IEEE TRANSACTIONS ON EDUCATION, VOL. 50, NO.

4, NOVEMBER 2007 367

Logical Reasoning Ability in Engineering

Students: A Case Study
Jaana Holvikivi

Abstract—Mathematical logic is one of the scientific foundations instruction. Many students in these programs come from Asia,
of information technology and computer engineering. Therefore, Africa, and other countries. The performances of international
engineering students must be proficient in logical operations. First- and home students differ in certain aspects. In particular, in-
year engineering students answered four logical reasoning tasks
as part of a larger survey on scientific thinking, producing more ternational students have specific problems learning computer
incorrect responses than previous studies would have predicted, science, programming, electronics, and other professional
most notably in international student groups. This paper presents engineering subjects. They seem to have no problems in math-
and analyzes these results, and compares and discusses the dif- ematics or physics, a pattern similar to reports from many other
ferences between natural human reasoning and engineering logic. countries [7]–[9]. The author conducted research to understand
It seeks to explain failures in logical reasoning that occur despite
formal training and considers several possible cultural and psycho- better the problems faced by international students in their
logical explanations to this phenomenon. The conclusion highlights professional studies [10]. As a first step, a survey on scientific
the need for stronger emphasis in developing problem-solving and thinking patterns and learning styles was conducted among
metacognitive skills in engineering education. The importance of first-year students. The survey produced surprising results.
fluency in the language of study emerges as a significant factor in Logical reasoning tasks that were assumed to be straightfor-
learning technical subjects.
ward solicited more incorrect responses than previous studies
Index Terms—Cognitive practices, decision making, engineering among college students would have predicted [11].
education, information technology education, international stu-
This paper presents an analysis of student performance in
dents, scientific thinking.
logical reasoning and discusses several possible explanations
for their performance. It first provides a brief overview of the
I. INTRODUCTION theory of logic and reasoning studies in a cross-cultural con-
text. Next, survey methods and results are described. A discus-
sion of reasons and explanations follows. Finally, this paper con-
E NGINEERING is a field of applied sciences based on
mathematics, physics, and chemistry. A wide, functional
understanding of processes, in addition to a good command of
cludes with provisionary implications to engineering education
curricula.
technical skills, is a prerequisite for successful professional en- II. REASONING AND LOGIC
gineering work [1], [2]. For example, fluency in programming
requires deep understanding of abstract concepts, capability in A. Mathematical Logic and Human Judgment
algorithmic thinking, and logical reasoning [3], [4]. Moreover, George Boole wrote in his work on mathematical logic about
many studies indicate that ability in logical reasoning is not 150 years ago: “The design of the following treatise is to in-
separate from overall intellectual capacity and that students vestigate the fundamental laws of those operations of the mind
who master reasoning and problem-solving tend to perform by which reasoning is performed; to give expression to them in
better in any science subject [5], [6]. Computer engineering, in the symbolic language of a Calculus ” [12]. Boole believed
particular, depends on mathematical logic on all levels of com- that human reasoning follows formal logic. The study of log-
puter system design: processing circuits make use of Boolean ical reasoning goes as far back as Aristotle, who is the creator
operations; programming is based on algorithmic thinking; of modal logic and syllogisms with modalities. He introduced
database operations follow Boolean logic; and so forth. There- the qualification of necessary and possible premises [13]. Until
fore, computer engineers are expected to be fluent in logical recently, understanding of an effective decision-making process
thinking. When research data suggest otherwise, a problem that has relied on formal logic and statistics. Braverman’s textbook
requires reflection and invites explanation emerges. on management decision making from 1972 clearly announces:
The present study focuses on engineering students at the “A real decision situation, no matter how complicated, can, with
EVTEK University of Applied Sciences, an engineering col- patience, be broken down into its component parts to any level
lege in the Helsinki area of Finland that has four-year Bachelor of detail desired, a process of reduction; solutions then can be
of Engineering degree programs in information technology found for the individual components; and the overall solution
(IT) and media engineering with English as the language of can be achieved by a continued application of the logic process”
[12].
Moreover, the scientific assumption that people have inherent
Manuscript received January 30, 2007; revised July 17, 2007. logic as a basis for rational thought was strongly influenced by
The author is with EVTEK University of Applied Sciences, FIN-02650
Espoo, Finland (e-mail: [email protected]). the psychological writings of Jean Piaget [14]. However, many
Digital Object Identifier 10.1109/TE.2007.906600 subsequent studies have shown that reality is different. Logical
0018-9359/$25.00 © 2007 IEEE
368 IEEE TRANSACTIONS ON EDUCATION, VOL. 50, NO. 4, NOVEMBER 2007

fallacies are common, and why human reasoning is prone to Naturally, the empirical difficulty in reasoning problems de-
them is of long-standing interest [13], [15]. Typical findings of pends on factors other than reasoning mechanisms per se. If
reasoning experiments show that people often make logical er- subjects fail to understand the statement of the problem in the
rors and draw inferences that are not necessary, but merely plau- intended way, fail to understand the task they are supposed to
sible, based on beliefs. The classic work on human judgment perform, fail to formulate a correct natural-language response,
by Tversky and Kahneman in the 1970s supports the view that or fail to remember the key parts of the problem, their answers
human reasoning systematically violates the standards for statis- do not reflect the correct reasoning process. Several methodical
tical reasoning, ignoring, among other things, base rates, sample difficulties are present in isolating these disparate types of error
size, and correlations. Tversky and Kahneman took a Bayesian in the data when reasoning is studied [19].
account of probabilistic reasoning as a normative standard for To perform something unnatural, such as deductive rea-
an ideally rational agent. They found that humans systematically soning, is difficult because a person has to ignore most normal
fall short of the standard, concluding that “man is apparently not thinking and adhere to the constrained case. Scientists, such as
a conservative Bayesian: he is not a Bayesian at all” [16]. physics researchers, are better trained to think in a restricted,
Consequently, Richardson [15] states that formal decision- artificial system than others. Mortimer and Wertsch [22] explain
making studies need to be replaced by bounded rationality that that the language of science has different grammar, and their
takes the complexity of real world situations into account. Hu- research shows that students are reluctant to switch from the
mans have difficulty isolating a reasoning task from the sur- natural language code to the theoretical speech genre used by
roundings, and concentrating on the given premises only. Rea- science teachers. Moreover, a study among secondary school
soning is highly influenced by context and content. Evans [17] students in three European countries (United Kingdom, France,
claims that the default mode for reasoning is, in fact, pragmatic and Finland) shows that argumentation skills depend largely on
rather than deductive or analytic. People tend to select believ- training; educational systems seem to concentrate on different
able alternatives, choosing something that would be plausible aspects of analytical skills [23].
in the real world instead of automatically following reasoning
rules. Unfortunately, failures in human reasoning are common
C. Cultural or Background Differences in Reasoning
also in critical real-life situations, such as clinical reasoning in
medicine [18]. Are the differences in reasoning and logical thinking mainly
among individuals or among cultures? The answer has signif-
B. Reasoning Explained icant consequences for multicultural engineering education,
particularly computer science education. Some researchers,
Reasoning is a central component of cognition that depends such as Nuckolls [24], claim that inference is an aspect of
on theories of comprehension, memory, learning, visual percep- reasoning that differs cross-culturally. He recalls a classic study
tion, planning, problem solving, and decision making. Rips [19] by the Soviet neuropsychologist Luria that seemed to show
pointed out in 1998 that the mechanics of reasoning were not that East and Central Asian peasants were generally unable
known. How does one belief become another in the brain? Re- to provide answers to syllogisms that contained unfamiliar
cent advances in neuroscience raise expectations that the mech- information. They were asked to conclude what color the
anisms of reasoning processes could soon be explained by the bears in Novaya Zemlya are, given the premises that in the
functioning of the brain. Antonio Damasio [20] proposes that Far North, covered with snow, all bears are white, and that
the brain has two complementary paths for decision making: one Novaya Zemlya is in the Far North, always covered with snow.
for reasoning and the other for instant activation of prior emo- Luria’s informants refused to answer the question insisting that
tional experiences in comparable situations. The second path is they have no firsthand knowledge. Only individuals from the
a kind of “gut feeling” reaction; it activates an emotional signal same culture who had received schooling responded correctly,
that increases the efficiency of the reasoning process and makes demonstrating that the difference was not developmental, but
that process speedier. The “gut feeling” is not a substitute for related to exposure to Western education and styles of reasoning
true reasoning, which takes a longer route. When a situation re- [25]. Accordingly, a study of practical intelligence in Kenya
quires a response, the brain prompts images related to the situ- illustrated how Western schooling influenced test results of
ation, the options for action, and the anticipation of future out- school children. Children from traditionally oriented families
comes. Reasoning strategies then operate on that knowledge to scored high in practical intelligence and low in abstract-rea-
produce a decision. In any situation, both paths are employed. soning-based abilities; whereas, children with more formal
The left hemisphere of the brain is known to be dominant schooling showed a reverse pattern [26].
for language and conscious problem-solving processes, though However, most studies on cross-cultural differences in
not exclusively [21]. The writing ability has been located in the cognitive processes [27] reach the opposite conclusion from
right hemisphere, and some reasoning studies indicate that in- Nuckolls; D’Andrade specifically stresses that reasoning ca-
tuition and heuristic problem-solving rely on right hemisphere pacity seems to be the same across cultures [11]. He cites
functions as well. At present, the author has not been able to several cross-cultural psychological studies, which show no
locate any brain study which combines reasoning and foreign difference in reasoning between people from different cultures
language processing. Nevertheless, the examples found thus far when the given task fits some framework in their world-view.
seem to indicate that neuroscience will, in the near future, prove However, when a reasoning task is purely abstract, most people
valuable in understanding the problems presented in this study. tend to fail everywhere.
HOLVIKIVI: LOGICAL REASONING ABILITY IN ENGINEERING STUDENTS: A CASE STUDY 369

Crump [28] has compared mathematical and number systems a) It must be the case that it is raining.
in many different cultures. He concludes that symbolic counting b) Maybe it is raining and maybe it is not.
is a universal skill that develops gradually with age and that c) It must be the case that it is not raining.
is relatively unaffected by variations in the counting system
1C GIVEN: If Roger drank tea then Tom sat down.
employed, culture, cognitive development, and schooling. The
human brain has a certain capability for mathematic operations, SUPPOSE: Roger did not drink tea.
and the diversity of skills and intelligence is on the individual,
THEN:
not cultural, level. Crump argues that the structure of mathe-
a) It must be the case that Tom sat down.
matics is in some sense isomorphic with the structure of reality.
b) Maybe Tom sat down and maybe he did not.
Numbers as a purely formal system are independent of external
c) It must be the case that Tom did not sit down.
reality.
1D GIVEN: If this rock is a garnet then it is a semiprecious
III. RESEARCH METHOD AND RESULTS stone.

A. Design of the Study SUPPOSE: This rock is not a semiprecious stone.

International students at EVTEK seem to have specific prob- THEN:
lems in learning professional engineering subjects compared to a) It must be the case that this rock is a garnet.
home students [10]. To achieve a preliminary understanding of b) Maybe this rock is a garnet or maybe this rock is
the nature of difficulties, a survey was conducted among engi- not a garnet.
neering students and teaching staff in international and national c) It must be the case that this rock is not a garnet.
degree programs. The survey was intended to be the first step in
a research project that uses a variety of methods, including in- 1A is a denial of antecedent problem with realistic content
terviews and videotaping laboratory sessions. The survey ques- because cutting oneself and bleeding are seen as naturally de-
tionnaire was designed to cover scientific thinking, world-view, pendent. 1B is an affirmation of the consequent problem with
and learning style broadly. It consisted of a four-page question- realistic content. 1C is a denial of antecedent problem with ar-
naire with approximately one hundred questions, starting with bitrary content because drinking tea is an action that has no ob-
reasoning and mathematical questions. vious connection to someone else sitting down. 1D is called
The first four questions were logical statements taken from a modus tollens problem with realistic content. In D’Andrade’s
study by D’Andrade [11]. He compared responses to different sample, 96% of the answers to a modus tollens problem were
kinds of syllogisms in American undergraduate students. The correct. However, with arbitrary content the success rate was
working hypothesis in the present survey was that the results only 33–57%. D’Andrade concludes that modus tollens is the
would be similar or even better because all the respondents study most difficult problem to solve if arbitrary because of the nega-
science and engineering. Another hypothesis was that faculty tion. Incidentally, the correct answers to the above are b, b, b,
would perform better than students. The logical questions were and c.
included as merely a check of overall validity, and no significant
C. Survey Procedure
differences among student groups were expected.
The questionnaire was given to the students in the classroom,
B. Test Statements and they were given as much time as needed to fill it in. The set-
ting was informal, and conversation or asking for clarifications
D’Andrade used four syllogisms to test American undergrad-
was allowed. However, students mostly kept quiet and concen-
uate students [11]. In formal logic, a syllogism is an inference
trated on answering, thus, the answers are individual responses
in which one proposition (the conclusion) follows of necessity
and not a result of group decision making. For faculty, the survey
from two others (known as premises). D’Andrade sought partic-
was conducted in staff meetings, with an introduction about the
ularly to determine whether realistic content solicited better re-
research project in general. Though responding was voluntary,
sults than arbitrary content. Three of the four statements selected
100% of those present participated.
for this study require different types of reasoning, including
The staff members are teachers in the IT and media engi-
cases with realistic and arbitrary content. The statements are as
neering departments. They have a Ph.D. or M.Sc. in engineering
follows [11].
or science; the average age is 50; and all of them are Finns with
the exception of one Estonian. The students were first-year stu-
1A GIVEN: If Jim cut himself then Jim would be bleeding.
dents from three programs. The international IT program had
SUPPOSE: We found out that Jim did not cut himself. a total of 24 respondents from a variety of nationalities: 14
Africans, four Asians, and the rest Finns or of mixed nation-
THEN:
ality. The Finnish logistical engineering group had ten respon-
a) It must be the case that Jim is bleeding.
dents, five of them female. The largest group was that of Finnish
b) Maybe Jim is bleeding and maybe he is not.
IT students: 77 respondents.
c) It must be the case that Jim is not bleeding.
1B GIVEN: If it is raining then the roof is wet. D. Results
When the total number of correct answers by each respondent
SUPPOSE: The roof is wet.
was first calculated, most respondents were expected to have
THEN: correct results because IT students have been trained in Boolean
370 IEEE TRANSACTIONS ON EDUCATION, VOL. 50, NO. 4, NOVEMBER 2007

TABLE I
NUMBER OF SYLLOGISMS SOLVED CORRECTLY IN ALL GROUPS

TABLE II
NUMBER OF CORRECT AND “PLAUSIBLE” ANSWERS IN FINNISH STUDENT GROUPS, AND THE CORRECT ANSWERS RATES IN D’ANDRADE’S STUDY [11]

TABLE III
NUMBER OF CORRECT AND “PLAUSIBLE” ANSWERS BY FACULTY MEMBERS

TABLE IV
NUMBER OF CORRECT AND “PLAUSIBLE” ANSWERS
IN THE INTERNATIONAL STUDENT GROUP
Fig. 1. Correct answering rate comparison.

operations when studying digital circuits. Thus, they are familiar

with AND, OR, and NOT operators; they know that 1 AND 0
yield 0, etc.
Table I shows a summary of the correct answers in all groups.
The surprisingly high overall failure rate prompted closer anal-
ysis. Next, success rates by question were compared in Fig. 1.
Because so many failed in following formal logic, the an- yielding a reasoning pattern similar to question 1A. For ques-
swers were analyzed to see if pragmatic reasoning was applied tion 1D, an indecisive answer was treated as “plausible.” The
instead, as Evans [17] suggests. Answers were taken one by one, results are presented in Tables II–IV.
counting the number of correct answers compared to a “plau- All groups performed worse in all syllogisms than the
sible” answer that would represent everyday reasoning based American student group. However, if people applied proba-
on generalization. Clearly, in the case of Jim who did not cut bilistic understanding, and chose plausible cases, then a good
himself, that he is not bleeding seems plausible. How many re- overall success rate for syllogisms A, B, and C appeared except
spondents gave this choice? Also, in the case of the wet roof, the in the international group. To confirm the results of international
natural explanation is rain, even though it is not certain. In ques- students, the reasoning tasks were given to another comparable
tions 1C and 1D, any plausibility is more difficult to see. 1C is group of 45 students in January 2007. The failure and success
a denial of antecedent problem with arbitrary content, but many rates were nearly the same as before (Table V). Moreover, 31%
solved it as if it were a problem with realistic content anyway, of the students failed in Boolean operations.
HOLVIKIVI: LOGICAL REASONING ABILITY IN ENGINEERING STUDENTS: A CASE STUDY 371

TABLE V level of difficulty. Additionally, analysis of think-aloud studies

NUMBER OF CORRECT ANSWERS IN A FOLLOW-UP STUDY IN 2007 has proven that problem solving is closely connected to repre-
sentation. Even minor differences in the wording of a problem
can lead problem-solvers to very different types of representa-
tions of that problem [30]. A more general conclusion is im-
plied: cross-cultural comparisons of intelligence testing are pre-
carious because uncontrolled language variables influence the
outcome.
Sternberg [27] has introduced the concept of successful
intelligence, which incorporates three aspects of intelligence:
analytical, practical, and creative. His research proves that
instruction which utilizes all three aspects is more successful
IV. DISCUSSION in teaching culturally diverse student groups compared to
The reliability of the results must be analyzed from several traditional instruction that emphasizes memorizing. The meta-
angles. Possible explanations include misunderstanding the task components of intelligent activity are executive processes,
or the problem instead of supposing an incapability of correct which plan and control activity, thus, being instrumental for
reasoning. However, because college students have already been successful performance. In the present study, the reasoning
exposed to a variety of surveys and tests of unclear purpose processes of many students appear to be lacking in these
in secondary school, general confusion is an unsatisfactory ex- metacomponents.
planation. Another general level explanation could be that this No significant difference can be seen in reasoning style
sample of respondents is simply less capable in reasoning than between mathematical logic and philosophical logic. They are
the American reference group. Students who attend universities so closely intertwined that even though students study Boolean
of applied sciences, such as EVTEK, are typically average-level algebra but not verbal reasoning they should see the connec-
high school graduates. As high school results tend to correlate tion and use similar mental procedures. Unfortunately, much
with IQ test results (where reasoning is one factor), one could evidence indicates that the transfer effect of learned mental
assume that this student population is of average intelligence patterns is small when mathematical skills are concerned
[26]. because schooling and abstract operations are not necessarily
Nevertheless, some difficulty in understanding the task is pos- transferred to everyday practice [28]. When the pass/fail rates
sible. Already known from earlier surveys and classroom ob- of equivalent courses for the international and home student
servations, international students often find given instructions groups were compared, they showed a pattern similar to this
difficult to follow [10]. Accordingly, 40% of international stu- study. The passing rate in digital circuits, where Boolean
dents misunderstood the instruction on the last page of the ques- logic is practiced, was 95% for home students and 38% for
tionnaire in this survey, a Kolb learning styles test [29]. The international students.
Kolb test has a list of four alternatives that are rated from 1 to
4, and each alternative must be given a different rating. How-
ever, nearly half of the students in the international group gave V. CONCLUSION
the same rating to several alternatives. In the Finnish student Analysis of the survey disclosed that engineering students,
group, only 5% filled the form incorrectly. The difference is in spite of formal training in mathematical logic, surprisingly
significant and suggests that the previous education system in often applied pragmatic reasoning when given syllogisms to
which an individual has been schooled plays an important part solve. A preference for pragmatic reasoning raises some con-
in his or her ability to follow given instructions. Another study cern over their ability to make good design decisions in their
that reported unsatisfactory reasoning ability in engineering stu- working lives. Though strict logical requirements are not neces-
dents in Palermo attributed the bad performance to the Italian
sary in everyday reasoning, scientific enquiry and professional
high school system, which focuses on humanistic knowledge
decision making in engineering and management require close
[6]. Similarly, many students of this study come from educa-
adherence to logical rules. Even though much of the communi-
tional systems where science is not prioritized.
cation in engineering makes intensive use of charts and formal-
Additionally, the language effect must be seriously consid-
ered. Did the respondents understand the statements correctly? ized graphical expression [31], the design must follow logical
Finnish students answered questions in Finnish, and interna- principles.
tional students in English. Most of the international students are The necessity of logical thinking in professional life coupled
second language speakers despite having attended an English with the poor performance of the student groups surveyed leads
medium high school. Half of the teachers answered the ques- to the conclusion that logical, systematic thinking should be
tions in English and half in Finnish. Teacher responses differed more heavily emphasized in education. Engineers need to be
only in the in the D “modus tollens” question, where the ques- able to select an appropriate way of thinking in each situation,
tion in Finnish gave more correct responses than the English ver- and to be able to alternate between everyday thinking modes,
sion. This pattern agrees with D’Andrade’s conclusion that the strict formal thinking, and creative, heuristic problem solving.
formulation of the question has a strong impact on the result in A good ability to reflect on one’s cognitive functions and high
this particular type of reasoning task. The questions in Finnish metacognitive skills are indispensable. Thus, the aim to develop
were direct translations from English, but as Finnish is struc- metacognitive abilities needs to be explicitly addressed in the
turally very different from English, translation could change the curriculum.
372 IEEE TRANSACTIONS ON EDUCATION, VOL. 50, NO. 4, NOVEMBER 2007

The study also implies that language affects ability in formal [14] J. Piaget and B. Inhelder, The Psychology of the Child. New York:
reasoning more than expected. People may apply different types Basic Books, 1969.
[15] R. C. Richardson, “Heuristics and satisficing,” in A Companion to Cog-
of thinking when using their native language and when using a nitive Science, W. Bechtel and G. Graham, Eds. Oxford, U.K.: Black-
second language. The findings of this study indicate that thor- well, 1998, pp. 566–575.
ough knowledge of language is needed for accurate performance [16] D. Kahneman, P. Slovic, and A. Tversky, Judgment Under Uncer-
tainty: Heuristics and Biases. Cambridge, U.K.: Cambridge Univ.
in reasoning tasks. Evidently, people should not be IQ tested in Press, 1982.
a language in which they are not fluent. The result suggests that [17] J. S. Evans, “Logic and human reasoning: An assessment of the deduc-
language as a medium of study has a stronger effect on learning tion paradigm,” Psych. Bul., vol. 128, pp. 978–996, Nov. 2002.
[18] J. I. Balla, “Logical thinking and the diagnostic process,” Methods Inf.
in science and engineering than is commonly believed. If this Med., vol. 19, pp. 88–92, 1980.
finding is confirmed by later studies, more attention should be [19] L. Rips, “Reasoning,” in A Companion to Cognitive Science, W.
paid to the modes of instruction in foreign language study pro- Bechtel and G. Graham, Eds. Oxford, U.K.: Blackwell, 1998, pp.
299–305.
grams in general. Nevertheless, further study is needed to con- [20] A. Damasio, Looking for Spinoza. Joy, Sorrow and the Feeling Brain.
firm and explain to what extent language influences learning. London, U.K.: Vintage/Random House, 2003.
[21] M. S. Gazzaniga, “The split brain revisited,” Sci. Amer., vol. 12, pp.
REFERENCES 27–31, Aug. 2002.
[22] E. Mortimer and J. Wertsch, “The architecture and dynamics of inter-
[1] J. Moss, K. Kotovsky, and J. Cagan, “The role of functionality in the
subjectivity in science classrooms,” Mind. Culture. Act., vol. 10, pp.
mental representations of engineering students: Some differences in the 230–244, 2003.
early stages of expertise,” Cogn. Sci., vol. 30, pp. 65–93, 2006. [23] M. Marttunen, L. Laurinen, L. Litosseliti, and K. Lund, “Compar-
[2] K. Hakkarainen, T. Palonen, S. Paavola, and E. Lehtinen, Communities ison of argumentation skills among secondary school students in
of Networked Expertise. Professional and Educational Perspectives. Finland, France and the United Kingdom,” in Proc. United Nations
Oxford, U.K.: Elsevier, 2004.
Educational, Scientific and Cultural Organization Conf. Intercultural
[3] R. Faux, “Impact of preprogramming course curriculum on learning in Education, Jyväskylä, Finland, 2003, CD-ROM.
the first programming course,” IEEE Trans. Educ., vol. 49, no. 1, pp.
[24] C. W. Nuckolls, “Cognitive anthropology,” in A Companion to Cogni-
11–16, Feb.. 2006. tive Science, W. Bechtel and G. Graham, Eds. Oxford, U.K.: Black-
[4] A. Eckerdal and A. Berglund, “What does it take to learn ‘Programming
well, 1998, pp. 140–145.
thinking’?,” in Proc. 1st Int. Computing Education Research Workshop, [25] J. V. Wertsch, Vygotsky and the Social Formation of Mind. Cam-
Seattle, WA, 2005, pp. 135–142. brigde, MA: Harvard Univ. Press, 1985.
[5] M. Johnson and A. Lawson, “What are the relative effects of reasoning
[26] R. J. Sternberg, “The theory of successful intelligence,” Revista Inter-
ability and prior knowledge on biology achievement in expository and americana de Psicología/Interamer. J. Psych., vol. 39, pp. 189–202,
inquiry classes?,” J Res. Sci. Teach., vol. 35, pp. 89–103, Jan. 1998.
2005.
[6] M. Capizzo, S. Nuzzo, and M. Zarcone, “The impact of the pre-in- [27] J.W. Berry, Y. H. Poortinga, M. H. Segall, and P. D. Dasden,
structional cognitive profile on learning gain and final exam of physics
Cross-Cultural Psychology. Research and Applications, 2nd
courses: A case study,” Eur. J. Eng. Educ., vol. 31, pp. 717–727, Dec. ed. Cambridge, U.K.: Cambridge Univ. Press, 2002.
2006.
[28] T. Crump, The Anthropology of Numbers. Cambridge, U.K.: Cam-
[7] M. L. Jacobson, R. A. Said, and H. Rehman, “Introducing design skills bridge Univ. Press, 1990.
at the freshman level: Structured design experience,” IEEE Trans. [29] J. Holvikivi, “Learning styles in engineering education: The quest to
Educ., vol. 49, no. 2, pp. 247–253, May 2006.
improve didactic practices,” Eur. J. Eng. Educ., vol. 32, pp. 401–408,
[8] S. Latu and A. Young, “Teaching ICT to Pacific Island background Aug. 2007.
students,” in Proc. 6th Australasian Computing Education Conf., R.
[30] K. Dunbar, “Problem solving,” in A Companion to Cognitive Science,
Lister and A. Young, Eds., Dunedin, New Zealand, 2004, vol. 30, pp. W. Bechtel and G. Graham, Eds. Oxford, U.K.: Blackwell, 1998, pp.
169–175.
289–298.
[9] D. Harnish, “An Arab-American project to improve technical educa- [31] C. Tenopir and D. W. King, Communication Patterns of Engineers.
tion: Sustaining intercultural educational collaboration during tense Piscataway, NJ: IEEE Press, 2004.
times,” in Proc. United Nations Educational, Scientific and Cultural
Organization Conf. Intercultural Education, Jyväskylä, Finland, 2003,
CD-ROM. Jaana Holvikivi received the M.S. degree in operations research and industrial
[10] J. Holvikivi, “Culture and cognition in information technology educa- engineering from Helsinki University of Technology (TKK), Espoo, Finland.
tion,” Eur. J. Eng. Educ., vol. 32, pp. 73–82, Mar. 2007. She is currently working toward the Ph.D. degree at TKK. She is a Senior
[11] R. D’Andrade, The Development of Cognitive Anthropology. Cam- Lecturer in computer science at EVTEK University of Applied Sciences, Espoo,
bridge, U.K.: Cambridge Univ. Press, 1995. Finland. She has worked for 15 years in systems analysis and research in public
[12] J. D. Braverman, Probability, Logic, and Management Decisions. administration in Finland, and as an Information Systems Consultant for the
New York: McGraw-Hill, 1972. United Nations Industrial Development Organization and the United Nations in
[13] J. Hintikka, “Inquiry as inquiry: A logic of scientific discovery,” many parts of the world before entering engineering education. Her interests
in Jaakko Hintikka Selected Papers. Dordrecht, The Netherlands: cover cultural anthropology and science of the mind. Her current research con-
Kluwer, 1996, vol. 5. centrates on improving multicultural information technology education.