BUSD2027 QualityMgmt Module2
BUSD2027 QualityMgmt Module2
BUSD2027 QualityMgmt Module2
of Quality Management
All blue slides have been inserted as supplements as created by subject matter
experts of NorQuest College. They are a part of NorQuest copyright.
Statistical Terms
Mean – add up all the numbers and divide by how many numbers there are
Image: Toews, M.W. (2007). Standard deviation diagram [Digital Graphic]. Retrieved from https://commons.wikimedia.org/wiki/File:Standard_deviation_diagram.svg. Used under CC BY 2.5.
Statistical Terms
Standard Deviation
• Note that xbar is not the average of the whole population, just your sample.
The true average would require 100% of the population. So to ensure that
xbar is useful, we have to ensure that we have a significant sample (i.e.
sample size).
Statistical Terms
Standard Deviation
Notice the dependency in this expression. Xbar is the average of all of the values in
the sample, while X is one of those values. The average value, Xbar, is therefore always
somewhere in the middle of that collection of values because it is dependent on your
sample. Xbar is not some magical number related to the population from which you
drew the samples. It is entirely dependent on the values you included in yours sample.
The true average of the population is not known until 100% of the values in the
population are known. To make the leap to assume that Xbar is a reasonable
approximation of the average value in the population, we must assume that every set
of samples we draw has an equal representation of values that are smaller than the
average and larger than the average - that is, the average deviation to the low side of
average is the same as the average deviation to the high side.
(Gorard, 2004)
Statistical Terms
Standard Deviation
So the formula for standard deviation includes a function that calculates the average
deviation. The average deviation is the sum of deviations divided by the number of
deviations that were included in the sum. Our challenge, is that some of those
deviations are negative (when Xbar is greater than X) and some of them are positive
(when Xbar is less than X).
We could have just added up the deviations and seen how closely the sum approaches
zero; that doesn’t tell us anything about how wide the spread is, just whether our
sample is balanced. To eliminate the negative sign in some of these deviations, we
have two choices. We could ignore the negative sign, and treat all deviations as
positive - that is, take the absolute value of the deviation, or we could square the
negative number, utilizing the rule that two negative numbers multiplied together
results in a positive. The latter, the “sum of squares” approach, has been the
preference of mathematicians because it lends itself to mathematical manipulation.
(Gorard, 2004)
Statistical Terms
Mean Deviation
However, the efficient and accurate approximation of the sample’s standard deviation
to the population’s standard deviation holds true only if the values are considered to
be true values. If there is measurement error associated with the values, then the two
methods are equally valid. In estimation, the mean deviation, calculating the average
absolute difference between individual values and the mean value is preferable to
standard deviation because:
• It is more efficient in calculating the spread of a normally distributed population
where there is a degree of measurement error in the results
• It is more efficient when the population is not perfectly normally distributed, or
normally distributed at all
(Gorard, 2004)
Statistical Terms
Mean Deviation
mean deviation
(Gorard, 2004)
Statistical Terms
z-score – The number of
standard deviations from the
mean
• To convert a value to a z-
score, subtract the mean
and then divide by the
standard deviation
http://www.chem.utoronto.ca/coursenotes/analsci/stats/ttest.html
The calculated values for t and d.o.f. (or d.f.) are used in a table of t values, and
if the calculated value is greater than the value in the table, the two samples, or
the sample and population, are different.
10-13
Overlapping normal distributions. Linacre JM. … Rasch
Measurement Transactions, 1996, 10:1 p.487
https://www.rasch.org/rmt/rmt101r.htm
10-14
Statistical Thinking
• Statistical thinking is a decision-making skill demonstrated
by the ability to draw conclusions based on data.
• Type 2 error
• Consumer’s risk
• Probability that a nonconforming product will be available for sale
• Process charts
• Graphs designed to signal process workers when nonrandom
variation is occurring in a process
• Samples are cheaper, take less time, are less intrusive, and allow
the user to frame the sample.
• When a lot of material has been rejected in the past and materials
must be sorted to keep good materials and return defective
materials for a refund
• Systematic samples
• To sample according to time or according to sequence
Figure 11-3
Copyright © 2017 Pearson Education, Inc. 11-31
Process Control Charts
Statistical process control charts:
• Tools for monitoring process variation
Figure 11-4
• Attribute
• An either-or situation, such as a motor starting or not, or a lens
being scratched or not
Table 11-1
• For example:
• Null Hypothesis: Ho: μ = 11 inches
• Alternative Hypothesis: H1: μ 11 inches
≠
Hypothesis Testing
Figure 11-5
Process Chart
Figure 11-6
Figure 11-7
Figure 11-8
Figure 11-9
Figure 11-10
Table 3.1
Hansen, Bertrand L. Quality Control: Theory and Applications. Upper Saddle River, NJ: Pearson Education
(1964). ISBN: 013745208X. ©1964, p.65. Reprinted and Electronically reproduced by permission of Pearson
Education, Inc., New York, NY.
Figure 11-11
Copyright © 2017 Pearson Education, Inc. 11-46
Interpreting Control Charts
Signals for
concern sent by
a control chart
(cont’d)
Hansen, Bertrand L. Quality Control: Theory and Applications. Upper Saddle River, NJ: Pearson Education
(1964). ISBN: 013745208X. ©1964, p.65. Reprinted and Electronically reproduced by permission of Pearson
Education, Inc., New York, NY.
Figure 11-11
Copyright © 2017 Pearson Education, Inc. 11-47
Interpreting Control Charts
Out-of-control situations:
• Two points in succession farther than two standard deviations
from the mean
• Process run – Five points in succession either above or below the
center line
• Process drift – Seven points, all increasing or decreasing
• Erratic behavior – Large jumps of more than three or four standard
deviations
Figure 11-13
Copyright © 2017 Pearson Education, Inc. 11-52
X and Moving Range (MR) Charts for
Population Data
X and MR charts are used if you have a variable
measurement that you want to monitor and do not have
enough observations to use sampling.
• Central limit theorem does not apply, which may result in the data
being non-normally distributed.
• Therefore, there is an increase in the likelihood that you will draw
an erroneous conclusion.
• It is best to first make sure that the data are normally distributed.
• Limits:
• MR limits
• Same as R chart (where n=2),except that the ranges are computed
as the differences from one sample to the next
• Control limits:
Where:
Table 11-3
Table 11-4
Figure 11-17
Image: SPC for Excel. (2017). Keeping the process on target: CUSUM charts [Digital Graphic]. Retrieved from https://www.spcforexcel.com/knowledge/variable-control-charts/keeping-process-target-cusum-charts.
Understanding Attributes Charts
• Defect – an irregularity or problem with a larger unit
• Countable; can be several within one unit
• Monitored using c and u charts
where:
• The uses for the u chart are the same as the c chart.
Figure 11-18
Figure 11-19
Figure 11-20
Figure 11-21
Where:
Figure 11-22
Where:
Over the years, they thought about the ways people use information in a
highly-automated data processing system.
If you do have a problem, then the fastest way to solve it is to follow these
steps IN ORDER. Order is important here, because often problem solving
becomes a very drawn-out affair with poor results because the process is
followed out of sequence.
i. The object is the item on which the gap exists. This can be an actual
object, like
i. a stapler that doesn’t work, or
ii. a piece of fruit that is bruised
or it be less physical, like a conversation or a concept. The object
represents the entity that is not meeting a standard or an ideal.
ii. The specific gap on, around, or about the object is known as the deviation
ii. The bent staple
iii. The bruise on the fruit
iv. The uncomfortable nature of the conversation, etc.
(Kepner & Tregoe, 1965)
Problem Solving: Step 2
Create a problem specification answering the following questions:
Question Is Is Not
Ask yourself for each statement, if the possible cause is the true cause of the
deviation, how does it explain both the IS and the IS NOT information? For the
possible cause to be the true cause, it must explain all of the IS and IS NOT
statements. List any assumptions you must confirm in order to make the possible
cause fit all of the IS and IS NOT statements.
Note the phrase “most probable”. This reflects the reality that we are suggesting a
true cause, as opposed to definitively stating a true cause. There is a bit of leeway
here that recognizes the imaginative conjecture that we are entertaining as we
move from the problem statement, a collection of facts, and the statements of
possible cause, that require assumptions, facts not yet in evidence, and
imagination.
We will test our assumptions to confirm the true cause, but testing costs time and
money. For now, we want to prioritize which possible cause to test first so that we
might arrive at the true cause most expeditiously.
For example, consider two points on a graph with a horizontal axis for cause, and
a vertical axis for effect. By definition, there is only one line that can pass through
two points on this graph, but there can exist any number of curved lines. Which
would we believe to describe the most probable relationship between cause and
effect represented by those two points? Logic would suggest that the straight line
is the most probable because there are no facts in evidence to support a curved
line, unless we make certain assumptions. The option with the fewest
assumptions, the straight line, is, therefore, the most probable cause.
Note the distinction between problem solving and decision making. The problem
is solved when you determine the most probable cause to be the true cause.
What you do with this information, in terms of eliminating or controlling the
cause, requires decision-making. Albeit subtle, it is an important distinction.
Often the problem solving team only has the mandate to find the cause, but
neither the resources nor the authority to dictate changes to operational policy or
procedure to effect the necessary changes to eliminate or control the cause.
Study – Evaluates the progress and success of the experiment, and areas for
further improvement.
Act - Summarizes the learnings from the study of the implemented plan, from
which a new plan is to be developed.
Other Problem Solving Tools
The 5 Whys
One drawback of the “5 Whys” method is that, for each why question, there may
be many possible causes. Without the ability to analyse the likelihood of each
cause (using distinctions and changes with the IS-IS NOT grid), each cause would
need to be investigated. This is likened to a “Shot Gun Approach”. Spray the area
with activity and hope you get the true cause while implementing a number of
improvements.
Improving the System
To be successful, a business or organization must balance the
needs of these different functional areas:
• Supply chain management
• Marketing
• Accounting
• Human resources
• Operations
• Engineering
• Strategy
Figure 10-1
Figure 10-2
Based on M. Brassard, The Memory Jogger II, published by GOAL/QPC, 2 Manor Parkway, Salem, New Hampshire,
2004. Reprinted with permission of GOAL/QPC.
Figure 10-3
Figure 10-4
Figure 10-5
Figure 10-8
Copyright © 2017 Pearson Education, Inc. 10-132
Histograms
• Graphical representations of data in bar format
Figure 10-10
Figure 10-11
Figure 10-12
Figure 10-13
Reprinted by permission from Patrick Shannon.
Figure 10-15
Figure 10-26
Figure 10-27
Figure 10-28
https://www.sciencedirect.com
/topics/engineering/quality-function-deployment
Quality Function Deployment (QFD)
Figure 7-3
Figure 7-4
Figure 7-5
Figure 7-6
Figure 7-7
Figure 7-8
Figure 7-9
Figure 7-10
Figure 7-11
Figure 7-12
Gorard, S. (2004). Revisiting a 90-year-old debate: The advantages of the mean deviation. Education Online. Retrieved from
http://www.leeds.ac.uk/educol/documents/00003759.htm
Kepner, C.H. & Tregoe, B.B. (1965). The rational manager: A systemmatic approach to problem solving and decision-making.
Princeton, NJ: Kepner-Tregoe, Inc.