Day 3:

Box plot, Capability Cp, Cpk, Pp, Ppk, Analysis

Contents
1 Box plot....................................................................................................................................1
1.1 Type of box plot................................................................................................................1
1.1.1 One Y: Simple............................................................................................................1
1.1.2 One Y: With Groups...................................................................................................1
1.1.3 Multiple Y’s: Simple..................................................................................................2
1.1.4 Multiple Y’s: With Groups........................................................................................2
1.2 Interpret the key results for Boxplot.................................................................................3
1.3 Practice..............................................................................................................................6
1.3.1 One Y: Simple - With Groups....................................................................................6
1.3.2 Multiple Y’s: Simple..................................................................................................7
1.3.3 Note: Exam Data View in Boxplot with data from section 1.3.2..............................7
2 Capability statistics for Between/Within Capability Sixpack..................................................8
2.1 Pp......................................................................................................................................8
2.2 Ppk..................................................................................................................................10
2.3 Cpm.................................................................................................................................12
2.4 PPM Total for Expected Overall Performance................................................................13
2.5 Z.Bench for Overall Capability.......................................................................................14
2.6 Cp (Between/Within)......................................................................................................15
2.7 Cpk (Between/Within)....................................................................................................17
2.8 PPM Total for Expected Between/Within Performance..................................................19
2.9 Z.Bench for Between/Within Capability.........................................................................20
2.10 Practice............................................................................................................................21
2.10.1 Example of Between/Within Capability Analysis...................................................21
2.10.2 Example of Normal Capability Analysis for Multiple Variables.............................21
2.10.3 Example of Between/Within Capability Sixpack....................................................22
2.10.4 Example of Nonnormal Capability Analysis...........................................................22
2.10.5 Example of Nonnormal Capability Analysis for Multiple Variables.......................22
2.10.6 Example of Nonnormal Capability Sixpack............................................................23

1
1 Box plot phát hiện outliers
1.1 Type of box plot
Choose Graph > Boxplot, then select a boxplot based on your data and how you want the plot to
look.
1.1.1 One Y: Simple
Create a boxplot that displays the distribution of values for one variable. The worksheet must
include at least one column of numeric or date/time data.

For example, in the following worksheet, Height contains the numeric data. The graph shows the
distribution of heights.

C1
Height
22
26
19
...

1.1.2 One Y: With Groups

Create a boxplot that displays the distributions of values for different groups. The worksheet
must include at least one column of numeric or date/time data and at least one column of
categorical data. Groups are defined by values in the categorical variable, or by unique
combinations of values in multiple categorical variables.

For example, in the following worksheet, Height contains the numeric data

and Fertilizer contains the categorical data. The graph shows the distributions of heights for each
fertilizer group.

C1 C2-T
Height Fertilizer
22 None
26 GrowFast
19 SuperPlant
... ...

2
1.1.3 Multiple Y’s: Simple
Create a boxplot that displays the distribution of values for different groups. The worksheet must
include multiple columns of numeric or date/time data. Each column is a group.

For example, the following worksheet contains the diameters of pipes produced each week for
three weeks. Week 1, Week 2, and Week 3 contain the numeric data. The graph shows the
distribution of pipe diameters for each week.

C1 C2 C3
Week 1 Week 2 Week 3
5.19 5.57 8.73
5.53 5.11 5.01
4.78 5.76 7.59
... ... ...

1.1.4 Multiple Y’s: With Groups

Create a boxplot that displays the distribution of values for multiple-level groups. Multiple-level
groups are displayed as clustered boxes. The worksheet must include multiple columns of
numeric or date/time data and at least one column of categorical data. Groups are defined by
separate columns and by values or unique combinations of values in categorical variables.

For example, the following worksheet contains the diameters of pipes produced each week for
three weeks, on two machines. Week 1, Week 2, and Week 3 contain the numeric
data. Machine contains the categorical data. The graph shows the distribution of pipe diameters
for each week and for each machine.

C1 C2 C3 C4
Week 1 Week 2 Week 3 Machine
5.19 5.57 8.73 1
5.53 5.11 5.01 2
4.78 5.76 7.59 1
... ... ... ...

1.2 Interpret the key results for Boxplot

Complete the following steps to interpret a boxplot.
Step 1: Assess the key characteristics

3
Examine the center and spread of the distribution. Assess how the sample size may affect the
appearance of the boxplot.
Center and spread
Examine the following elements to learn more about the center and spread of your sample data.
Median
The median is represented by the line in the box. The median is a common measure of the center
of your data. Half the observations are less than or equal to it, and half are greater than or equal
to it.
Interquartile range box
The interquartile range box represents the middle 50% of the data. It shows the distance between
the first and third quartiles (Q3-Q1).
Whiskers
The whiskers extend from either side of the box. The whiskers represent the ranges for the
bottom 25% and the top 25% of the data values, excluding outliers.
Hold the pointer over the boxplot to display a tooltip that shows these statistics. For example,
this boxplot of resting heart rates shows that the median heart rate is 71. Most subjects have a
resting heart rate that is between 64 and 80, but some subjects have heart rates that are as low as
48 and as high as 100.

Investigate any surprising or undesirable characteristics on the boxplot. For example, a boxplot
may show that the median length of wood boards is much lower than the target length of 8 feet.
Sample size (n)
The sample size can affect the appearance of the graph. For example, although these boxplots
seem quite different, both of them were created using randomly selected samples of data from
the same population.

4
 n = 15 n = 500

A boxplot works best when the sample size is at least 20. If the sample size is too small, the
quartiles and outliers shown by the boxplot may not be meaningful. If the sample size is less than
20, consider using an Individual value plot instead. Có thể giải thích dc outlier thì mới loại ra
khỏi bộ dữ liệu.
Step 2: Look for indicators of nonnormal or unusual data
Skewed data indicate that data may be nonnormal. Outliers may indicate other conditions in your
data.
Skewed data
When data are skewed, the majority of the data are located on the high or low side of the graph.
Skewness indicates that the data may not be normally distributed.
These boxplots illustrate skewed data. The boxplot with right-skewed data shows wait times.
Most of the wait times are relatively short, and only a few wait times are long. The boxplot with
left-skewed data shows failure time data. A few items fail immediately and many more items fail
later.

Right-skewed Left-skewed

If you know that your data are not naturally skewed, investigate possible causes. If you want to
analyze severely skewed data, read the data considerations topic for the analysis to make sure
that you can use data that are not normal.
Outliers
Outliers, which are data values that are far away from other data values, can strongly affect your
results. Often, outliers are easiest to identify on a boxplot.
On a boxplot, outliers are identified by asterisks (*).

5
TIP
Hold the pointer over the outlier to identify the data point.
Try to identify the cause of any outliers. Correct any data entry or measurement errors. Consider
removing data values that are associated with abnormal, one-time events (special causes). Then,
repeat the analysis.
Step 3: Assess and compare groups
If your boxplot has groups, assess and compare the center and spread of groups.
Centers
Look for differences between the centers of the groups. For example, this boxplot shows the
thickness of wire produced by four suppliers. The median thicknesses for some groups seem to
be different.

Spreads
Look for differences between the spreads of the groups. For example, this boxplot shows the fill
weights of cereal boxes from four production lines. The median weights of the groups of cereal
boxes are similar, but the weights of some groups are more variable than others.

6
1.3 Practice
1.3.1 One Y: Simple - With Groups.
A plant fertilizer manufacturer wants to develop a formula of fertilizer that yields the most
increase in the height of plants. To test fertilizer formulas, a scientist prepares three groups of 50
identical seedlings: a control group with no fertilizer, a group with the manufacturer's fertilizer,
named GrowFast, and a group with fertilizer named SuperPlant from a competing manufacturer.
After the plants are in a controlled greenhouse environment for three months, the scientist
measures the plants' heights.

As part of the initial investigation, the scientist creates a boxplot of the plant heights from the
three groups to evaluate the differences in plant growth between plants with no fertilizer, plants
with the manufacturer's fertilizer, and plants with their competitor's fertilizer.

 Open the sample data, PlantGrowth.MTW.

 Choose Graph > Boxplot > One Y > With Groups.
 In Graph variables, enter Height.
 In Categorical variables for grouping (1-4, outermost first), enter Fertilizer.
 Click OK.
1.3.2 Multiple Y’s: Simple
• Open the sample data, PlantGrowth_2.MTW.
• Choose Graph > Boxplot > One Y > Multiple Y’s: Simple.
• In Graph variables, enter 'Zone A' 'Zone B' 'Zone C'.
• Click OK.

7
1.3.3 Note: Exam Data View in Boxplot with data from section 1.3.2.

WORKSHEET 2

Boxplot of Zone A, Zone B, Zone C

Click 1 by one

8
2 Capability statistics for Between/Within Capability Sixpack
2.1 Pp
Pp is a measure of the overall capability of the process. Pp is a ratio that compares two values:
The specification spread (USL – LSL)
The spread of the process (the 6-σ variation) based on the overall standard deviation
Pp evaluates overall capability based on the variation in the process, but not on its location.
You must provide both a lower specification limit (LSL) and an upper specification limit (USL)
to calculate the Pp index.

Interpretation
Use Pp to evaluate the overall capability of your process based on the process spread. Overall
capability indicates the actual performance of your process that your customer experiences over
time.
Because Pp does not consider the location of the process, it indicates the overall capability that
your process could achieve if it were centered. Generally, higher Pp values indicate a more
capable process. Lower Pp values indicate that your process may need improvement.

Low Pp
In this example, the specification spread is less than the overall process spread. Therefore, Pp is
low (0.40), and the overall capability of the process is poor based on its variability.

9
High Pp
In this example, the specification spread is considerably greater than the overall process spread.
Therefore, Pp is high (1.80), and the overall capability of the process is good based on its
variability.
You can compare Pp to other values to get more information about the capability of your process.
Compare Pp to a benchmark value to assess the overall capability of your process.  Many industries
use a benchmark value of 1.33. If Pp is lower than your benchmark, consider how to improve your
process by reducing its variation.
Compare Pp and Ppk. If Pp and Ppk are approximately equal, then the process is centered between
the specification limits. If Pp and Ppk differ, then the process is not centered.
Caution
Because the Pp index does not consider the process location, it does not indicate how close the
process is to the target region that is defined by the specification limits. For example, the following
graphs show two processes with the same Pp value, yet one process falls within the specification
limits and the other does not.

Pp = 2.27

Pp = 2.27

For a complete and accurate analysis, use graphs in combination with other capability indices (such
as Ppk) to draw meaningful conclusions from your data.

10
2.2 Ppk
Ppk is a measure of the overall capability of the process and equals the minimum of PPU and
PPL. Ppk is a ratio that compares two values:
The distance from the process mean to the closest specification limit (USL or LSL). The one-sided
spread of the process (the 3-σ variation) based on its overall variation Ppk evaluates both the location
and the overall variation of the process.

Interpretation
Use Ppk to evaluate the overall capability of your process based on both the process location and
the process spread. Overall capability indicates the actual performance of your process that your
customer experiences over time.
Generally, higher Ppk values indicate a more capable process. Lower Ppk values indicate that
your process may need improvement.

Low Ppk
In this example, the distance from the process mean to the nearest specification limit (USL) is
less than the one-sided process spread. Therefore, Ppk is low (0.66), and the overall capability of
the process is poor.

High Ppk
In this example, the distance from the process mean to the nearest specification limit (LSL) is
greater than the one-sided process spread. Therefore, Ppk is high (1.68), and the overall
capability of the process is good.

11
You can compare Ppk to other values to get more information about the capability of your process.
Compare Ppk to a benchmark value that represents the minimum value that is acceptable for your
process. Many industries use a benchmark value of 1.33. If Ppk is lower than your benchmark,
consider ways to improve your process.
Compare Pp and Ppk. If Pp and Ppk are approximately equal, then the process is centered between
the specification limits. If Pp and Ppk differ, then the process is not centered.
Compare Ppk and Cpk. When a process is in statistical control, Ppk and Cpk are approximately
equal. The difference between Ppk and Cpk represents the improvement in process capability that
you could expect if shifts and drifts in the process were eliminated.

Caution
The Ppk index represents only one side of the process curve and does not measure how the process
performs on the other side of the process curve.
For example, the following graphs display two processes that have identical Ppk values. However,
one process violates both specification limits, and the other process violates only the lower
specification limit.

Ppk = min {PPL = 4.01, PPU = 0.64} = 0.64

Ppk = PPL = PPU = 0.64

If your process has nonconforming parts that fall on both sides of the specification limits, consider
using other indices, such as Z.bench, to more fully assess process capability.

12
2.3 Cpm
Cpm is a measure of the overall capability of the process. Cpm compares the specification spread to
the spread of your process data while taking into account how much the data deviate from the target
value.
You must provide a target value for Minitab to calculate the Cpm.

Interpretation
Use Cpm to evaluate the overall capability of your process relative to both the specification spread
and the target. Overall capability indicates the actual performance of your process that your customer
experiences over time.
Generally, higher Cpm values indicate that your process is more capable. Lower values indicate that
your process may need improvement.

High Cpm
In this example, the data are on target and fall within the specification limits. Therefore, Cpm is high (1.60).

Low Cpm
In this example, the data fall inside of the specification limits, but the process is off target. Therefore, Cpm is low
(1.03).

13
Low Cpm
In this example, the process is on target, but all of the data do not fall within the specification limits. Therefore, Cpm
is low (0.48).

You can compare Cpm to other values to get more information about the capability of your process.
Compare Cpm with a benchmark value to assess the overall capability of your process. Many
industries use a benchmark value of 1.33. If Cpm is lower than your benchmark, consider ways to
improve your process, such as reducing its variation or shifting its location.
Compare Ppk and Cpm. If the process is centered on target, the Ppk and Cpm values are
approximately equal.
2.4 PPM Total for Expected Overall Performance
PPM Total for expected overall performance is the expected number of parts per million that are
outside the specification limits. Expected overall performance values are calculated using overall
standard deviation. PPM Total for expected overall performance is 1,000,000 times the probability
that the measurement of a randomly selected part from the overall process distribution is outside the

specification limits. The probability that a randomly

selected part is outside the specification limits is shown by the shaded regions under the overall
normal curve.

Interpretation
Use PPM Total for expected overall performance to estimate the number of nonconforming items,
represented in parts per million, that you can expect to be outside the specification limits based on
the overall variation of your process. Overall performance values indicate the actual process
performance that your customer experiences over time.

14
Lower values of PPM Total indicate greater process capability. Ideally, few or no parts have
measurements that are outside the specification limits.
You can also use PPM to estimate the percentage of conforming and nonconforming parts in your
process.

% Nonconforming
PPM Parts % Conforming

66807 6.6807% 93.3193%

6210 0.621% 99.379%

233 0.0233% 99.9767%

3.4 0.00034% 99.99966%

2.5 Z.Bench for Overall Capability

Z.bench (overall) is the percentile on a standard normal distribution that translates the estimated
probability of defects in the process to an upper tail probability. It is calculated based on overall
process performance, using the overall standard deviation.

The defects for the process fall on both sides of the specification limits. The overall standard deviations are shown
by the tick marks.

If you put all the defects on the right tail of the distribution, and then measure the number of overall standard
deviations from the center (vertical line) to the point that defines the total defects, you obtain the Z-bench (overall)
value.

NOTE

To display Z.bench measures, you must click Options and change the default output from capability
statistics to benchmark Z's when you perform the capability analysis.

15
Interpretation
Use Z.Bench (overall) to evaluate the overall sigma capability of your process.
Generally, higher Z.bench values indicate a more capable process. Lower Z.bench values indicate
that your process may need improvement. When possible, compare Z.bench with a benchmark value
based on process knowledge or industry standards. If Z.Bench is lower than your benchmark,
consider ways to improve your process.
Compare Z.Bench (within) and Z.Bench (overall). When a process is in statistical control, Z.Bench
(within) and Z.Bench (overall) are approximately equal. The difference between the two values
represents the improvement in process capability that you could expect if the process were brought
into control. Zbench (overall) is sometimes referred to as Z.Bench Long-Term (LT).
2.6 Cp (Between/Within)
Cp is a measure of the between/within capability of the process. Cp is a ratio that compares two
values:
The specification spread (USL– LSL)
The spread of the process (the 6-σ variation) based on the between/within standard deviation
Cp evaluates capability based on the variation in your process but not on its location.

Interpretation
Use Cp to evaluate the between/within capability of your process. Between/within capability
indicates the capability that your process could achieve if systemic sources of overall process
variation, besides the variation within and between the subgroups, were eliminated.
Because Cp does not consider the location of the process, it indicates the capability that your process
could achieve if it were centered. Generally, higher Cp values indicate a more capable process.
Lower values indicate that your process may need improvement.

Low Cp

16
In this example, the specification spread is less than the overall process spread. Therefore, Cp is low (0.40) and the
capability of the process is poor based on its variability.

High Cp
In this example, the specification spread is considerably greater than the overall process spread. Therefore, Cp is
high (1.80) and the overall capability of the process is good based on its variability.

You can compare Cp to other values to get more information about the capability of your process.
Compare Cp to a benchmark value to assess the potential capability of your process. Many industries
use a benchmark value of 1.33. If Cp is lower than your benchmark, consider how to improve your
process by reducing its variation.
Compare Cp and Pp. If Pp is considerably greater than Cp, there may be other sources of systemic
variation in the process, in addition to the variation between and within the subgroups.
Compare Cp and Cpk. If Cp and Cpk are approximately equal, then the process is centered between
the specification limits. If Cp and Cpk differ, then the process is not centered.

Caution
Because the Cp index does not consider the process location, it does not indicate how close the
process is to the target region that is defined by the specification limits. For example, the following
graphs show two processes with the same Cp value, yet one process falls within the specification
limits and the other does not.

17
Cp = 3.13

Cp = 3.13

For a complete and accurate analysis, use graphs in combination with other capability indices (such
as Cpk) to draw meaningful conclusions from your data.
2.7 Cpk (Between/Within)
Cpk is a measure of the between/within capability of the process and equals the minimum of CPU
and CPL. Cpk is a ratio that compares two values:
The distance from the process mean to the closest specification limit (USL or LSL)
The one-sided spread (the 3-σ variation) of the process based on between/within subgroup standard
deviation
Cpk valuates both the location and the variation (between and within subgroups)of the process.

Interpretation
Use Cpk to evaluate the between/within capability, considering both its location and spread.
Between/within capability indicates the capability that your process could achieve if systemic
sources of overall process variation, besides the variation within and between the subgroups, were
eliminated.
Generally, higher Cpk values indicate a more capable process. Lower Cpk values indicate that your
process may need improvement.

Low Cpk

18
In this example, the distance from the process mean to the nearest specification limit is less than the one-sided
process spread. Therefore, Cpk is low (0.66) and the between/within capability of the process is poor.

High Cpk
In this example, the distance from the process mean to the nearest specification limit is greater than the one-sided
process spread. Therefore, Cpk is high (1.68) and the between/within capability of the process is good.

You can compare Cpk to other values to get more information about the capability of your process.
Compare Cpk with a benchmark that represents the minimum value that is acceptable for your
process. Many industries use a benchmark value of 1.33. If Cpk is lower than your benchmark,
consider ways to improve your process, such as reducing its variation or shifting its location.
Compare Cpk and Ppk. If Cpk is considerably greater than Ppk, there may be other sources of
systemic variation in the process, in addition to the variation between and within the subgroups. For
example, tool wear is a systemic source of variation that could cause a manufacturing process to
exhibit greater between/within capability (Cpk) than overall capability (Ppk).
Compare Cp and Cpk. If Cp and Cpk are approximately equal, then the process is centered between
the specification limits. If Cp and Cpk differ, then the process is not centered.

Caution
The Cpk index represents only one side of the process curve, and does not measure how the process
performs on the other side of the process curve.
For example, the following graphs display two processes with identical Cpk values. However, one
process violates both specification limits, and the other process only violates the lower specification
limit.

19
Cpk = min {CPL = 4.58, CPU = 0.93} = 0.93

Cpk = CPL = CPU = 0.93

If your process has nonconforming parts that fall on both sides of the specification limits, consider
using other indices, such as Z.bench, to more fully assess process capability.
2.8 PPM Total for Expected Between/Within Performance
PPM Total for expected between/within performance is the expected number of parts per million that
have measurements that are outside the specification limits. Expected between/within performance
values are calculated using the between/within standard deviation. PPM Total for expected
between/within performance is 1,000,000 times the probability that the measurement of a randomly
selected part from the within-subgroup process distribution is outside the specification limits.

The probability that a randomly selected part is outside the specification limits is shown by the shaded regions under
the between/within normal curve.

Interpretation
Use PPM Total for expected between/within performance to estimate the number of nonconforming
items, represented in parts per million, that you can expect to be outside the specification limits based
on the variation between/within the subgroups. Between/within performance values indicate the
performance that could be achieved if additional sources of systemic process variation, besides the
variation between and within subgroups, could be eliminated.
Lower values of PPM Total indicate greater process capability. Ideally, few or no parts have
measurements that are outside the specification limits.
You can also use PPM to estimate the percentage of conforming and nonconforming parts in your
process.

20
 % Nonconforming
PPM Parts % Conforming

66807 6.6807% 93.3193%

6210 0.621% 99.379%

233 0.0233% 99.9767%

3.4 0.00034% 99.99966%

2.9 Z.Bench for Between/Within Capability

Z.bench (between/within) is the percentile on a standard normal distribution that translates the
estimated probability of defects in the process to an upper tail probability. It is calculated based on
the between/within process performance, using the between/within standard deviation.

The defects for the process fall on either side of the specification limits. The between/within subgroup standard
deviations are shown by the tick marks.

If you put all the defects on the right tail of the distribution, and then measure the number of between/within-
subgroup standard deviations from the center (red line) to the point that defines the total defects, you obtain the Z-
bench (between/within) value.

NOTE

To display Z.bench, you must click Options and change the default output from capability statistics
to benchmark Z's when you perform the capability analysis.

Interpretation
Use Z.Bench (between/within) to evaluate the between/within sigma capability of your process.
Generally, higher Z.bench (between/within) values indicate that the process is more capable. Lower
values indicate that your process may need improvement. When possible, compare Z.bench
21
(between/within) with a benchmark value based on process knowledge or industry standards. If
Z.bench (between/within) is less than your benchmark, consider ways to improve your process.
2.10 Practice
YOU SHOULD CHECK WHETHER YOUR DATA IS NORMAL OR NOT BEFORE
GOING TO THE NEXT STEP.
2.10.1 Example of Between/Within Capability Analysis
An engineer wants to assess the capability of a process that coats large rolls of paper with a thin
film. The engineer collects three measurements of coating thickness from 25 consecutive rolls.
Because the machine is reset for each new roll, the engineer needs to consider variation between
rolls in addition to variation within a roll. The film thickness must be 50 ± 3 microns to meet
engineering specifications.
The engineer performs between/within capability analysis to evaluate how well the coating
thickness meets the customer requirements of 50 ± 3 microns.
 Open the sample data, FilmThickness.MTW.
 Choose Stat > Quality Tools > Capability Analysis > Between/Within.
 In Single column, enter Coating.
 In Subgroup size, enter Roll.
 In Lower spec, enter 47.
 In Upper spec, enter 53.
 Click Options.
 In Target (adds Cpm to table), enter 50.
 Click OK in each dialog box.
2.10.2 Example of Normal Capability Analysis for Multiple Variables
A manufacturer uses two machines to produce connector pins. A quality engineer wants to
compare the process capability of each machine. The engineer randomly samples five pins per
box from each machine. The specification range for the length of the pin is 13 mm to 25 mm.
The engineer performs normal capability analysis for multiple variables to evaluate how well the
lengths of the connector pins meet customer requirements for the two machines.
 Open the sample data, PinLength.MTW.
 Choose Stat > Quality Tools > Capability Analysis > Multiple Variables (Normal).
 In Variables, enter Length.
 In Subgroup sizes, enter 5.
 Select By variables and enter Machine.
 In Lower spec, enter 13.
 In Upper spec, enter 25.
 Click OK.

22
2.10.3 Example of Between/Within Capability Sixpack
An engineer wants to assess the capability of a process that coats large rolls of paper with a
thin film. The engineer collects three measurements of coating thickness from 25 consecutive
rolls. Because the machine is reset for each new roll, the engineer needs to consider variation
between rolls in addition to variation within a roll. The film thickness must be 50 ± 3 microns
to meet engineering specifications.
The engineer performs between/within capability sixpack to check the assumptions for
between/within capability analysis and to evaluate how well the coating thickness meets
customer requirements.
 Open the sample data, FilmThickness.MTW.
 Choose Stat > Quality Tools > Capability Sixpack > Between/Within.
 In Single column, enter Coating.
 In Subgroup size, enter Roll.
 In Lower spec, enter 47.
 In Upper spec, enter 53.
 Click Options.
 In Target (adds Cpm to table), enter 50.
 Click OK in each dialog box.
2.10.4 Example of Nonnormal Capability Analysis
Quality engineers at a company that manufactures floor tiles investigate customer
complaints about warping in the tiles. To ensure production quality, the engineers
measure warping in 10 tiles each working day for 10 days. The upper specification limit
for the warping measurement is 6 mm. The engineers know that the distribution of
warping follows the Weibull distribution.
The engineers perform nonnormal capability analysis to evaluate how well the tile
manufacturing process meets customer requirements.
 Open the sample data, TileWarping.MTW.
 Choose Stat > Quality Tools > Capability Analysis > Nonnormal.
 In Single column, enter Warping.
 From Fit distribution, select Weibull.
 In Upper spec, enter 6.
 Click OK.
2.10.5 Example of Nonnormal Capability Analysis for Multiple Variables
A manufacturer uses two machines to make plastic pellets. A quality engineer wants to assess the
process capability of each machine. The engineer randomly samples 50 pellets from each
machine and records the weight of each pellet. The specification range is 12 g to 25 g. The
engineer knows that the weight of the pellets follows a Weibull distribution.

23
The engineer performs nonnormal capability analysis for multiple variables to evaluate how well
the weight of the pellets meets customer requirements for the two machines.
 Open the sample data, PelletWeight.MTW.
 Choose Stat > Quality Tools > Capability Analysis > Multiple Variables (Nonnormal).
 In Variables, enter Weight.
 Select By variables and enter Machine.
 From Fit distribution, select Weibull.
 In Lower spec, enter 12.
 In Upper spec, enter 25.
 Click OK.
2.10.6 Example of Nonnormal Capability Sixpack
Quality engineers at a company that manufactures floor tiles investigate customer complaints
about warping in the tiles. To ensure production quality, the engineers measure warping in 10
tiles each working day for 10 days. The upper specification limit for the warping
measurement is 6 mm. The engineers know that the distribution of warping follows the
Weibull distribution.
The engineers perform nonnormal capability sixpack to check the assumptions for nonnormal
capability analysis and to evaluate how well the tile manufacturing process meets customer
requirements.
 Open the sample data, TileWarping.MTW.
 Choose Stat > Quality Tools > Capability Sixpack > Nonnormal.
 In Single column, enter Warping.
 In Subgroup size, enter 10.
 From Fit distribution, select Weibull.
 In Upper spec, enter 8.
 Click OK.
2.10.7 Cpk/ ppk - what is the difference?

24