistics
incomplete, as averages play an important role in understanding and There are many functions of statistics. Let us consider the following
comparing data and statistics provide more measures. five important functions.
What is the definition of statistical data? Condensation

Statistics may be defined as the science of collection, presentation Generally speaking by the word ‘to condense’, the researcher mean
analysis and interpretation of numerical data from the logical analysis. to reduce or to lessen. Condensation is mainly applied at embracing
It is clear that the definition of statistics by the understanding of a huge mass of data by providing only few
Croxton and Cowden is the most scientific and realistic one. observations. If in a particular class in Chennai School, only marks
in an examination are given, no purpose will be served. Instead if the
According to this definition there are four stages researcher are given the average mark in that particular examination,
1. Collection of Data definitely it serves the better purpose. Similarly the range of marks
It is the first step and this is the foundation upon which the entire is also another measure of the data. Thus, Statistical measures help
data set. Careful planning is essential before collecting the data. There are to reduce the complexity of the data and consequently to understand
different methods of collection of data such as census, sampling, primary, any huge mass of data.
secondary, etc., and the investigator should make use of correct method.
Comparison
2. Presentation of data Classification and tabulation are the two methods that are used
The mass data collected should be presented in a suitable, concise to condense the data. They help us to compare data collected from
form for further analysis. The collected data may be presented in the different sources. Grand totals, measures of central tendency measures
form of tabular or diagrammatic or graphic form. of dispersion, graphs and diagrams, coefficient of correlation etc
3. Analysis of data provide ample scope for comparison. If the researcher have one group
of data, the researcher can compare within it. If the rice production
The data presented should be carefully analyzed for making (in Tonnes) in Tanjore district is known, then the researcher can
inference from the presented data such as measures of central tendencies, compare one region with another region within the district. Or if
dispersion, correlation, regression etc., the rice production (in Tonnes) of two different districts within
4. Interpretation of data Tamilnadu is known, then also a comparative study can be made.
The final step is drawing conclusion from the data collected. A As statistics is an aggregate of facts and figures, comparison is
valid conclusion must be drawn on the basis of analysis. A high degree always possible and in fact comparison helps us to understand the
of skill and experience is necessary for the interpretation. data in a better way.
♦ ♦
Forecasting sphere of human activity – social as well as physical – like Biology,
By the word forecasting, the researcher mean to predict or to Commerce, Education, Planning, Business Management, Information
estimate beforehand. Given the data of the last ten years connected to Technology, etc. It is almost impossible to find a single department
rainfall of a particular district in Tamilnadu, it is possible to predict of human activity where statistics cannot be applied. the researcher
or forecast the rainfall for the near future. In business also forecasting now discuss briefly the applications of statistics in other disciplines.
plays a dominant role in connection with production, sales, profits etc.
Statistics and Industry
The analysis of time series and regression analysis plays an important
role in forecasting. Statistics is widely used in many industries. In industries, control
charts are widely used to maintain a certain quality level. In production
Estimation engineering, to find whether the product is conforming to specifications
One of the main objectives of statistics is drawn inference about a or not, statistical tools, namely inspection plans, control charts, etc.,
population from the analysis for the sample drawn from that population. are of extreme importance. In inspection plans the researcher have to
The four major branches of statistical inference are resort to some kind of sampling – a very important aspect of Statistics.
1. Estimation theory Statistics and Commerce
2. Tests of Hypothesis Statistics are lifeblood of successful commerce. Any businessman
3. Non Parametric tests cannot afford to either by under stocking or having overstock of his
4. Sequential analysis goods. In the beginning the researcher estimates the demand for his
In estimation theory, the researcher estimate the unknown value of goods and then takes steps to adjust with his output or purchases. Thus
the population parameter based on the sample observations. Suppose statistics is indispensable in business and commerce.
the researcher are given a sample of heights of hundred students in a As so many multinational companies have invaded into our Indian
school, based upon the heights of these 100 students, it is possible to economy, the size and volume of business is increasing. On one side
estimate the average height of all students in that school. the stiff competition is increasing whereas on the other side the tastes
are changing and new fashions are emerging. In this connection, market
Tests of Hypothesis survey plays an important role to exhibit the present conditions and to
A statistical hypothesis is some statement about the probability forecast the likely changes in future.
distribution, characterising a population on the basis of the information
available from the sample observations. In the formulation and testing Statistics and Agriculture
of hypothesis, statistical methods are extremely useful. Whether crop Analysis of variance (ANOVA) is one of the statistical tools
yield has increased because of the use of new fertilizer or whether the developed by Professor R.A. Fisher, plays a prominent role in agriculture
new medicine is effective in eliminating a particular disease are some experiments. In tests of significance based on small samples, it can
examples of statements of hypothesis and these are tested by proper be shown that statistics is adequate to test the significant difference
statistical tools. between two sample means. In analysis of variance, the researcher are
concerned with the testing of equality of several population means.
3.3. SCOPE OF STATISTICS
Statistics is not a mere device for collecting numerical data, but as Statistics and Economics
a means of developing sound techniques for their handling, analyzing Statistical methods are useful in measuring numerical changes in
and drawing valid inferences from them. Statistics is applied in every complex groups and interpreting collective phenomenon. Nowadays the
♦ ♦
uses of statistics are abundantly made in any economic study. Both in Statistics and Modern applications
economic theory and practice, statistical methods play an important role. Recent developments in the fields of computer technology
Alfred Marshall said, “Statistics are the straw only which one and information technology have enabled statistics to integrate
like every other economists have to make the bricks”. It may also their models and thus make statistics a part of decision making
be noted that statistical data and techniques of statistical tools are procedures of many organisations. There are so many software
immensely useful in solving many economic problems such as packages available for solving design of experiments, forecasting
wages, prices, production, distribution of income and wealth and simulation problems etc.
so on. Statistical tools like Index numbers, time series Analysis, SYSTAT, a software package offers mere scientific and technical
Estimation theory, Testing Statistical Hypothesis are extensively graphing options than any other desktop statistics package. SYSTAT
used in economics. supports all types of scientific and technical research in various
diversified fields as follows
Statistics and Education
1. Archaeology: Evolution of skull dimensions
Statistics is widely used in education. Research has become a
common feature in all branches of activities. Statistics is necessary 2. Epidemiology: Tuberculosis
for the formulation of policies to start new course, consideration of 3. Statistics: Theoretical distributions
facilities available for new courses etc. There are many people engaged 4. Manufacturing: Quality improvement
in research work to test the past knowledge and evolve new knowledge. 5. Medical research: Clinical investigations.
These are possible only through statistics.
6. Geology: Estimation of Uranium reserves from ground water
Statistics and Planning
3.4. IMPORTANCE OF STATISTICS
Statistics is indispensable in planning. In the modern world,
which can be termed as the “world of planning”, almost all the These days statistical methods are applicable everywhere. There is
organizations in the government are seeking the help of planning no field of work in which statistical methods are not applied. According
for efficient working, for the formulation of policy decisions and to A L. Bowley, ‘A knowledge of statistics is like a knowledge of
execution of the same. foreign languages or of Algebra, it may prove of use at any time
In order to achieve the above goals, the statistical data relating under any circumstances”. The importance of the statistical science is
to production, consumption, demand, supply, prices, investments, increasing in almost all spheres of knowledge, e g., astronomy, biology,
income expenditure etc and various advanced statistical techniques meteorology, demography, economics and mathematics. Economic
for processing, analyzing and interpreting such complex data are of planning without statistics is bound to be baseless. Statistics serve in
importance. In India statistics play an important role in planning, administration, and facilitate the work of formulation of new policies.
commissioning both at the central and state government levels. Financial institutions and investors utilize statistical data to summaries
the past experience. Statistics are also helpful to an auditor, when
Statistics and Medicine the researcher uses sampling techniques or test checking to audit the
In Medical sciences, statistical tools are widely used. In order accounts of his client.
to test the efficiency of a new drug or medicine, t - test is used or to
compare the efficiency of two drugs or two medicines, t - test for the 3.4. LIMITATIONS OF STATISTICS
two samples is used. More and more applications of statistics are at Statistics with all its wide application in every sphere of human
present used in clinical investigation. activity has its own limitations. Some of them are given below.
♦ ♦
1. Statistics is not suitable to the study of qualitative phenomenon: 3.5. SOURCE OF DATA

Since statistics is basically a science and deals with a set of numerical The facts and figures which can be numerically measured are studied
data, it is applicable to the study of only these subjects of enquiry, in statistics. A numerical measure of same characteristic is known as
which can be expressed in terms of quantitative measurements. observation and collection of observations is termed as data. Data are
As a matter of fact, qualitative phenomenon like honesty, poverty, collected by individual research workers or by organization through
beauty, intelligence etc, cannot be expressed numerically and any sample surveys or experiments, keeping in view the objectives of the
statistical analysis cannot be directly applied on these qualitative study. The data collected may be:
phenomenon’s. Nevertheless, statistical techniques may be applied 1. Primary Data
indirectly by first reducing the qualitative expressions to accurate
2. Secondary Data
quantitative terms. For example, the intelligence of a group of
students can be studied on the basis of their marks in a particular Primary and Secondary Data in Statistics
examination.
The difference between primary and secondary data in Statistics
2. Statistics does not study individuals: is that Primary data is collected firsthand by a researcher (organization,

Statistics does not give any specific importance to the individual person, authority, agency or party etc) through experiments, surveys,
items; in fact it deals with an aggregate of objects. Individual items, questionnaires, focus groups, conducting interviews and taking (required)
when they are taken individually do not constitute any statistical measurements, while the secondary data is readily available (collected
data and do not serve any purpose for any statistical enquiry. by someone else) and is available to the public through publications,
3 . Statistical laws are not exact: journals and newspapers.
Primary data means the raw data (data without fabrication or not

It is well known that mathematical and physical sciences are exact.
tailored data) which has just been collected from the source and has not
But statistical laws are not exact and statistical laws are only
gone any kind of statistical treatment like sorting and tabulation. The term
approximations. Statistical conclusions are not universally true.
primary data may sometimes be used to refer to firsthand information.
They are true only on an average.
4. Statistics table may be misused: Sources of Primary Data

Statistics must be used only by experts; otherwise, statistical The sources of primary data are primary units such as basic
methods are the most dangerous tools on the hands of the inexpert. experimental units, individuals, households. Following methods are
The use of Statistical tools by the inexperienced and untraced used to collect data from primary units usually and these methods
persons might lead to wrong conclusions. Statistics can be easily depend on the nature of the primary unit. Published data and the data
misused by quoting wrong figures of data. As King says aptly collected in the past is called secondary data.
‘statistics are like clay of which one can make a God or Devil as • Personal Investigation
one pleases’.

The researcher conducts the experiment or survey himself/herself
5. Statistics is only, one of the methods of studying a problem: and collected data from it. The collected data is generally accurate

Statistical method do not provide complete solution of the problems and reliable. This method of collecting primary data is feasible
because problems are to be studied taking the background of the only in case of small scale laboratory, field experiments or pilot
countries culture, philosophy or religion into consideration. Thus surveys and is not practicable for large scale experiments and
the statistical study should be supplemented by other evidences. surveys because it take too much time.
♦ ♦
• Through Investigators Sources of Secondary Data

The trained (experienced) investigators are employed to collect The secondary data may be available from the following sources:
the required data. In case of surveys, they contact the individuals • Government Organizations
and fill in the questionnaires after asking the required information,

Federal and Provincial Bureau of Statistics, Crop Reporting Service-
where a questionnaire is an inquiry form having a number of
Agriculture Department, Census and Registration Organization
questions designed to obtain information from the respondents.
etc
This method of collecting data is usually employed by most of
the organizations and its gives reasonably accurate information • Semi-Government Organization
but it is very costly and may be time taking too.
Municipal committees, District Councils, Commercial Financial
• Through Questionnaire Institutions like banks etc

The required information (data) is obtained by sending a questionnaire • Teaching and Research Organizations
(printed or soft form) to the selected individuals (respondents) (by • Research Journals and Newspapers
mail) who fill in the questionnaire and return it to the investigator. • Internet
This method is relatively cheap as compared to “through investigator”
method but non-response rate is very high as most of the respondents 3.6. POPULATIONS PARAMETERS, SAMPLES, AND STATISTICS
don’t bother to fill in the questionnaire and send it back to investigator. A population is any large collection of objects or individuals, such
• Through Local Sources as Americans, students, or trees about which information is desired.
A parameter is any summary number, like an average or percentage

The local representatives or agents are asked to send requisite
that describes the entire population.
information who provide the information based upon their own
The population mean μ (the greek letter “mu”) and the population
experience. This method is quick but it gives rough estimates only.
proportion p are two different population parameters. For example:
• Through Telephone
• We might be interested in learning about μ, the average weight of

The information may be obtained by contacting the individuals on all middle-aged female Americans. The population consists of all
telephone. Its a Quick and provide accurate required information. middle-aged female Americans, and the parameter is µ.
• Through Internet • Or, the researcher might be interested in learning about p, the

With the introduction of information technology, the people may proportion of likely American voters approving of the president’s
be contacted through internet and the individuals may be asked to job performance. The population comprises all likely American
provide the pertinent information. Google survey is widely used voters, and the parameter is p.
as online method for data collection now a day. There are many The problem is that 99.999999999999... % of the time, the researcher
paid online survey services too. don’t — or can’t — know the real value of a population parameter.
It is important to go through the primary data and locate any The best the researcher can do is estimate the parameter! This is where
inconsistent observations before it is given a statistical treatment. samples and statistics come in to play.
Secondary Data 3.7. SAMPLES AND STATISTICS
Data which has already been collected by someone, may be sorted, A sample is a representative group drawn from the population.
tabulated and has undergone a statistical treatment. It is fabricated or A statistic is any summary number, like an average or percentage,
tailored data. that describes the sample.
♦ ♦
The sample mean, x¯, and the sample proportion p^ are two different • Range, Variance, Standard Deviation
sample statistics. For example: • Identifies the spread of scores by stating intervals
• We might use x¯, the average weight of a random sample of 100 • Range = High/Low points
middle-aged female Americans, to estimate µ, the average weight
• Variance or Standard Deviation = difference between observed
of all middle-aged female Americans.
score and mean
• Or, the researcher might use p^, the proportion in a random sample
• Use this when the researcher want to show how “spread out”
of 1000 likely American voters who approve of the president›s job
the data are. It is helpful to know when the researcher data
performance, to estimate p, the proportion of all likely American
are so spread out that it affects the mean
voters who approve of the president›s job performance.
4. Measures of Position
Because samples are manageable in size, the researcher can
determine the actual value of any statistic. the researcher use the known • Percentile Ranks, Quartile Ranks
value of the sample statistic to learn about the unknown value of the • Describes how scores fall in relation to one another. Relies
population parameter. on standardized scores
• Use this when the researcher need to compare scores to a
3.8. DESCRIPTIVE STATISTICS normalized score (e.g., a national norm)
Descriptive statistics are ways of summarizing large sets of
quantitative (numerical) information. If the researcher have a large Central tendency
number of measurements, the best thing the researcher can do is to Central tendency refers to the idea that there is one number that
make a graph with all the possible scores along the bottom (x axis), best summarizes the entire set of measurements, a number that is in
and the number of times the researcher came across that score recorded some way “central” to the set.
vertically (y axis) in the form of a bar. But such a graph is just plain The mean
hard to do statistical analyses with, so the researcher have other, more
The mean is just the average. It is the sum of all the researcher
numerical ways of summarizing the data.
measurements, divided by the number of measurements. This is the
Types of Descriptive Statistics The mean
most used measure of central tendency, because of its mathematical
1. Measures of Frequency: qualities. It works best ifThe
the mean
data isis just
distributed very
the average. It isevenly
the sumacross theresearcher measu
of all the
• Count, Percent, Frequency range, or is distributed
number inofthe form of aThis
measurements. normal or bell-shaped
is the most used measure ofcurve
central tendency, be
• Shows how often something occurs (see below). One interesting thing about the mean is that it represents
qualities. It works best if the data is distributed very evenly across the range, or
• Use this when the researcher want to show how often a response the expected valueofifathe distribution
normal of measurements
or bell-shaped curve (see below). were random!thing about the
One interesting
is given Here is what the formula looks like:
the expected value if the distribution of measurements were random! Here is wha
2. Measures of Central Tendency

• Mean, Median, and Mode
• Locates the distribution by various points
The median The median
• Use this when the researcher want to show how an average
or most commonly indicated response The median is the number at which
The median half
is the the researcher
number at which halfmeasurements
the researcher measurements ar
are more than thatand
number
half are and halfthatare
less than less than
number. that number.
The median is actually aThe
better measure of c
3. Measures of Dispersion or Variation
the researcher data are skewed, meaning lopsided. If, for example, the research
folks and one millionaire, the distribution of their wealth would be lopsided tow
and the millionaire would be an outlier, or highly deviant member of the group
influence the mean a great deal, making it seem like all the members of the group
The range is the measure from the smallest measurement to the largest one
measure of statistical dispersion or "spread."
The range for our example is 2.2, the distance from the lowest score, 1.8, to the highe
Interquartile range.
♦ ♦ A slightly more sophisticated measure is the interquartile range. If the resea
median is actually a better measure of centrality than the mean if The interquartile range for example
into quartiles, is.one9,fourth
meaning that because the quartiles
of the measurements are in quartile 1, one fou
3, and one fourth in 4, the researcher will get a number that divides 1 and 2 and a num
the researcher data are skewed, meaning lopsided. If, for example, divide roughly at 2.45 and 3.35. The reason for the odd dividing lines
4. The researcher then measure the distance between those two numbers, which the
the researcher have a dozen ordinary folks and one millionaire, the is because there are 15 pieces of data, which, of course, cannot be
the data. Notice that the number between quartile 2 and 3 is the median!
distribution of their wealth would be lopsided towards the ordinary neatly divided into quartiles!
people, and the millionaire would be an outlier, or highly deviant The interquartile range for example is .9, because the quartiles divide roughly at 2.45
The standard deviation
member of the group. The millionaire would influence the mean a for the odd dividing lines is because there are 15 pieces of data, which, of course, ca
great deal, making it seem like all the members of the group are doing The standard deviation is the “average” degree to which scores
into quartiles!
quite well. The median would actually be closer to the mean of all the deviate from the mean. More precisely,
The standard the researcher measure how far
deviation.
people other than the millionaire. all the researcher measurements are from the mean, square each one,
The standard deviation is the "average" degree to which scores deviate f
The median for our example is 3.0. Half the people scored lower, and add them all up. The result is called the variance. Take the square
precisely, the researcher measure how far all the researcher measurements are from
and half higher (and one exactly). root of the variance, and the researcher has the standard deviation. Like
one, and add them all up. The result is called the variance. Take the square root o
the mean, it is the “expected value” of far the scores deviate from the
researcher has the standard deviation. Like the mean, it is the "expected value"
The mode mean. Here is what the formula
deviate fromlooks like:
the mean. Here is what the formula looks like:
The mode is the measurement that has the greatest frequency, the
one the researcher found the most of. Although it isn’t used that much,
it is useful when differences are rare or when the differences are non-
numerical. The prototypical example of something is usually the mode. The normal curve
The normal curve
The example for the mode is 3.2. It is the grade with the most
people (3). At its simplest, the central At its simplest, the central tendency and the measure of dispersion describ
tendency and the measure of dispersion
summary of the set of data. On a more sophisticated level, these measures describ
describe a rectangle that is a summary of the set of data. On a more
normal curve, that contains the data most efficiently.
Statistical dispersion sophisticated level, these measures describe a curve, such as the normal
105
Dispersion refers to the idea that there is a second number which tells curve, that contains the data most efficiently.
us how “spread out” all the measurements are from that central number. This curve, also called the bell-shaped curve, represents a distribution
that reflects certain probabilistic events when extended to an infinite
The range number of measurements. It is an idealized version of what happens
The range is the measure from the smallest measurement to the largest in many large sets of measurements: Most measurements fall in the
one. This is the simplest measure of statistical dispersion or “spread.” middle, and fewer fall at points farther away from the middle. A simple
The range for our example is 2.2, the distance from the lowest example is height: Very few people are below 3 feet tall; very few
score, 1.8, to the highest, 4.0. are over 8 feet tall; most of us are somewhere between 5 and 6. The
same applies to weight, IQs, and SATs! In the normal curve, the mean,
Interquartile range median, and mode are all the same.
A slightly more sophisticated measure is the interquartile range. If
the researchers divide the data into quartiles, meaning that one fourth 3.9. INFERENTIAL STATISTICS
of the measurements are in quartile 1, one fourth in 2, one fourth in 3, With inferential statistics, the researcher is trying to reach
and one fourth in 4, the researcher will get a number that divides 1 and conclusions that extend beyond the immediate data alone. For instance,
2 and a number that divides 3 and 4. The researcher then measure the the researcher use inferential statistics to try to infer from the sample
distance between those two numbers, which therefore contains half of data what the population might think. Or, the researcher use inferential
the data. Notice that the number between quartile 2 and 3 is the median! statistics to make judgments of the probability that an observed difference
♦ ♦
between groups is a dependable one or one that might have happened • The researcher have two separate data sets that are provided by
by chance in this study. Thus, the researcher use inferential statistics the same people, just at different times (e.g. pre/post)
to make inferences from our data to more general conditions; the
researcher use descriptive statistics simply to describe what’s going ANOVA (Analysis of Variance)
on in our data. An ANOVA is a statistical test that is also used to compare means.
The difference between a t-test and an ANOVA is that a t-test can only
t-tests compare two means at a time, whereas with an ANOVA, the researcher
A t-test is a statistical test that can be used to compare means. can compare multiple means at the same time. ANOVAs also allow
There are three basic types of t-tests: one-sample t-test, independent- the researcher to compare the effects of different factors on the same
samples t-test, and dependent-samples (or paired-samples) t-test. measure. ANOVAs can become very complicated, and the analysis
For all t-tests, the researcher are simply looking at the difference should only be done by someone who has been trained in statistics.
between the means and dividing that difference by some measure There are several types of ANOVAs, including: one-way ANOVA,
of variation. within-groups (or repeated-measures) ANOVA, and factorial ANOVA.
One-sample t-test One-way ANOVA

A one-sample t-test can be used to compare the researcher data to A one-way ANOVA is used to compare three or more groups/levels
the mean of some known population. along the same dimension. It is similar to an independent-samples
Thus, use a one-sample t-test when: t-test, just with more groups.
• The researcher have one data set or one mean that the researcher Thus, use a one-way ANOVA when:
are interested in • The researcher has three or more separate, non-overlapping groups
• The researcher know the mean of the population (the entire or data sets that the researcher wants to compare.
population, not a sample!) the researcher wish to compare the
researcher mean to Within-groups (Repeated measures) ANOVA
A within-groups ANOVA is used to compare data from related
Independent-samples t-test groups or the same people over time. This is similar to a dependent-
An independent-samples t-test can be used to compare data from samples t-test, just with more data sets. This is most often used when
two separate, non-related samples. the researcher are doing a longitudinal study that tracks the same
Thus, use an independent-samples t-test when: people across time.
• The researcher has two separate, non-overlapping groups or data Thus, use a within-group ANOVA when:
sets that the researcher wants to compare. That is, different people • The researcher have separate data sets that are provided by the
provided the data for each group. same people over time
Dependent samples t-test Factorial ANOVA

A dependent-samples t-test can be used to compare data from A factorial ANOVA is used when the researcher has two or more
related groups or the same people over time. This is most often used variables/factors/dimensions, and the researcher want to explore whether
when the researcher have a pretest/posttest setup. there are interactions between these factors. Essentially, the researcher
Thus, use a dependent-samples t-test when: are comparing the means of the various combinations of factors.
♦ ♦
Thus, use a factorial ANOVA when: of the population distribution(s) from which one›s data are drawn,
• The researcher are interested in the interaction between two or while a non-parametric test is one that makes no such assumptions.
more variables/factors/dimensions In this strict sense, «non-parametric» is essentially a null category,
One thing that is important to note about ANOVAs is that because since virtually all statistical tests assume one thing or another about
there are more than two groups that are being compared, follow-up (or the properties of the source population(s).
post-hoc) tests are often required to further interpret the data. For instance, For practical purposes, the researcher can think of “parametric”
if the researcher compare Freshmen, Sophomores, Juniors, and Seniors on as referring to tests, such as t-tests and the analysis of variance, that
a measure of leadership skills and find a statistically significant difference, assume the underlying source population(s) to be normally distributed;
the researcher will have to conduct follow-up tests to determine which they generally also assume that one’s measures derive from an equal-
groups are significantly different from each other. These follow-up tests interval scale. And the researcher can think of “non-parametric” as
may show that Freshmen and Sophomores are no different from each referring to tests that do not make on these particular assumptions.
other, nor are Juniors and Seniors, but Juniors and Seniors both have Examples of non-parametric tests include
better leadership skills than either Freshmen or Sophomores. • the various forms of chi-square tests
• the Fisher Exact Probability test
Regression
• the Mann-Whitney Test
A regression analysis is a statistical procedure that allows the researcher
to make a prediction about an outcome (or criterion) variable based on • the Wilcoxon Signed-Rank Test
knowledge of some predictor variable. To create a regression model, the • the Kruskal-Wallis Test
researcher first need to collect (a lot of) data on both variables, similar to • and the Friedman Test
what the researcher would do if the researcher were conducting a correlation. Non-parametric tests are sometimes spoken of as “distribution-
Then the researcher would determine the contribution of the predictor free” tests, although this too is something of a misnomer.
variable to the outcome variable. Once the researcher has the regression
model, the researcher would be able to input an individual’s score on the 3.11. VARIABLE
predictor variable to get a prediction of their score on the outcome variable.
A variable is any characteristics, number, or quantity that
Thus, use regression when:
can be measured or counted. A variable may also be called a data
• The researcher want to be able to make a prediction about an item. Age, sex, business income and expenses, country of birth, capital
outcome given what the researcher already know about some related expenditure, class grades and vehicle type are examples of variables.
factor. It is called a variable because the value may vary between data units
Another option with regression is to do a multiple regression, in a population, and may change in value over time. For example;
which allows the researcher to make a prediction about an outcome ‘income’ is a variable that can vary between data units in a population
based on more than just one predictor variable. Many retention models (i.e. the people or businesses being studied may not have the same
are essentially multiple regressions that consider factors such as GPA, incomes) and can also vary over time for each data unit (i.e. income
level of involvement, and attitude towards academics and learning. can go up or down).
3.10. PARAMETRIC AND NON PARAMETRIC STATISTICS 3.12. TYPES OF THE VARIABLES
In the literal meaning of the terms, a parametric statistical test is There are different ways variables can be described according to the
one that makes assumptions about the parameters (defining properties) ways they can be studied, measured, and presented. Numeric variables have
♦ ♦
values that describe a measurable quantity as a number, like ‘how many’ • Confounding variable: extra variables that have a hidden effect
or ‘how much’. Therefore numeric variables are quantitative variables. on the researcher experimental results.
Numeric variables may be further described as either continuous or discrete: • Control variable: a factor in an experiment which must be held
A continuous variable is a numeric variable. Observations can take constant. For example, in an experiment to determine whether light
any value between a certain set of real numbers. The value given to makes plants grow faster, the researcher would have to control for
an observation for a continuous variable can include values as small soil quality and water.
as the instrument of measurement allows. Examples of continuous
• Dependent variable: the outcome of an experiment. As the
variables include height, time, age, and temperature.
researcher change the independent variable, the researcher watch
A discrete variable is a numeric variable. Observations can take a
what happens to the dependent variable.
value based on a count from a set of distinct whole values. A discrete
variable cannot take the value of a fraction between one value and the • Independent variable: a variable that is not affected by anything
next closest value. Examples of discrete variables include the number that the researcher, the researcher, does. Usually plotted on the x-axis.
of registered cars, number of business locations, and number of children • A measurement variable has a number associated with it. It’s an
in a family, all of of which measured as whole units (i.e. 1, 2, 3 cars). “amount” of something, or a “number” of something.
The data collected for a numeric variable are quantitative data. • Qualitative variable: a broad category for any variable that can’t
Categorical variables have values that describe a ‘quality’ or be counted (i.e. has no numerical value). Nominal and ordinal
‘characteristic’ of a data unit, like ‘what type’ or ‘which category’. variables fall under this umbrella term.
Categorical variables fall into mutually exclusive (in one category or • Quantitative variable: A broad category that includes any variable
in another) and exhaustive (include all possible options) categories. that can be counted, or has a numerical value associated with it.
Therefore, categorical variables are qualitative variables and tend to Examples of variables that fall into this category include discrete
be represented by a non-numeric value. variables and ratio variables.
Categorical variables may be further described as ordinal or nominal
An ordinal variable is a categorical variable. Observations can take a • Random variables are associated with random processes and
value that can be logically ordered or ranked. The categories associated give numbers to outcomes of random events.
with ordinal variables can be ranked higher or lower than another, but • A ranked variable is an ordinal variable; a variable where every
do not necessarily establish a numeric difference between each category. data point can be put in order (1st, 2nd, 3rd, etc.).
Examples of ordinal categorical variables include academic grades • Ratio variables: similar to interval variables, but has a meaningful
(i.e. A, B, C), clothing size (i.e. small, medium, large, extra-large) zero.
and attitudes (i.e. strongly agree, agree, disagree, strongly disagree).
A nominal variable is a categorical variable. Observations can take a Less Common Types of Variables
value that is not able to be organized in a logical sequence. Examples of
• Active Variable: a variable that is manipulated by the researcher.
nominal categorical variables include sex, business type, eye color, religion
and brand. The data collected for a categorical variable are qualitative data. • Attribute variable: another name for a categorical variable (in
statistical software) or a variable that isn’t manipulated (in design
Common Types of Variables of experiments).
• Categorical variable: variables than can be put into categories. • Binary variable: a variable that can only take on two values,
For example, the category “Toothpaste Brands” might contain the usually 0/1. Could also be yes/no, tall/short or some other two-
variables Colgate and Aqua fresh. variable combination.
♦ ♦
• Collider Variable: a variable represented by a node on a causal • Manifest variable: a variable that can be directly observed or
graph that has paths pointing in as well as out. measured.
• Covariate variable: similar to an independent variable, it has an • Manipulated variable: another name for independent variable.
effect on the dependent variable but is usually not the variable of • Mediating variable: variables that explain how the relationship
interest. between variables happens. For example, it could explain the
• Criterion variable: another name for a dependent variable, when difference between the predictor and criterion.
the variable is used in non-experimental situations. • Moderating variable: changes the strength of an effect between
• Dichotomous variable: Another name for a binary variable. independent and dependent variables. For example, psychotherapy
• Dummy Variables: used in regression analysis when the researcher may reduce stress levels for women more than men, so sex moderates
want to assign relationships to unconnected categorical variables. the effect between psychotherapy and stress levels.
For example, if the researcher had the categories “has dogs” and • Nuisance Variable: an extraneous variable that increases variability
“owns a car” the researcher might assign a 1 to mean “has dogs” overall.
and 0 to mean “owns a car.” • Observed Variable: a measured variable (usually used in SEM).
• Endogenous variable: similar to dependent variables, they are • Outcome variable: similar in meaning to a dependent variable,
affected by other variables in the system. Used almost exclusively but used in a non-experimental study.
in econometrics.
• Polychotomous variables: variables that can have more than two
• Exogenous variable: variables that affect others in the system. values.
• Explanatory Variable: a type of independent variable. When • Predictor variable: similar in meaning to the independent variable,
a variable is independent, it is not affected at all by any other but used in regression and in non-experimental studies.
variables. When a variable isn’t independent for certain, it’s an
• Responding variable: an informal term for dependent variable,
explanatory variable.
usually used in science fairs.
• Extraneous variables are any variables that the researcher are
• Scale Variable: basically, another name for a measurement variable.
not intentionally studying in the researcher experiment or test.
• Test Variable: another name for the Dependent Variable.
• A grouping variable (also called a coding variable, group variable
or by variable) sorts data within data files into categories or groups. • Treatment variable: another name for independent variable.
• Identifier Variables: variables used to uniquely identify situations.
Classification of Data
• Indicator variable: another name for a dummy variable.
The process of arranging data into homogenous groups or classes
• Interval variable: a meaningful measurement between two according to some common characteristics present in the data is called
variables. Also sometimes used as another name for a continuous classification.
variable.
• Intervening variable: a variable that is used to explain the Bases of Classification
relationship between variables. There are four important bases of classification:
• Latent Variable: a hidden variable that can’t be measured or (1) Qualitative Base (2) Quantitative Base (3) Geographical Base
observed directly. (4) Chronological or Temporal Base
♦ ♦
(1) Qualitative Base Construction of a Statistical Table
When the data are classified according to a quality or attribute A statistical table has at least four major parts and some other
such as sex, religion, literacy, intelligence, etc. minor parts.
(2) Quantitative Base
1. The Title
When the data are classified by quantitative characteristics like
height, weight, age, income, etc. 2. The Box Head (column captions)
(3) Geographical Base 3. The Stub (row captions)
When the data are classified by geographical regions or location, 4. The Body
like states, provinces, cities, countries, etc. 5. Prefatory Notes
(4) Chronological or Temporal Base
When the data are classified or arranged by their time of occurrence, 6. Foot Notes
such as years, months, weeks, days, etc. 7. Source Notes
Tabulation of Data Frequency Distribution

The process of placing classified data into tabular form is known as A frequency distribution is a tabular arrangement of data into
tabulation. A table is a symmetric arrangement of statistical data in rows classes according to size or magnitude along with corresponding class
and columns. Rows are horizontal arrangements whereas columns are frequencies (the number of values which fall in each class).
vertical arrangements. It may be simple, double or complex depending
upon the type of classification. Ungrouped Data or Raw Data
Data which have not been arranged in a systemic order is called
3.13. TYPES OF TABULATION ungrouped or raw data.
(1) Simple Tabulation or One-way Tabulation
When the data are tabulated to one characteristic, it is said to be Grouped Data
a simple tabulation or one-way tabulation. Data presented in the form of a frequency distribution is called
(2) Double Tabulation or Two-way Tabulation grouped data.
When the data are tabulated according to two characteristics at a
time, it is said to be a double tabulation or two-way tabulation. Array
(3) Complex Tabulation The numerical raw data arranged in ascending or descending order
When the data are tabulated according to many characteristics, it is called an array.
is said to be a complex tabulation.
lass Limits
Difference Between Classification And Tabulation The variant values of the classes or groups are called the class
• First the data are classified and then they are presented in tables, limits. The smallest value of the class is called the lower class limit
and classification and tabulation in fact go together. So classification and the largest value of the class is called the upper class limit. Class
is the basis for tabulation. limits are also called inclusive classes.
• Tabulation is a mechanical function of classification, because in
Class Boundaries
tabulation classified data are placed in rows and columns.
• Classification is a process of statistical analysis whereas tabulation The true values which describe the actual class limits of a class are
is a process of presenting the data in suitable forms. called class boundaries. The smallest true value is called the lower class
♦ ♦
boundary and the largest true value is called the upper class boundary
(h)=Range / Number of Classes(h)
of the class. It is important to note that the upper class boundary of a
In the case of fractional results, the next higher whole number is
class coincides with the lower class boundary of the next class. Class taken as the size of the class interval.
boundaries are also known as exclusive classes.
4. Deciding the starting point:
Open-end Classes
The lower class limit or class boundary should cover the smallest
A class which has no lower class limits or no upper class limit in a value in the raw data. It is a multiple of class intervals.
frequency table is called an open-end class. the researcher do not like to use 5. Determining the remaining class limits (boundary):
open-end classes in practice, because they create problems in calculation.
When the lowest class boundary has been decided, by adding the
lass Mark or Mid-Point class interval size to the lower class boundary the researcher
The class mark or mid-point is the mean of the lower and upper can compute the upper class boundary. The remaining lower and
class limits or boundaries. It divides the class into two equal parts. It upper class limits may be determined by adding the class interval
is obtained by dividing the sum of the lower and upper class limits or size repeatedly till the largest value of the data is observed in
class boundaries of a class by 2. the class.
6. Distribution of data into respective classes:
Size of Class Interval

All the observations are divided into respective classes by using
The difference between the upper and lower class boundaries (not the tally bar (tally mark) method, which is suitable for tabulating
between the class limits) of a class or the difference between two the observations into respective classes. The number of tally bars
successive mid points is called the size of class interval. is counted to get the frequency against each class. The frequency
3.14. CONSTRUCTION OF FREQUENCY DISTRIBUTION of all the classes is noted to get the grouped data or frequency
distribution of the data. The total of the frequency columns must
The following steps are involved in the construction of a frequency
be equal to the number of observations.
distribution.
1. Finding the range of the data: The range is the difference between 3.15. FREQUENCY DISTRIBUTION OF DISCRETE DATA
the largest and the smallest values.
Discrete data is generated by counting, and each and every
2. Deciding the approximate number of classes in which the data observation is exact. When an observation is repeated, it is counted.
are to be grouped. There are no hard and first rules for number of The number for which the observation is repeated is called the frequency
classes. In most cases the researcher have 55 to 2020 classes. H.A. of that observation. The class limits in discrete data are true class limits;
Sturges provides a formula for determining the approximation there are no class boundaries in discrete data.
number of classes.

K=1+3.322logNK=1+3.322logN Cumulative Frequency Distribution

where KK= Number of classes and logNlogN = Logarithm of the The total frequency of all classes less than the upper class boundary
total number of observations. of a given class is called the cumulative frequency of that class.
3. Determining the approximate class interval size: The size of A table showing the cumulative frequencies is called a cumulative
class interval is obtained by dividing the range of data by the frequency distribution. There are two types of cumulative frequency
number of classes and is denoted by hh class interval size distributions.
♦ ♦
Less Than Cumulative Frequency Distribution Simple Bar Chart
It is obtained by adding successively the frequencies of all the A simple bar chart is used to represent data involving only one
previous classes including the class against which it is written. The variable classified on a spatial, quantitative or temporal basis. In a simple
cumulate is started from the lowest to the highest size. bar chart, the researcher make bars of equal width but variable length,
i.e. the magnitude of a quantity is represented by the height or length
More Than Cumulative Frequency Distribution of the bars. The following steps are used to draw a simple bar diagram:
It is obtained by finding the cumulate total of frequencies starting • Draw two perpendicular lines, one horizontally and the other
from the highest to the lowest class. vertically, at an appropriate place on the paper.
3.16. DIAGRAMS AND GRAPHS OF STATISTICAL DATA • Take the basis of classification along the horizontal line (X−X−
axis) and the observed variable along the vertical line (Y−Y− axis),
We have discussed the techniques of classification and tabulation that
or vice versa.
help us organize the collected data in a meaningful fashion. However,
this way of presenting statistical data does not always prove to be • Mark signs of equal breadth for each class and leave equal or not
interesting to a layman. Too many figures are often confusing and fail less than half a breadth between two classes.
to convey the message effectively. • Finally mark the values of the given variable to prepare required bars.
One of the most effective and interesting alternative ways to present
Multiple Bar Chart
statistical data is through diagrams and graphs. There are several ways
in which statistical data may be displayed pictorially, such as different In a multiple bars diagram two or more sets of inter-related data
types of graphs and diagrams. The most commonly used diagrams and are represented (multiple bar diagram facilitates comparison between
graphs shall be discussed in subsequent posts, as listed below. more than one phenomena). The technique of making a simple bar chart
is used to draw this diagram but the difference is that the researcher
Types of Diagrams/Charts use different shades, colors, or dots to distinguish between different
phenomena. the researcher use to draw multiple bar charts if the total
1. Simple Bar Chart
of different phenomena is meaningless.
2. Multiple Bar Chart or Cluster Chart
3. Stacked Bar Chart or Sub-Divided Bar Chart or Component Bar Component Bar Chart
Chart A sub-divided or component bar chart is used to represent data
• Simple Component Bar Chart in which the total magnitude is divided into different or components.
• Percentage Component Bar Chart In this diagram, first the researcher make simple bars for each
class taking the total magnitude in that class and then divide these
• Sub-Divided Rectangular Bar Chart simple bars into parts in the ratio of various components. This type
• Pie Chart of diagram shows the variation in different components within each
class as well as between different classes. A sub-divided bar diagram
Types of Diagrams/Charts is also known as a component bar chart or stacked chart.
1. Histogram
Percentage Component Bar Chart
2. Frequency Curve and Polygon
A sub-divided bar chart may be drawn on a percentage basis. To
3. Lorenz Curve
draw a sub-divided bar chart on a percentage basis, the researcher
122 Research Methodology
♦
express each component as the percentage of its respective total. In
drawing a percentage bar chart, bars of length equal to 100 for each
class are drawn in the first step and sub-divided into the proportion
of the percentage of their component in the second step. The diagram
so obtained is called a percentage component bar chart or percentage
stacked bar chart. This type of chart is useful to make comparisons in
components holding the difference of total constants. REFERENCES
Pie Chart
A pie chart can used to compare the relation between the whole
and its components. A pie chart is a circular diagram and the area of • Ackerly, B. A., & True, J. (2010). Doing feminist research in political
the sector of a circle is used in a pie chart. Circles are drawn with radii and social science. Basingstoke, England: Palgrave Macmillan. H62.
proportional to the square root of the quantities because the area of a A5127 2010
circle is 2πr22πr2. • Agresti, A. (1996) Introduction to Categorical Data Analysis. New York:
To construct a pie chart (sector diagram), the researcher draw Wiley.
a circle with radius (square root of the total). The total angle of the • Aitkin, M. (1978) The analysis of unbalanced cross-classifications (with
circle is 360°360°. The angles of each component are calculated by discussion).Journal of the Royal Statistical Society, A, 41, 195-223.
the formula. • Altman, D. G. (1991) Practical Statistics for Medical Research. London:
Angle of Sector=Component Part/Total×360° Chapman and Hall.
These angles are made in the circle by means of a protractor to • Altman, D. G. (1998) Categorizing continuous variables. In Encyclopedia
show different components. The arrangement of the sectors is usually of Biostatistics Volume 1 (P. Armitage and T. Colton, Eds.). Chichester:
anti-clock wise. Wiley.
• American Psychological Association. (2010). Publication manual of the
American Psychological Association. Washington DC: APA. REFERENCE
AC1.P83 2010
• Anderson, R., & Braud, W. (2011). Transforming self and others through
research: Transpersonal research methods and skills for the human
sciences and humanities. Albany, NY: State University of New York
Press. BF76.5.A68 2011
• Arshan, H. (2007) Excel For Statistical Data Analysis Retrieved from
http://home.ubalt.edu/ntsbarsh/excel/excel.htm 21 June 2009
• Babbie, E.R. (2010). The practice of social research. Belmont, CA:
Wadsworth. H62.B2 2010
• Barone, T., & Eisner, E. W. (2012). Arts Based Research. Los Angeles,
CA: Sage. H62.B3365 2012
• Beck, A. T., Steer, A., and Brown G. K. (1996) Beck Depression Inventory
Manual (2nd ed). San Antonio: The Psychological Corporation.
123
124 Research Methodology References 125
♦ ♦
• Belsley, Kuh, and Welsh (1980). Regression Diagnostics: Identifying • Breslow, N. E. (1970) A generalized Kruskal-Wallace test for comparing
Influential Data and Sources of Collinearity. New York: Wiley. K samples subject to unequal patterns of censorship. Biometrika, 57,
• Berger, R.L., Boos, D.D., and Guess, F.M. (1988) Tests and confidence 579–594.
sets for comparing two mean residual life functions. Biometrics, 44, • Burns, P. (2009) Spreadsheet addiction Retrieved from http://www.
103–115. burnsstat.com/pages/Tutor/spreadsheetaddiction.html, 21 June 2009
• Bernard, H. R. (1998). Handbook of methods in cultural anthropology. Walnut • Chang, H., Ngunjiri, F. W., & Hernandez, K.-A. C. (2013). Collaborative
Creek, CA: AltaMira Press. GN345.H37 1998 autoethnography.Walnut Creek, CA: Left Coast Press. GN307.7.C43
• Bernard, H. R. (2002). Research methods in anthropology: Qualitative 2013
and quantitative approaches. Walnut Creek, CA: Alta Mira Press. GN345. • Cleveland, W. S. (1979) Robust locally weighted regression and smoothing
B36 2002 scatterplots. Journal of the American Statistical Association, 74, 829–836.
• Best, J. (2004). More damned lies and statistics: How numbers confuse • Cleveland, W. S. (1993) Visualizing Data. Summit, NJ: Hobart.
public issues. Berkeley, CA: U of California P. HM535.B474 • Collett, D. (2003) Modeling Binary Data (2nd ed). Boca Raton, FL:
• Bishop, W. (1999). Ethnographic writing research. Portsmouth, NH: Chapman and Hall/CRC.
Heinemann. GN307.7.B57 1999
• Collett, D. (2003) Modelling Survival Data in Medical Research (2nd
• Bloomberg, L. D., & Volpe, M. (2012). Completing your qualitative ed). Boca Raton, FL: Chapman and Hall/CRC.
dissertation: A road map from beginning to end. Thousand Oaks, CA:
• Conover, W. J. (1998) Practical Nonparametric Statistics. New York:
Sage. H62.B58555 2012
John Wiley & Sons.
• Bloor, M. and Wood, F. (2006). Keywords in qualitative methods: A
• Cook, R. D. (1977) Detection of influential observations in linear regression.
vocabulary of research concepts. Thousand Oaks, CA: Sage. H62.B5856
Technometrics, 19, 15–18.
2006
• Cook, R. D. and Weisberg, S. (1982) Residuals and Influence in Regression.
• Bogdan, R.C. and Biklen, S.K. (2007). Qualitative research for education:
London: Chapman and Hall.
An introduction to theories and methods. Boston, MA: Pearson. LB1029.
B56 2007 • correlations between errors in the two-way classification. Annals of
Mathematical Statistics, 25, 484–498.
• Booth, A., Papaioannou, D., & Sutton, A. (2012). Systematic approaches
to a successful literature review. Los Angeles, CA: Sage. LB1047.3.B66 • Cox, D. R. (1972) Regression models and life tables. Journal of the Royal
2012 Statistical Society, B., 34, 187–220.
• Booth, W.C. et al. (2008). The craft of research (3rd ed.). Chicago: U of • Cox, D. R. and Snell, E. J. (1968) A general definition of residuals. Journal
Chicago P. Q180.55.M4B66 2008 of the Royal Statistical Society, B, 30, 248-275.
• Bourque L.B. and Fielder, E.P. (2002). How to conduct self-administered • Creswell, J. W. (2009). Research design: qualitative, quantitative, and
and mail surveys. Thousand Oaks, CA: Sage. HN29.S724 2002. mixed methods approaches. Thousand Oaks, CA: Sage. H62.C6963 2009
• Bourque L.B. and Fielder, E.P. (2002). How to conduct telephone surveys. • Creswell, J. W. (2013). Qualitative inquiry and research design: Choosing
Thousand Oaks, CA: Sage. HN29.S724 2002. among five approaches (3rd ed.). Los Angeles, CA: Sage. H61.C73 2013
• Box, G. E. P. (1954) Some theorems on quadratic forms applied in the • Creswell, J. W., & Plano, C. V. L. (2011). Designing and conducting
study of analysis of variance problems. II. Effects of inequality of variance mixed methods research. Los Angeles, CA: Sage. H62.C6962 2011
and of • Davidson, M. L. (1972) Univariate versus multivariate tests in repeated
• Boyd, N. A., & Roque, R. H. N. (2012). Bodies of evidence: The practice measures experiments. Psychological Bulletin, 77, 446–452.
of queer oral history. New York, NY: Oxford University Press. HQ76.3.U5 • Denzin, N. K., & Lincoln, Y. S. (2013). Strategies of qualitative inquiry (4th
B63 2012 ed.). Los Angeles, CA: Sage. H61.S8823 2013
♦ ♦
• Denzin, N.K. and Lincoln, Y.S. (eds.). (2011). The Sage handbook of • Everitt, B. S. and Wykes, T. (1999) A Dictionary of Statistics for
qualitative research. Thousand Oaks, CA: Sage. H62.H2455 2011 Psychologists. London: Arnold.
• Denzin, N.K., and Lincoln, Y.S. (eds.). (2008). Collecting and Interpreting • Everitt, B. S., Landau, S., and Leese, M. (2001) Cluster Analysis (4th
qualitative materials. Los Angeles, CA: Sage. H62.C566 2008 ed). London: Arnold.
• Der, G. and Everitt, B. S. (2001) A Handbook of Statistical Analysis • Feldman, M.S. et al. (2003). Gaining access: A practical and theoretical
Using SAS (2nd ed). Boca Raton, FL: Chapman and Hall/CRC. guide for qualitative researchers. Walnut Creek, CA: Altamira Press,.
• DeWalt, K. M., & DeWalt, B. R. (2011). Participant observation: A guide H62.G235 2003
for fieldworkers. Lanham, MD: Rowman & Littlefield. GN346.4.D48 2011 • Fetterman, D.M. (2010). Ethnography: Step-by-step. Los Angeles, CA:
• Diggle, P. J. (1988) An approach to the analysis of repeated measures. Sage. GN345.F47 2010
Biometrics, 44, 959–971. • Fink, A. (2002). How to ask survey questions.. Thousand Oaks, CA:
• Dizney, H. and Gromen, L. (1967) Predictive validity and differential Sage. HN29.S724 v.2 2002
achievement on three MLA comparative foreign language tests. Educational • Fink, A. (2002). How to design survey studies. Thousand Oaks, CA:
and Psychological Measurement, 27, 1959-1980. Sage. HN29. S724 v.6 2002
• Draper, N. R. and Smith, H. (1998) Applied Regression Analysis (3rd • Fink, A. (2002). How to sample in surveys. Thousand Oaks, CA: Sage.
ed). New York: Wiley. HN29.S724 v.7 2002
• Dunn, G. and Everitt, B. S. (1995) Clinical Biostatistics: An Introduction • Fink, A. (2002). The survey handbook. Thousand Oaks, CA: Sage. HN29.
to Evidence-Based Medicine. London: Arnold. S724 v.1 2002
• Dytham, C. (2003). Choosing and using statistics: A biologist’s guide. • Fink, A. (2003). How to report on surveys. Thousand Oaks, CA: Sage.
Malden, MA: Blackwell. QH323.5.D98 2003 HN29.S724 v.10 2003
• Everitt, B. S. (1992) The Analysis of Contingency Tables (2nd ed). Boca • Fink, A. (2005). Conducting research literature reviews: From the Internet
Raton, FL: Chapman and Hall/CRC. to paper. Thousand Oaks, CA: Sage. Q180.55.M4F56 2005
• Everitt, B. S. (1999) Chance Rules. New York: Springer Verlag.
• Fink, A. (2006). How to conduct surveys: A step-by-step guide. Thousand
• Everitt, B. S. (2001a) A Handbook of Statistical Analysis Using S-PLUS Oaks, CA: Sage. HN29. F53 2006
(2nd ed).cBoca Raton, FL: Chapman and Hall/CRC.
• Fisher, R. A. (1936) The use of multiple measurements on taxonomic
• Everitt, B. S. (2001b) Statistics for Psychologists. Mahwah, NJ: Lawrence problems. Annals of Eugenics, 7, 179–188.
Erlbaum.
• Foster, J., Barkus, E., and Yavorsky, C. (2006). Understanding and using
• Everitt, B. S. (2002a) The Cambridge Dictionary of Statistics (2nd ed). advanced statistics. Thousand Oaks, CA: Sage. HA29.F583 2006
Cambridge:cCambridge University Press.
• Galman, S. C. (2007). Shane, the lone ethnographer: A beginner’s guide
• Everitt, B. S. (2002b) Modern Medical Statistics: A Practical Guide. to ethnography. Lanham, MD: AltaMira Press. GN345.G35 2007
London: Arnold.
• Galman, S. C. (2013). The good, the bad, and the data: Shane the Lone
• Everitt, B. S. and Dunn, G. (2001) Applied Multivariate Data Analysis ethnographer’s basic guide to qualitative data analysis. Walnut Creek,
(2nd ed). London: Arnold. CA: Left Coast Press. GN345.G34 2013
• Everitt, B. S. and Hay, D. (1992) Talking About Statistics. London: Arnold. • Gamerman, D. (1991) Dynamic Bayesian models for survival data. Applied
• Everitt, B. S. and Pickles, A. (2000) Statistical Aspects of the Design and Statistics, 40, 63-79.
Analysis of Clinical Trials. London: ICP.
• Goldberg, D. (1972) The Detection of Psychiatric Illness by Questionnaire.
• Everitt, B. S. and Rabe-Hesketh, S. (2001) Analysing Medical Data Using Oxford: Oxford University Press.
S-PLUS. New York: Springer.
♦ ♦
• Goldwater, E. (2007) Using Excel for Statistical Data Analysis - Caveats • Howell Major, C. and Savin-Baden, M. (2010). An introduction to
Retrieved from http://wwwunix.oit.umass.edu/evagold/excel.html 27 June qualitative research synthesis: Managing the information explosion
2009 in social science research. New York, NY: Routledge. H62.M23583
• Grbich, C. (2013). Qualitative data analysis: An introduction. London, 2010
England: Sage. H62.G722 2013 • Howell, D. C. (2002) Statistical Methods for Psychology (5th ed). Pacific
• Hancock, D. R., & Algozzine, R. (2011). Doing case study research: Lodge, CA: Duxbury.
A practical guide for beginning researchers. New York, NY: Teachers • Hubbuch, S.M. (2005). Writing research papers across the curriculum.
College Press. LB1028.H313 2011 Boston, MA: Thomson Wadsworth. LB2369.H83 2005
• Hand, D. J. (1997) Construction and Assessment of Classification Rules. • Huberty, C. J. (1994) Applied Discriminant Analysis. New York: Wiley.
Chichester: Wiley. • Huynh, H. and Feldt, L. S. (1976) Estimates of the correction for degrees
• Hand, D. J. (1998) Discriminant Analysis, Linear. In Encyclopedia of of freedom for sample data in randomised block and split-plot designs.
Biostatistics Volume 2 (P. Armitage and T. Colton, Eds.). Chichester: Journal of Educational Statistics, 1, 69–82.
Wiley. • Ivankova, N. V. (2015). Mixed methods applications in action research:
• Hand, D. J., Daly, F., Lunn, A. D., et al. (1994) Small Data Sets. Boca From methods to community action.Thousand Oaks, CA: Sage. LB1028.24.
Raton, FL: Chapman and Hall/CRC. I83 2015
• Harrell, E. E. (2000) Regression Modelling Strategies with Applications • Jaccard, J. and Becker, M.A. (2010). Statistics for the behavioral
to Linear Models, Logistic Regression and Survival Analysis. New York: sciences (5th ed.). Belmont, CA; Wadsworth. BF39.J28 2010
Springer. • Johnson, A.P. (2005). A short guide to action research. Boston, MA:
• Hennink, M. M., Hutter, I., & Bailey, A. (2011). Qualitative research Pearson Education. LB1028.24.J65
methods. London, England: Sage. H85.H44 2011 • Johnson, B. & Christensen, L. (2012). Educational research: Quantitative,
• Hennink, M., Hutter, I., and Bailey, A. (2011). Qualitative research qualitative, and mixed approaches. Los Angeles, CA: Sae. LB1028.J59
methods. Los Angeles, CA: Sage. H85.H44 2011 2012
• Herr, K. and Anderson, G.L. (2005). The action research dissertation. • Jolliffe, I. T. (2002) Principal Components Analysis (2nd ed). New York:
Thousand Oaks, CA: Sage. H62.H447 2005 Springer.
• Hesse-Biber, S. N. (ed.). (2012). Handbook of feminist research: Theory • Kaler, A., & Beres, M. A. (2010). Essentials of field relationships. Walnut
and praxis. Thousand Oaks, CA: Sage. Creek, CA: Left Coast Press. GN346.K36 2010
• Hill, C. E. (2012). Consensual qualitative research: A practical resource • Kapor, M. (1981) Efficiency on Erogocycle in Relation to Knee-Joint
for investigating social science phenomena. Washington, DC: American Angle and Drag. Delhi: University of Delhi.
Psychological Association. H62.C58497 2012 • Keller, D.K. (2006). The Tao of statistics. Thousand Oaks, CA: Sage.
• Hollander, M. and Wolfe, D. A. (1999) Nonparametric Statistical Methods. QA276.K253 2006
New York: Wiley. • Kindon, S., Pain, R., and Kesby, M. (eds). (2007). Participatory action
• Holman, J. S. L., Adams, T. E., & Ellis, C. (Eds.) (2013). Handbook of research approaches and methods. NY: Routledge. HM711.P37 2007
autoethnography.Walnut Creek, CA: Left Coast Press. GN307.7.H65 2013 • Kleinbaum, D. G. and Klein, M. (2002) Logistic Regression — A Self
• Hosmer, D. W. and Lemeshow, S. (1989) Applied Logistic Regression Learning Text. New York: Springer.
(2nd ed). New York: Wiley. • Knüsel, L. (1989) Computergestützte Berechnung Statistischer Verteilungen
• Hosmer, D. W. and Lemeshow, S. (1999) Applied Survival Analysis. Oldenburg, München-Wien (An English version of the program can be
New York: Wiley. obtained at http://www.stat.uni-muenchen.de/˜/knuesel/elv)
♦ ♦
• Knüsel, L. (1998) On the accuracy of statistical distributions in Microsoft • McCullough, B.D. (2008) Microsoft Excel’s ‘Not The Wichmann-Hill’
Excel 97 Computational Statistics & Data Analysis 26 (3) 375-377 random number generators Computational Statistics & Data Analysis 52
• Knüsel, L. (2002) On the reliability of Microsoft Excel XP for statistical (10) 4587-4593
purposes Computational Statistics & Data Analysis 39 (1) 109-111 • Merriam, S. (2009). Qualitative research: A guide to design and
• Knüsel, L. (2005) On the accuracy of statistical distributions in Microsoft implementation. San Francisco, CA: Jossey - Bass. LB 1028.M396 2009
Excel 2003. Computational Statistics and Data Analysis. 48 (3) 445-449 • Metzler, K. (1997). Creative interviewing: The writer’s guide to gathering
• Krzanowski, W. J. and Marriott, F. H. C. (1995) Multivariate Analysis information by asking questions. PN4784.I6M4 1997
Part 2. London: Arnold. • Miles, M. B. (2013). Qualitative data analysis: A methods sourcebook.
• Leedy, P.D. and Ormond, J.E. (2005). Practical research: planning and Los Angeles, CA: Sage. H62.M437 2014
design. Upper Saddle River, NJ: Merrill. Q180.55.M4L43 2005 • Mitchell. M.L. and Jolley, J.M. (2010). Research design explained. Belmont,
• Levene, H. (1960a) Robust tests for the equality of variance. In CA: Wadsworth. BF76.5.M57 2010
Contributions to Probability and Statistics (O. Aikin, Ed.). Stanford, CA: • Nagata, D. K., Kohn-Wood, L., & Suzuki, L. A. (2012). Qualitative
Stanford University Press. strategies for ethnocultural research. Washington, DC: American
• SPSS Inc. (2001a) SPSS 11.0 Advanced Models: Englewood Cliffs, NJ: Psychological Association. BF76.5.Q355 2012
Prentice Hall. • Oishi, S.M. (2002). How to conduct in-person interviews for surveys.
• Litwin, M.S. (2002). How to assess and interpret survey Thousand Oaks, CA: Sage. HN29.S724 v.5 2002
psychometrics. Thousand Oaks, CA: Sage. HN29.S724 2002 v.8 • O’Leary, Z. (2014). The essential guide to doing your research project. Los
• Locke, L.F., Spirduso, W.W., and Silverman, S.J. (2007). Proposals that Angeles, CA: Sage. Q180.55.M4O54 2014
work: A guide for planning dissertations and grant proposals. Thousand • Pan, M. L., & Lopez, M. (2013). Preparing literature reviews: Qualitative
Oaks, CA: Sage. Q180.55.P7L63 2007 and quantitative approaches (4th ed.). Glendale, CA: Pyrczak Pub.
• Madison, D. S. (2012). Critical ethnography: Method, ethics, and Q180.55.E9P36 2013
performance. Thousand Oaks, CA: Sage. GN345.M324 2012 • Pascale, C.-M. (2011). Cartographies of knowledge: Exploring qualitative
• Maxwell, J.A. (2012). A realist approach for qualitative research. Los epistemologies. Thousand Oaks, CA: Sage. H62.P313 2011
Angeles, CA: Sage. H62.M324 2012 • Patten, M. (2007). Understanding research methods: An overview of the
• McCullough, B.D. & Heiser, D.A. (2008) On the accuracy of statistical essentials. (6th ed.). Glendale, CA: Pyrczak Publishing. Q180.55.M4
procedures in Microsoft Excel 2007 Computational Statistics & Data P316 2007
Analysis 52 (10) 4570-4578 • Patton, M.L. (2005). Understanding research methods. Glendale, CA:
• McCullough, B.D. & Wilson, B. (1999) On the accuracy of statistical Pyrczak. Q180.55.M4P38 2005
procedures in Microsoft Excel 97 Computational Statistics & Data Analysis • Ravens, T. (2004) Wissenschaftlich mit Excel arbeiten Verlag Pearson
31 (1) 27-37 Studium
• McCullough, B.D. & Wilson, B. (2002) On the accuracy of statistical • Rea, L.M. (1997). Designing and conducting survey research: a
procedures in Microsoft Excel 2000 and Excel XP Computational Statistics comprehensive guide. San Francisco, CA: Jossey-Bass. HC31.2.R43 1997
& Data Analysis 40 (4) 713-721 • Reason, P. and Bradbury, H. (eds.). (2006). Handbook of action research.
• McCullough, B.D. & Wilson, B. (2005) On the accuracy of statistical Thousand Oaks, CA: Sage. HM571.H36 2006
procedures in Microsoft Excel 2003 Computational Statistics & Data • Ripley, B.D. (2002) Retrieved from http://www.stats.ox.ac.uk/ ripley/
Analysis 49 (4) 1244-1252 RSS2002.pdf 28 June 2009
• McCullough, B.D. (1998) Assessing the reliability of statistical software • Ritchie, D.A. (2003). Doing oral history: A practical guide. New York,
part I, Amer. Statist. 52 358-366 NY: Oxford UP. D16.14.R57 2003
♦ ♦
• Roulston, K. (2010). Reflective interviewing: A guide to theory and practice. • SPSS Inc. (2001c) SPSS 11.0 Syntax Reference Guide: Englewood Cliffs,
Los Angeles, CA: Sage. H62.R678 2010 NJ:Prentice Hall.
• Rubin, A. (2008). Practioner’s guide to using research for evidence-based • SPSS Inc. (2001d) SPSS Base 11.0 for Windows User’s Guide: Englewood
practice. Hoboken, NJ: John Wiley. RC337.R73 2008 Cliffs, NJ:Prentice Hall.
• Saldana, J. (2009). The coding manual for qualitative researchers. Los • Stone Sunstein, B. and Chiseri-Strater, E. (2007). FieldWorking: Reading
Angeles, CA: Sage. H62.S343 2009 and writing research. Boston: Bedford/St. Martin’s. GN 346.S86 2007
• Saldaña, J. (2013). The coding manual for qualitative researchers (2nd • Stringer, E.T. (2007). Action Research. Thousand Oaks, CA: Sage. HV11.
ed.). Los Angeles, CA: Sage. H62.S31856 2013 S835 2007
• Salkind, N. J. (2012). 100 questions (and answers) about research methods. • Therneau, T. M. and Grambsch, P. M. (2000) Modeling Survival Data.
Los Angeles, CA: Sage. H62.S3195 2012 New York: Springer.
• Sanders, L.D. (2010). Discovering research methods in psychology: • Therneau, T. M., Grambsch, P. M., and Fleming, T. R. (1990) Martingale-
A student’s guide. Malden, MA : British Psychological Society/ based residuals for survival models. Biometrika, 77, 147–160.
Blackwell. BF76.5.S245 2010. • Thomas, R.M. & Brubaker, D.L. (2000). Theses and dissertations: A
• Sawitzki, G. (1994) Numerical reliability of data analysis systems. guide to planning, research, and writing. Westport, CT: Bergin & Garvey.
Computational Statistics & Data Analysis 18 (2) 269-286 LB2369.T458 2000
• Sawitzki, G. (1994) Report on the reliability of data analysis systems. • Vogt, W. P., Gardner, D. C., & Haeffele, L. M. (2012). When to use what
Computational Statistics & Data Analysis 18 (2) 289-301 research design. New York, NY: Guilford Press. H62.V623 2012
• Schmuller, J. (2009) Statistical Analysis with Excel For Dummies, 2nd • Wechsler, D. (1974) Wechsler Intelligence Scale for Children — Revised.
Edition John Wiley & Sons New York: Psychological Corp.
• Seidman, I. (2006). Interviewing as qualitative research: A guide for • Witkin, H. A., Oftman, P. K., Raskin, E., and Karp, S. A. (1971) Group
researchers in education and the social sciences.New York, NY: Teachers Embedded Figures Test Manual, Palo Alto, CA: Consulting Psychologist
College Press. H61.28.S45 2006 Press.
• Seidman, I. (2013). Interviewing as qualitative research: A guide for • Yalta, A.T. (2008) On the accuracy of statistical distributions in Microsoft
researchers in education and the social sciences. New York, NY: Teachers Excel 2007 Computational Statistics & Data Analysis 52 (10) 4579-4586
College. H61.28.S45 2013 • Yin, R.K. (2009). Case study research: Design and methods. Los Angeles,
• Silverman, D. (2013). Doing qualitative research. Thousand Oaks, CA: CA: Sage. H62.Y56 2009
Sage. H62.S472 2013
• Simons, H. (2009). Case study research in practice. Los Angeles, CA:   
Sage. H62.S476 2009
• Singleton, Jr., R.A. and Straits, B.C. (2005). Approaches to social
research. New York, NY: Oxford UP. H62.S4776 2005
• Sprague, J. (2005). Feminist methodologies for critical researchers:
Bridging differences. Walnut Creek, CA: AltaMira Press. HQ1180.S65
2005
• SPSS Inc. (2001b) SPSS 11.0 Regression Models: Englewood Cliffs,
NJ: Prentice Hall.
View publication stats

Research Methodology

Uploaded by

Copyright:

Available Formats

Research Methodology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Research Methodology

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Book · December 2018

A. Sivakumar GOVINDARAJAN SINGARAVELU

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

A.P.H. PUBLISHING CORPORATION

Educational research endeavors to eliminate the problems of scholars,

Dr. G. Singaravelu CONTENTS

Chapter-3: Introduction to Statistics 92

Dr. G. Singaravelu holds the Degrees M.A(English).,M.Ed.,M.

• Double negatives in questions should be avoided. The Allocation of Serial Numbers

Independent Information and Restrictions on Values

Degrees of Freedom and Probability Distributions

2.11. FORMULATION OF HYPOTHESIS

This device provides two examples of Ratio Scales (height and

What is the definition of statistical data? Condensation

2. Measures of Central Tendency

One-sample t-test One-way ANOVA

Dependent samples t-test Factorial ANOVA

Tabulation of Data Frequency Distribution

View publication stats

You might also like