BU L NG

FR M HE
B SI S
Master these quality tools
and do your job better
 QUALITY TOOLS

QUALITY CONTROL is about models, methods,

measuring and managing. It’s about uncovering a problem
and finding the solution. It’s about using the right tech-
niques at the right time to make things better.
One of the forefathers of quality, Kaoru Ishikawa,
knew how to make things better. He taught quality and
promoted it in Japan for decades. He believed that 95% of
a company’s problems could be solved by a select number
The Tools
of quality tools.
Histograms p. 20
The powerful collection of tools that Ishikawa had in
mind is referred to by different names: “the old seven,” Control charts p.21
“the first seven” or “the basic seven.” Whatever you call
Pareto analysis p. 22
them, it’s imperative to know them all—inside and out—if
you are to succeed as a quality professional. Cause and effect
We asked seven of QP’s frequent contributors to each
diagrams p. 24
cover one of the tools in about 500 words. Their respec-
tive explanations (presented here in no particular order) Check sheets p. 26
get to the heart of these tools. True, they could have writ- Scatter plots p. 27
ten much, much more, but these illuminating snapshots Stratification p. 28
give you the basics you need to understand—or explain to
others—how these tools are used.
Chances are, as a quality practitioner, you’re famil-
iar with most of the seven. If you feel the need to brush
up more on one or two, however, there are additional
resources listed at the end of each article. QP’s website
(www.qualityprogress.com) offers articles on the basic
tools, and ASQ’s website is stocked with publications
(www.asq.org/books-and-publications.html) and other
resources to help you learn about quality.
Each of these seven tools is indispensable and can
make a difference in the way you work. As the American
psychologist Abraham Maslow noted, “If the only tool you
have is a hammer, everything starts to look like a nail.”

January 2009 • QP 19
 Histograms
Histograms
Statistics and data analysis procedures can be divided that yield numeric or tabular output. Examples of
into two general categories: quantitative techniques quantitative techniques include hypothesis testing,
and graphical techniques. analysis of variance, point estimation, confidence in-
Quantitative techniques are statistical procedures tervals and least-squares regression. Graphical tech-
niques include histograms, scatter plots, proba-
bility plots, residual plots, box plots, block plots
Histogram example / FIGURE 1 and ballots.
Worksheet example Exploratory data analysis (EDA) relies heavily
The Bulldogs bowling team wants to improve its standing in the league. Team members on these and other similar graphical techniques.
decided to study their scores for the past month. The 55 bowling scores are: Graphical procedures are not just tools used
103 107 111 115 115 118 119 121 122 124 124 within an EDA context; they are the shortest path
125 126 127 127 129 134 135 137 138 139 141 to gaining insight into a data set in terms of test-
142 144 145 146 147 148 148 149 150 151 152 ing assumptions, model selection and statistical
153 153 154 155 155 155 156 157 159 160 161 model validation, estimator selection, relation-
163 163 165 165 167 170 172 176 177 183 198 ship identification, factor effect determination
Using the table on the histogram worksheet, they estimate B (the number of bars) and outlier detection. In addition, good statistical
to be seven. The highest score was 198, and the lowest was 103, so the range of
graphics can effectively communicate the under-
values is:
lying message that is present within the data.
R = largest – smallest
A histogram is a graphical display of tabu-
R = 198 – 103 = 95
lated frequencies, which are shown as bars. It
The width of each bar is:
illustrates what proportion of cases fall into
W=R÷B
each of several categories. A histogram differs
W = 95 ÷ 7 = 13.6
from a bar chart in that it is the area, not the
The bowling scores have no decimal places, so the bar width must not have
decimal places. They round 13.6 up to 14. Because 14 is an awkward number to height, of the bar that denotes the value—a
work with, they decide to adjust W to 15. Choosing 100 to be the lower edge of the crucial distinction when the categories are not
first bar, the lower edges of the other bars are
of uniform width. The categories are usually
100 + 15 = 115
specified as nonoverlapping intervals of a vari-
115 + 15 = 130, and so on
able. The categories (bars) must be adjacent.
The histogram they Bowling histogram Figure 1 is an example of a histogram.
drew seems to indicate a 20
double-peaked, or bimodal, Although A. M. Guerry published a histogram
distribution: a group of
Number of players

in 1833, Karl Pearson (1857-1936) first used the

players who score in the 15 word “histogram” in 1891. Pearson was a scien-
low 100s and another
more talented group that tist in Victorian London. As a student at Cam-
scores in the mid-100s. 10 bridge University, Pearson learned to use applied
To improve the team’s
mathematics as a pedagogical tool for determin-
standing, members can try
to improve everyone’s score, 5 ing the truth (in other words, one that provided
which would shift the entire the standards and the means of producing reli-
histogram to the right. Or,
they could focus their efforts 0 able knowledge).
100 115 130 145 160 175 190
on improving the poorer Pearson’s passionate interest in mathemati-
players, which would narrow Score cal statistics was a means to the truth. He estab-
the distribution, making the
team as a whole more consistent. lished the foundations of contemporary math-
Source: Nancy R. Tague, The Quality Toolbox, second edition, ASQ Quality Press, 2005. ematical statistics and helped create the modern

20 QP • www.qualityprogress.com
 QUALITY TOOLS

world view. His statistical method not only transformed reigns or sovereigns or periods of different prime min-
our vision of nature, but also gave scientists a set of isters.”
quantitative tools with which to conduct research. Figure 1 is an example of a histogram with bimodal
Pearson introduced the histogram Nov. 18, 1891. distribution. Other histogram distributions are comb,
While presenting a lecture on maps and chartograms, truncated or heart-cut, and dog food.1
he coined the term to describe a time diagram. He ex- —James J. Rooney
plained that the histogram could be used for historical
REFERENCE
purposes to illustrate blocks of time for “charts about 1. Nancy R. Tague, The Quality Toolbox, ASQ Quality Press, 2005, pp. 298-299.

Control Charts
Control Charts
Control charts are statistically based graphical tools 4. Control limits: Control limits typically appear
used to monitor the behavior of a process. Walter A. at
3 from the process average. More details about
Shewhart developed them in the mid-1920s while work- control limits can be found in the online version of this
ing at Bell Laboratories. More than 80 years later, con- article at www.qualityprogress.com.
trol charts continue to serve as the foundation for sta- 5. Zones: The zones represent the distance be-
tistical quality control. tween each standard deviation and are useful when
The graphical and statistical nature of control charts discussing specific out-of-control rules.
helps us: 6. Rational subgroups: The variation within sub-
• Quantify the variation of a process. groups should be as small as possible to make it easier
• Center a process. to detect subgroup-to-subgroup variation.
• Monitor a process in real time. Notice there are no specification limits present on
• Determine whether to take action on a process. the control chart in Figure 2. This is by design, not by
accident.
The structure of control charts The presence of specification limits on control
Constructing control charts is straightforward and,
more often than not, aided by computer software de-
signed specifically for this purpose. Minitab and JMP,
Control chart structure / FIGURE 2
among others, are commonly used.
Figure 2 illustrates the general form for a control 5 6 4
Measurement unit for quality
characteristic being charted

chart. Its critical components listed here correspond UCL

A
with the numbers in the figure: 2σ
B 3σ
1. X-axis: This axis represents the time order of 1σ
V
subgroups. Subgroups represent samples of data taken C CL
O
from a process. It is critical that the integrity of the C P
1σ
time dimension be maintained when plotting control B 3σ
2σ
charts. A
LCL
2. Y-axis: This axis represents the measured value of
2 1 3 4
the quality characteristic under consideration when using
variables charts. When attributes charts are used, this axis
1 2 3 4 … k
is used to quantify defectives or defects. Subgroup number
3. Center line: The center line represents the pro-
VOP = voice of process LCL = lower control limit
cess average. UCL = upper control limit CL = center line

January 2009 • QP 21
 charts could easily lead to inaction, particularly when fresh and access to documentation for process chang-
a process is out of control but within specification. es is readily available.
As the time between the out-of-control event and
Types of control charts the beginning of the investigation increases, the like-
Control charts can be categorized into two types: vari- lihood of determining root causes diminishes greatly.
ables and attributes. Charts fall into the variables catego- Hence, the motto “time is of the essence” is most ap-
ry when the data to be plotted result from measurement propriate.
of a variable or continuous scale. Attributes charts are
used for count data in which each data element is classi- Highly effective tool
fied in one of two categories, such as good or bad. A control chart is relatively easy to develop and use,
Generally, variables charts are preferred over at- and it can be a highly effective statistical tool when
tributes charts because the data contain more infor- selected properly and used correctly. Its selection and
mation and are typically more sensitive to detecting use alone, however, is not sufficient. When so indicat-
process shifts than attributes charts. A more detailed ed, control charts must be acted on in a timely manner
discussion on variables and attributes charts can be so the root causes may be identified and removed from
found in the online version of this article. the process.
One last thing: When in doubt, avoid tampering with
Using control charts the process.
A control chart that has not triggered any out-of-control —T.M. Kubiak
condition is considered stable, predictable and operating
in a state of statistical control. The variation depicted on BIBLIOGRAPHY
the chart is due to common-cause variation. Chambers, David S., and Donald J. Wheeler, Understanding Statistical Process
Control, second edition, SPC Press, 1972.
Points falling outside the limits are attributed to Kubiak, T. M., and Donald W. Benbow, The Certified Six Sigma Black Belt Hand-
book, second edition, ASQ Quality Press, 2008.
special-cause variation. Such points, regardless of Minitab 15 software, Minitab Inc., State College, PA, 2007.
whether they constitute “good” or “bad” occurrences, Montgomery, Douglas C., Introduction to Statistical Quality Control, fifth edi-
tion, John Wiley & Sons Inc., 2005.
should be investigated immediately while the cause- Western Electric Co., Inc., Statistical Quality Control Handbook, second edi-
tion, Delmar Printing Co.,1958.
and-effect relationships and individual memories are

Pareto
Analysis Pareto Analysis
In 1950, Joseph M. Juran rephrased the theories of Ital- the “vital few” versus the “useful many.”
ian economist Vilfredo Pareto (1848-1923) as the Pareto A Pareto chart graphically displays the relative im-
principle, often referred to as the 80-20 rule. The rule portance of differences among groups of data within a
postulates that in any series of variables (problems or set—a prioritized bar chart. Depicting values from the
errors), a small number will account for most of the ef- highest to the lowest in the form of bars (left to right),
fect (for example, 80% of customer complaints come the Pareto chart has many potential uses for decision
from 20% of customers, or 80% of a company’s profit making, for example:
comes from 20% of products made). Juran referred to • Relative frequency of categories of occurrences.

22 QP • www.qualityprogress.com
 QUALITY TOOLS

• Which 20% of sources caused 80% of the er-

rors.
Monthly complaint summary / TABLE 1

• Relative costs incurred in producing different

types of defectives. Number of Percentage
Complaint type Category complaints of total
• Determining which category or categories
should be the focus of improvement efforts. Too soft (break easily) 1 475 38

Too hard (difficult to chew) 2 5 0

For example
Crackers Are Us (CAU) is a fictitious bakery that Too crumbly (disintegrate easily) 3 65 5
produces crackers for the consumer market. Too bland (not enough flavor or
4 15 1
Crackers are sold to distributors, which sell to seasoning)
retail stores. The product package and the com- Too irregular in shape (not
5 12 1
pany’s website provide contact information for standardized shape)

submitting consumer complaints. CAU’s com- Too unappealing (color, smell,

6 25 2
appearance)
plaint unit logs every complaint. Overall, com-
plaints (the numbers reflect units) for the past Too much salt (or other seasoning) 7 19 2
month were the highest on record. The month’s
Too much fat (trans fat or other fat) 8 85 7
summary is shown in Table 1.
A Pareto chart graphically displays the data Too few beneficial nutrients (too
9 45 4
(see Figure 3). It appears that 78% of the com- much filler)
plaints came from 20% of the complaint catego- Too damaged (soaked, burned,
10 497 40
ries. Further analysis may indicate a need for dented)
nested Pareto charts, which are more discrete
Total complaints 1,243 100
breakdowns of the top 80% or weighting of the
categories by dollar cost.
Ensure the categories chosen are clearly dif-
ferentiated to avoid overlap. The intention of the
graphic is to clarify the data represented. Remem-
ber the data source: Garbage in is garbage out. Pareto chart / FIGURE 3
The Pareto chart is a valuable means for visu-
alizing the relative importance of data.
500 100%
—Russ Westcott

Cumulative percentage line

375 75%
BIBLIOGRAPHY
Hartman, Melissa G., “Separate the Vital Few From the Trivial Many,”
Quality Progress, September 2001, p. 120.
Juran, Joseph M., and A. Blanton Godfrey, eds., Juran’s Quality Hand-
book, fifth edition, McGraw-Hill, 1999, section 5.20-5.24.
250 50%
Stevenson, William J., “Supercharging Your Pareto Analysis, Frequency 80% line
Approach Isn’t Always Appropriate,” Quality Progress, October
2000, pp. 51-55.

125 25%

OTHER RESOURCES
http://personnel.ky.gov/nr/rdonlyres/d04b5458-97eb-4a02-
bde1-99fc31490151/0/paretochart.pdf.
http://quality.dlsu.edu.ph/tools/pareto.html. 0
www.asq.org/learn-about-quality/cause-analysis-tools/overview/ Category 10 1 8 3 9 6 7 2, 4, 5 = 10
pareto.html.
www.au.af.mil/au/awc/awcgate/navy/bpi_manual/pareto.pps. Complaints 497 475 85 65 45 25 19 32 = 1,243
www.ncdot.org/programs/CPI/download/CPIToolbox/PARETO.pdf. Percentage of total 40% 38% 7% 5% 4% 2% 2% 2% = 100%

January 2009 • QP 23
 Cause and
Effect Cause and Effect Diagrams

DiagramS This now-familiar tool was reportedly developed by Ka-

oru Ishikawa of Tokyo University. The Japanese name
is Tokusei Yoin Zu, or “characteristics diagram.”
The cause and effect diagram has been defined as
a “tool for analyzing process dispersion. It is also re-
ferred to as the Ishikawa diagram and the fishbone dia-
1
number of theories in a manner that allows the user
to better understand interrelations.4
Figure 5 illustrates a traditional manufacturing
example in which the tool “identifies many possible
causes for an effect or problem. It can be used to struc-
ture a brainstorming session. It immediately sorts ideas
gram because the complete diagram resembles a fish into useful categories.”5 In a service industry example,
skeleton. The diagram illustrates the main causes and the main characteristic groupings can include people,
sub-causes leading to an effect (symptom).”2 processes, policies or technology. Characteristic group-
Ishikawa defined the effect to control or improve as ings will vary.
the quality characteristic (Y in Figure 4) and the po- Wikipedia adds that “a common use of the Ishikawa
tential causes as the factors (X variables in Figure 4).3 diagram is in product design to identify desirable fac-
He first used the diagram to show the relationship be- tors leading to an overall effect.”6
tween cause and effect in 1943. Chrysler, Ford and General Motors included a high-
When developing the tool, Ishikawa reported that level example of the diagram while discussing control
in almost half of the cases, the reason for variation, or charts for variables in their Statistical Process Control
dispersion, was due to: (SPC) Reference Manual.7
1. Raw materials. Then, they applied the diagram in their Measure-
2. Machinery or equipment. ment Systems Analysis Reference Manual third edi-
3. Work method. tion, (see Online Figure 4) to identify some potential
Joseph M. Juran described the diagram in the sources for measurement system variability,8 using the
context of quality improvement, indicating that it is following characteristic groupings: standard, work-
an example of a graphical method used to arrange a piece, instrument, person and procedure, and environ-
ment.
I have used an adaptation or variation of the dia-
gram over the years. Figure 6 uses the typical fish-
High-level fishbone bone diagram groupings to define inputs that need
for key characteristics / FIGURE 4 verification in the process model graphic. The figure
illustrates that “fundamental quality management
x Material Equipment science recognizes the need for appropriate controls
characteristics
Key product

Key process on the quality of the inputs as well as controls for the
parameters process itself and the output.”9
People y
I have also used a more traditional application of
the diagram (Figure 4) to depict the relationship be-
x Methods Environment tween key product (Y) and key process characteris-
and systems tics (X variables) in a Six Sigma context.
The Society of Automotive Engineers published sev-

24 QP • www.qualityprogress.com
 QUALITY TOOLS

Fishbone diagram / FIGURE 5

Measurement Materials Methods

Lab error Analytical procedure
Raw materials
Ca

Im rat
An

Ca

No
ca
pro ion
lcu

aly

lib

t fo
Supplier City Supplier one
lib n

AK

rat
lat

st

DB
pe

llo
Plant

W-
Wa
Truck Supplier two

ion
io

we
system

2
ter
Sampling

d
Solvent contamination Lab solvent contamination

Dir

Iro
Su

nt
In

ty
Su
In
pp
lab

oo
bo
pp
lab
lie

ls
ttle
lie
Iron in
r

r
Iron in

s
product
product
Rust near Inexperienced Rusty pipes

At

At
sample point analyst In

rea

sa
Materials of construction
ols
ipe

Maintenance Out

mp
cto
To

P584
s

s
rs
dp

E583 o2 tor

mp

le
es

ls

ge

r
oo
se

P560
lin

po
ac

Pu
an

#3 Heat exchanger leak

po

nt

Re

int
E470
ing

ch

P573
Ex

Iro

E5
E4
Ex
en

83
70
Op

Environment Manpower Machines

Source: ASQ

eral papers on Six Sigma in 2007 that applied the dia- Process model / FIGURE 6
gram as a diagnostic tool for problem solving that can
improve service quality (noise)10 and identify potential Verification Verification
Inputs activities activities
causes of manufacturing process variation.11
As with many quality tools, such as failure mode Machinery and equipment
Manpower
effects analysis and control plans, this diagram should
Methods Process Outcomes
be constructed using brainstorming methods by a Materials
cross-functional team to capture broad organizational Measurements
input.
Feedback
–R. Dan Reid
Source: R. Dan Reid, “Auto Industry Drives to Improve Healthcare,” Quality Progress,
November 2007, pp. 56-58.
REFERENCES
1. Joseph M. Juran, ed., Quality Control Handbook, third edition, 1979, pp.
16-20.
2. QP Staff, “Quality Glossary,” Quality Progress, June 2007, www.asq.org/
quality-progress/2007/06/quality-tools/quality-glossary.html.
3. Kaoru Ishikawa, Industrial Engineering and Technology Guide to Quality Transactional Area, SAE International, 2007.
Control, Asian Productivity Organization, 1976, p. 18. 11. Helio Maciel Junior and Luciano Ferreira Rodrigo Castro, Using the Six
4. Juran, Quality Control Handbook, see reference 1. Sigma Methodology for Process Variation Reduction, SAE International,
5. American Society for Quality, “Quality Tools,” www.asq.org/learn-about- 2007.
quality/cause-analysis-tools/overview/fishbone.html.
6. Wikipedia, “Ishikawa Diagram,” http://en.wikipedia.org/wiki/Ishikawa_ OTHER RESOURCES
diagram. A Google web search on this tool will yield many results, including:
7. Chrysler Corp., Ford Motor Co. and General Motors Corp., Statistical Inoue, Michael S. and James L. Riggs, “Describe Your System with Cause and
Process Control (SPC) Reference Manual, 1998, Figure 6, p. 26. Effect Diagrams,” Industrial Engineering, April 1971, pp. 26-31.
8. DaimlerChrysler Corp., Ford Motor Co., General Motors Corp., Ishikawa, Kaoru, “Cause and Effect Diagram,” Proceedings, International
Measurement Systems Analysis Reference Manual, 2002, p. 15. Conference on Quality Control, JUSE, Tokyo, 1969, pp. 607-610.
9. R. Dan Reid, “Auto Industry Drives to Improve Healthcare,” Quality Ishikawa, Kaoru, “Industrial Engineering and Technology Guide to Quality
Progress, November 2007, pp. 56-58. Control,”Asian Productivity Organization, 1976, chapter 3, cause-and-
10. Patrick Garcia, Alfred Baumann and Roland Kölsch, Six Sigma Applied for effect diagram.

January 2009 • QP 25
 Check Sheets Check Sheets
Stop arm, horn, brakes and vacuum! bus and starting the route. If any of these four items
As a 16-year-old going through school bus driver weren’t working, the bus did not move, and the me-
training, this was a very important checklist. As sim- chanic was to be called.
ple as this seems, the instructor boiled down starting So what about the checklist, otherwise known as a
the bus route to these four import checks (see Figure check sheet? What role does this tool play in execut-
7). Yes, there were other inspections to be made. But ing processes? Is there still relevance for such a sim-
when you sat down in that driver’s seat, these were the ple tool in today’s high-tech world? Here is a simple
last things that were to be done prior to moving the review of a quality tool with a very high return on in-
vestment.

Memory jogger
School bus route start Joseph M. Juran considered the check sheet a type of
check sheet / FIGURE 7 lesson learned. He likened the check sheet to a mem-
ory jogger, as a reminder of what to do and what not
Parameter check Yes No to do.1 In environments in which repetitive activity is
1. Stop arm: Does the stop arm deploy and do the stop commonplace, the check sheet is a perfect tool to do
lights work? just that: jog the memory to make sure processes are
2. Horn: Does the horn sound when pressed? followed completely.
3. Brakes: Does the pedal stay off the floor when the
brakes are pumped?
Making the connection between a memory jogger
4. Vacuum: With the stop arm out and the brake and a process is a logical role the check sheet can play.
pedal depressed, does the vacuum gauge read the Perhaps the strongest message ISO 9000:2008 deliv-
appropriate level?
ers is that the reliance on memory is not the way busi-
ness is conducted.

Historical uses
Automated check sheet / FIGURE 8 In case the phrase “check sheet” is a new term for you,
here are a couple of points to consider. If you’ve ever
Training made a grocery list, completed a form of some type
material or executed an inspection plan, you’ve used a check
Task 1
Training sheet.
material From manufacturing to medicine, from the public
domain to the private sector, check sheets have been
Training used to ensure that what is to be accomplished is com-
material pleted in a reproducible and repeatable fashion, oppor-
Click on
Go to website Task 2 tunity after opportunity. Check sheets help drive con-
process
Training
sistency in execution on every occasion.
material

Uses and forms

Training
material Check sheets can be created by watching an individual
Task 3 do a series of tasks and jotting them down as they’re
Training performed. Check sheets can be created by breaking
material
down a process into its critical tasks and capturing

26 QP • www.qualityprogress.com
 QUALITY TOOLS

them in a work instruction. Get started

Today’s technology enables the user to go to a web- Whether the opportunity is in manufacturing, quality
site, call up the process that is to be executed and eas- inspections, medical or IT, check sheets are a helpful
ily access the process (check sheet) for the task that tool. Whether they are simple memory joggers or so-
must be completed. phisticated applications, check sheets accomplish the
Figure 8 shows how automation can assist in devel- same purpose: They help remind the person doing the
oping and implementing a check sheet. As long as the tasks what must be done.
individuals doing the tasks have access, they can get all ––Keith Wagoner
the information they need to execute the process flaw-
lessly. If needed, a simple click can give them access to REFERENCE
1. Joseph M. Juran, Juran on Leadership for Quality: An Executive Handbook,
training material for the elements of the check sheet. The Free Press, 2003, p. 148.

Scatter Plots
Scatter Plots
Plot the data. Is there a statistics professor anywhere find yourself wondering if one is there, chances
on the planet who doesn’t stress that? If you are inves- are there isn’t a relationship worth pursuing.
tigating a potential relationship between two variables, A few words of warning: You may see a
then a scatter plot is the tool to use. relationship, but that does not always mean
A scatter plot is a simple visual form of graphical one variable drives the other (cause and
analysis. Let’s use a general example to flesh out its effect). Both may be driven by a third fac-
usefulness. tor. Just because your grass grows and
You are wondering if temperature at a point in your your neighbor’s grass grows at about the
process is related to the number of defects you observe.
Data from 20 lots have been collected and recorded in
Table 2. Lot data / TABLE 2
Using that data, you can set up a plot with the inde-
pendent variable (temperature) on the x-axis and the Lot Temperature Defects
1 22 2
dependent variable (number of defects in the lot) on
2 29 9
the y-axis. You can plot the 20 observations in this ex-
3 21 3
ample (or in any situation) on a piece of grid paper or 4 30 8
by using a tool such as Excel. 5 25 4
6 23 4
Seeing is believing 7 22 3
8 29 7
In Excel, put the data in columns, highlight the tem-
9 26 7
perature and defects columns, click the Chart Wizard
10 25 5
icon, select XY Scatter and follow the menu prompts 11 28 7
until you’re finished. When the plot is complete, take a 12 27 8
look at the result (Figure 9, p. 28). 13 24 4
Because this is meant to be a visual tool, believe 14 26 5
your eyes. If there appears to be a pattern, such as a 15 25 6
16 27 6
line or a curve, then you have a relationship. What do
17 23 2
you see in the example? There seems to be pretty clear 18 28 8
evidence there is a direct relationship between the two 19 21 1
factors. If you need to look very hard for a pattern or 20 24 5

January 2009 • QP 27
 same rate doesn’t mean one causes the other. They are Next step
both driven by other factors, such as the weather or What do you do next? Talk with the process experts,
fertilization practices. The indication of a relationship search the literature and gather and plot more data.
merely means additional investigation is worthwhile. You may want to better define the relationship by using
additional quality tools more advanced than the ones
dealt with in this collection.
You can quantify a relationship by establishing a corre-
Defects vs. temperature / FIGURE 9 lation coefficient. You also can create a predictive model
of the relationship: It may be linear (for straight lines),
10 quadratic (for curved lines) or some combination.

9
If you want to learn more, look up correlation coeffi-
cient, predictive model, linear relationship or quadratic
8 relationship in Excel using the help function, in your
7 favorite textbook or on the internet.
––Peter E. Pylipow
6
Defects

5
ADDITIONAL INFORMATION
ASQ, “Scatter Diagram,” www.asq.org/learn-about-quality/cause-analysis-
4
tools/overview/scatter.html.
Carillon Technologies Ltd., “Scatter Diagrams, Correlation and Measurement
3 Analysis,” www.carillontech.com/Charts/Scatter%20folder/Scatter.htm.
Joseph D. Conklin, “Test Drives and Data Splits,” Quality Progress, www.asq.
2 org/quality-progress/2008/04/six-sigma/34-per-million-test-drives-and-data-
splits.html.
1 Christine M. Cook, “More Is Not Always Better,” Quality Progress, www.asq.
org/quality-progress/2008/10/statistics-roundtable/more-is-not-always-
better.html.
0
Robert L. Mason and John C. Young, “Transforming Data,” Quality Progress,
20 21 22 23 24 25 26 27 28 29 30 31
www.asq.org/quality-progress/2008/09/statistics-roundtable/statistics-
Temperature roundtable-transforming-data.html.

Stratification
Stratification
Stratification is a fancy word for a simple concept: the problem was worse or better. That way, the airline
breaking down data into categories so you can make could pinpoint a time and investigate what may have
sense of it. The concept is as old as rational thinking been going on then.
itself. For decades, giants of quality improvement, such The team stratified the data by week and produced
as Walter Shewhart, W. Edwards Deming, Joseph M. Ju- a control chart of the weekly data, only to find random
ran and Kaoru Ishikawa, have recommended its use. variation. Further stratification by day of the week and
To illustrate how handy it can be to have this tool in time of day also failed to pinpoint the problem.
your arsenal, consider the following scenario: The team asked whether it was possible that only
certain airports accounted for the high error rate. To
Get a handle on the problem answer the question, it stratified the data by airport and
An airline company was trying to understand its rela- whether it was the departure or arrival city. But, again,
tively high rate of baggage-handling errors. Someone the team found nothing unusual.
asked whether there were certain time periods when Finally, someone came up with the theory that, be-

28 QP • www.qualityprogress.com
 QUALITY TOOLS

Authors
cause passengers were checking in earlier and earlier
these days, the bags were being misplaced in stor-
age areas while awaiting the arrival of the incoming
aircraft. When the team stratified the data by time of
check-in, they found that the majority of the errors
had occurred on bags checked in more than three
JAMES J. ROONEY is a senior risk and reliability engineer with ABS
hours prior to departure.
Consulting, Public Sector Division, in Knoxville, TN. He earned a
master’s degree in nuclear engineering from the University of
Tennessee. Rooney is a fellow of ASQ and holds the following ASQ
Take your pick certifications: biomedical auditor, hazard analysis and critical con-
The airline team stratified—broke down, categorized trol point auditor, manager of quality/organizational excellence,
quality auditor, quality engineer, quality improvement associate,
and separated—the data several ways during their ex- quality process analyst, quality technician, reliability engineer and
Six Sigma Green Belt.
ploration to get closer to the root cause of the problem.
They could have stratified the numbers in many other
T.M. KUBIAK is an author and consultant in Weddington, NC. He
ways: by weight or size of bag, by whether check-in is a co-author of The Certified Six Sigma Black Belt Handbook.
was at curbside or the inside counter, by agent or by Kubiak, a senior member of ASQ, serves on many ASQ boards and
is a past chair of ASQ’s Publication Management Board.
baggage handling crew at the departure site or desti-
nation airport.
In practice, there are always many ways to stratify
data. Knowledge of the system and intuition are the RUSS WESTCOTT is president of R.T. Westcott & Associates in Old
Saybrook, CT. He is an ASQ fellow, a certified manager of quality/
best guides. A cause-effect diagram can also be used
organizational excellence and a quality auditor. He is editor of
to guide this thinking. the third edition of The CMQ/OE Handbook, co-editor of The
Quality Improvement Handbook and author of Simplified Project
Stratification is an underlying tool that is often Management for the Quality Professional and Stepping Up to ISO
used with the other six basic quality tools. In the early 9004:2000. Westcott is also an instructor of the ASQ CMQ/OE
refresher course.
stages of this scenario, the airline team produced a
control chart after stratifying the data by week to see
if there were any abnormal weeks. R. DAN REID, an ASQ fellow and certified quality engineer, is a pur-
chasing manager at General Motors Powertrain. He is co-author
Check sheets often contain columns or rows to of the three editions of QS-9000 and ISO/TS 16949; the Chrysler,
Ford, GM Advanced Product Quality Planning With Control Plan;
tally stratification information, such as time of day or
Production Part Approval Process and Potential Failure Modes
operator. A Pareto diagram stratifies data by catego- and Effects Analysis manuals; ISO 9001:2000; ISO IWA 1, and its
replacement AIAG’s new Business Operating Systems for Health
ries, such as cause or location. Histograms and scatter Care Organizations. Reid also was the first delegation leader of
diagrams can also be used to display and compare the the International Automotive Task Force.

data after stratification.

KEITH WAGONER is a senior service operations manager for
Wachovia Bank in Charlotte, NC. He holds a bachelor’s degree in
Think ahead liberal studies from State University of New York, Albany. He is a
senior member of ASQ and a certified quality engineer.
The most important time to think about stratification
is before you collect data. If the airline team had not
collected the basic data involving time of check-in,
they would not have been able to look at whether that PETER E. PYLIPOW is a senior design excellence engineer with
Vistakon—Johnson and Johnson Vision Care Inc. He earned a
was a factor in the baggage-handling problem without master’s degree in management engineering from Rensselaer
Polytechnic Institute in Troy, NY. Pylipow is a senior member of
taking several months to collect the necessary data. ASQ and an ASQ certified quality manager, quality engineer and
As a result, the team discovered what many qual- quality auditor.

ity professionals already know: A little bit of thinking

and planning in the present will save you—and your
PAUL PLSEK is president of Paul E. Plsek and Associates in
customers—lots of headaches in the future. QP Roswell, GA. He is a full-time consultant on improvement, innova-
tion and change in complex systems. Plsek earned a master’s in
––Paul Plsek
engineering from the Polytechnic Institute of New York in Brook-
lyn. He is a member of ASQ and an author whose books include
ADDITIONAL INFORMATION Creativity, Innovation and Quality (ASQ Quality Press) and Quality
Jack B. ReVelle, “All About Data,” Quality Progress, www.asq.org/ Improvement Tools (Juran Institute Publication).
quality-progress/2006/01/problem-solving/all-about-data.html.

January 2009 • QP 29