LEAN SIX SIGMA

ANALYZE PHASE
“If you torture the data long enough, it will confess.” - Ronald H. Coase

1
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 COURSE CONTENT

Coverage:
• ANALYZE PHASE TOOLS
 7 QC Tools (Histogram, Cause-and-Effect Diagrams / Fishbone
Diagram, Pareto charts, Control Charts, Scatter diagrams,
Stratification, Check Sheet) and PFMEA
 5 Why Analysis, Tree Diagram, Affinity clustering
 Data visualization - Box Plot , Multi Vari charts Trends and
comparison charts
 Hypothesis testing - Parametric test (Continuous Data , Discrete
Data) Alpha & Beta errors

2
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

3
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Analyze Phase
The Analyze phase of DMAIC project helps PROCESS DATA
project teams to identify “Vital X” form the DOOR DOOR
list of “Potential Xs”.

In this phase project teams use a variety of

Process door and Data door analysis tools
to validate the root cause of the problem

4
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools

5
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Process Flow

Process Flow is a graphical representation of a business process through a flowchart. It's
used as a means of getting a top-down understanding of how a process works, what steps it
consists of, what events change outcomes, and so on

6
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Process Flow

Stage 1
•Request for a documented process (SOP) form the process owner
•Study the documented process
•Create a process flow based on your understanding
•Check with the process owner if your process flow is inline with the document process
Stage 2
• Spend a “Day in the Life of (DILO)” the processor and work-shadow the processor
•Pick up random transaction and travel to different departments and track how the
transaction is processed
Stage 3
•Highlight the deviations observed in the shop floor and the documented process to the
process owner
Stage 4
• Request the process owner to confirm the final version with written approval
Stage 5
• Document the final version of process flow in project PPT

7
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Check Sheet

Check sheet one of the 7QC tools it is a form or data collection template used to collect
data in real time at the location where the data is generated. The data it captures can be
quantitative or qualitative

8
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Cause & Effect Diagram

Kaoru Ishikawa
A Cause and Effect Diagram is a graphical tool for displaying a list of causes associated with a
specific effect. It is also known as a fishbone diagram or an Ishikawa diagram (created by Dr.
Kaoru Ishikawa, an influential quality management innovator). The graph organizes a list of
potential causes into categories.

9
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Cause & Effect Diagram

What is Cause And What is effect ?
Effect is the problem or the outcome and case is the principle reason for the outcome
Sequence of events
Brainstorming – Clustering – Cause & Effect Diagram

Brainstorming Clustering Fishbone diagram

10
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Cause & Effect Diagram - Brainstorming

Brainstorming
To brainstorm is to think about and try to come up with Potential
cause to a problem or ideas or solutions to a problem.
Brainstorming combines a relaxed, informal approach to problem
solving with lateral thinking. It encourages people to come up
with thoughts and ideas that can, at first, seem a bit crazy
Types of Brainstorming Dos and Don’ts of brainstorming
•Analogy •Ambience & atmosphere during discussion Brainstorming
•Brain Writing must be perfect
•Try Storming •Focus on the theme.
•Anti-Storming •Give everybody the opportunity to speak.
•Step ladder approach •No criticism of ideas as they are expressed.
•Questioning •Do not get into action with only one idea.
Orthodoxy •Do not get into details.
•Do not evaluate an idea.
•Do Not Paraphrase

11
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Cause & Effect Diagram –clustering

Clustering is a method which can help you gather large amounts of data and organize them into
groups or themes based on their relationships. The affinity process is great for grouping data
gathered during research or ideas generated during Brainstorming

Brainstorming Outcome Clustering

12
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Cause & Effect Diagram

13
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Histogram
Histogram is a graphical representation of the
distribution of numerical data

Purpose : The shape, mean & spread of the

histogram gives additional information about
the process and guides the managers to take
the right decision to improve the process.

How to construct histogram manually

Step 1 : Calculate the range (Maximum Value – Class Width=2
Minimum Value) Classes Frequency Total
Step 2 : Create class intervals (about 8 or 10) by 34-35 8
determining appropriate class intervals 36-37 25
Step 3 : Create frequency distribution table 38-39 31
Step 4 : Draw histogram based on the frequency 40-41 35
42-43 20
distribution table
44-45 5
46-47 1

14
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Pareto chart

Vilfredo Pareto was an Italian Economist in the 19th Century In 1906, he
made the famous observation that twenty percent of the population owned
eighty percent of the property in Italy.

This is famously referred as “80-20 rule” or “Vital few trivial many” and
graphical represented as “Pareto Chart”

The tool was popularized by Joseph M Juran and Kaoru Ishikawa. Vilfredo Pareto
How to construct Pareto chart
Step 1 : Root cause identified during fishbone diagram can be converted as defect
categories
Step 2 : Collect the data to identify the frequency of each defect category
Step 3 : Calculate the contribution of each category to the total number of defects Minitab exercise
Step 4 : Calculate Cumulative Contribution
Step 5 : Plot the number of defect counts in primary access (As bar-graph)
Step 6 : Plot the cumulative contribution in Secondary access (As Line-graph)
Excel Analysis

15
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Stratification
Stratification is to classify or group data with matching characteristics in groups or strata. It
serves to facilitate the work before using other tools such as histograms or scatter diagrams

Joseph M Juran in his famous book “Quality Hand Book”

Stratification is the separation of data into categories. It is used to
identify which categories contribute to the problem being solved
and which categories are worthy of further investigation

Joseph M. Juran

16
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Stratification - Subdivisions

Level 1 Level 2 Level 3

Multiple drill down of the same data

17
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Stratification - Multi-perspective analysis

Same return data analyzed in count and $value

18
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Stratification - Repeat analysis

Return data analyzed for two consecutive timeperiod

19
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Control Chart

Statistical process control (SPC) is a method of quality control which
employs statistical methods to monitor and control a process. Control
Chart is one of the important SPC tool

Control Chart also known as Shewhart charts or process-behavior

charts, are a statistical process control tool used to distinguish
between common cause variation and special cause variation Walter A. Shewhart
Special Cause Common Cause

UCL
_
Plan
X

LCL

Do Check and Act Special Cause

20
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

7 QC Tools – Scatter Diagram

Correlation Correlation is a statistical technique that can show whether and how strongly
pairs of variables are related. The strength of correlation is statistically represented
“correlation coefficient” and graphically represented by “Scatter Diagram”

Correlation Coefficient level of significance

•0.8 to 1 (Strong Positive correlation)
•-0.8 to -1 (Strong Negative correlation)
• Between 0.7 to 0.5 and -0.7 to -0.5 (week correlation)
• Between 0.5 t0 -0.5 (No correlation)

21
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Failure Mode Effects Analysis - FMEA

•FMEA is a structured approach to Identifying the areas and ways in which a
process or system can fail (failure mode)
•Estimating risk associated with specific causes
•Identifying and prioritizing the actions that should be taken to reduce those risks
•Evaluating and documenting proposed process plans or current control plans
•FMEA is used to reduce risk, and therefore unintended consequences, in the
implementation.
PFMEA Process FMEA : A Process Failure Mode Effects Analysis (PFMEA) is a
structured analytical tool used by an organization, business unit, or cross-
functional team to identify and evaluate the potential failures of a process

22
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Failure Mode Effects Analysis - FMEA

PFMEA Process FMEA : A Process Failure Mode Effects Analysis (PFMEA) is a
structured analytical tool used by an organization, business unit, or cross-
functional team to identify and evaluate the potential failures of a process

23
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Failure Mode Effects Analysis - FMEA

•Identify all stake holders

•Map the process (Material flow as well as
information flow) at a most granular level

24
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Failure Mode Effects Analysis - FMEA

•For each process step determine the ways in which the input to the step or the
function of the step can go wrong (failure mode – People , Process , Technology ).
•For each failure mode, determine effects - Select a severity rating for each effect
(1-10 scale)
•Identify potential causes of each failure mode - Select an occurrence rating for
each cause (1-10 scale)
•List current monitoring and controls for each cause -Select a detection rating for
each cause (1-10 scale)
•Multiple these three ratings to calculate Risk Priority Number (S * O * D = RPN)
•Identify the process steps with high RPN (>100) and take required actions to bring
down the RPN

25
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Failure Mode Effects Analysis - FMEA

•FMEA should be a living document , Keep updating the actions taken and
subsequent reduction in RPA

Assume that the Severity number Item Severity Occurrence Detection RPN
cannot be reduced. Indicate the a 8 10 2 160
order of importance that you b 10 8 2 160
would assign as far as addressing
c 8 2 10 160
these processes so as to reduce
overall risk. d 10 2 8 160

26
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

5 why Analysis
The 5 Whys technique is a
simple and effective tool for
solving problems. Its primary
goal is to find the exact reason
that causes a given problem by
asking a sequence of “Why”
questions.

27
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

5 why Analysis – What’s wrong with this

28
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

5 why Analysis – How it should be done

29
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Tree Diagram – How it should be done

Another way to find structure in potential causes is to use a tree diagram, which is a tool
used to arrange related ideas in sequence from broad and general to narrow and specific

30
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Tree Diagram – How it should be done

31
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Data visualization - Box Plot

A box plot summarizes information about the shape, dispersion and centering; it also helps
spot outliers in the data set

•BOX – represents the middle 50% values of the data set

•MEDIAN – represents the point for which 50% of the data points
are above and 50% of the data points are below in the line
•Q1 and Q3 – Q1 represents the point for which 25% of the data
points are below and 75% of the data points are above in the
line; Q3 represents 75% of the data points are below and 25% of
the data points are above in the line
•AESTRIX – represents an outlier and is a point which does not
belong to the family of Xs in the data; beyond Q3+1.5(Q3-Q1) or
Q1-1.5(Q3-Q1)
•LINES – These vertical lines represent a whisker which joins Q1
or Q3 with the farthest data point but other than the outlier ;
Maximum Value = Q3+1.5(Q3-Q1) ; Minimum Value = Q1-1.5(Q3-
Q1)

32
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Data visualization – Multi Vari charts

Multi-Vari Chart graphically displays patterns of variation. It is used to identify possible Xs or
families of variation, such as variation within a subgroup, between subgroups, or over time

33
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
A hypothesis test evaluates two mutually exclusive statements about a population
to determine which statement is best supported by the sample data. Hypothesis
testing is the process of using statistics to determine the probability that a specific
hypothesis is true or false.

To prove that a hypothesis is true, or false, with absolute certainty, we would need
absolute knowledge. That is, we would have to examine the entire population.
Instead, hypothesis testing concerns on how to use a random sample to judge if it
is evidence that supports or not the hypothesis.

This text book definition may sound a bit complex ….

Lets understand this with a help of a real life example

34
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Scenario 1
Operations Manager of Tier 1 automobile component
manufacturing company makes a statement in weekly
Production Planning and Control meeting that
“ Morning shift take less time to produce one unit compared
to the night shift… hence morning shift is more productive
compared to night shift”

What operations Manager trying to prove is called “Real Life Hypothesis” in this
case, Morning shift is performing better than the night shift
If we try and convert this “Real Life Hypothesis” in to “Statistical Hypothesis” we
will rephrase as “The Average handling time of morning shift is lesser than the
average handling time of night shift”

35
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Now we have perfect opportunity to perform a
hypothesis testing
There is no difference… every thing is good and no problem is called Null
Hypothesis (H o)
“The Average handling time of morning shift is same as the average handling time
of night shift”

What we want to prove is called Alternate Hypothesis (H a)

“The Average handling time of morning shift is lesser than the average handling
time of night shift”
Test result should either tell us Ho is true or Ha is true

36
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Step1 : Understand the scenario
Step2 : Frame “real life Hypothesis”
Step3 : Convert “real life Hypothesis” in to “statistical Hypothesis”
Step4 : This “statistical Hypothesis” is called “Alternate Hypothesis” (Ha)
Step5 : Based on “Alternate Hypothesis” (Ha) frame “Null Hypothesis” (Ho)
Step6 : Collect sample
Step7 : Perform test statistics
Step8 : Based on the test result either go with Ho or Ha
Step9 : Convert test result in to Business context
Going Back in time

In a courtroom, the person is assumed innocent until proven guilty. In a hypothesis test, we
assume the null hypothesis is true until the data proves otherwise

It is only later, in 1933, that “hypothesis tests” were invented by two statisticians: Polish Jerzy
Neyman and British Egon Pearson
37
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Now we have decided to test this hypothesis… Operations manager (OM)
requested the Team Lead (TL) of both the shift to collect AHT of 60 sample
each from their respective teams
After few days both the TLs reverted back to the OM with 60 samples of AHT
and the corresponding sample averages are
Morning shift : 17 Min 42 Sec Most of us would have made your
Night Shift : 20 Min 35 Sec decision by looking at these values
but failed to notice
Question to audience is … Will you go with Ho or Ha
Null Hypothesis (H o) : “The Average handling time of morning shift is same as the
average handling time of night shift”
Alt Hypothesis (H a) : “The Average handling time of morning shift is lesser than the
average handling time of night shift”

38
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
What is the relevance of “Sample average” in previous slide?
Understanding the DATA under study is very important in hypothesis testing

Continuous Count
Population
Binary
Discrete
Sample
Ordinal

Primary DATA Normal Nominal

Secondary Non Normal

Measured Derived

39
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing Considering this area of

Let us relook at the scenario 1 uncertainty (overlapping CIs) we
Morning shift : 17Min 42 Sec will go with Null hypothesis Ho
Night Shift : 20 Min 35 Sec
18.35 20.35 22.35
( )
ES
x̅
15.42 17.42 19.42
(
MS
x̅ )

14 15 16 17 18 19 20 21 22 23

Question to audience is … Will go with Ho or Ha

Null Hypothesis (H o) : “The Average handling time of morning shift is same as the
average handling time of night shift”
Alternate Hypothesis (H a) : “The Average handling time of morning shift is lesser
than the average handling time of night shift”
40
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Now we understand confidence interval plays a very important role deciding the out
come of Hypothesis testing
Based on these formulas we can make the
following statements

If we take less samples Confidence interval

will be wider
If we take more samples Confidence interval
will be Narrow

So it becomes very clear that we need to

identify the accurate sample size and
perform hypothesis testing

41
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
What can go wrong while doing Hypothesis testing

Truth
Ho Ha

Correct
Type II
Ho Decision
Error
1– =
Confidence
Accept
1 - = Probability of
Correct correctly accepting
Type I Decision the alternative
Ha Error 1 – = Power
hypothesis
(detecting a change
when there is one).

1 - = Probability of correctly accepting the null

hypothesis (detecting no change when there is none).

42
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing
Identifying the correct Hypothesis test based on scenario
What
Discrete Continuous
type of
Data Data
data

Statistics of interest
Data
: proportion
distribution

1 Proportion Test Statistics of

interest

2 Proportion Test
Variance or Mean or
Chi square Test standard deviation Average Median

One sample
One variance t-test , z-test One sample Sign test
Test Two sample t-
test, paired T Mann - Whitney test
Test for Two or
more variance One way
Moods Median test
ANOVA

43
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE
Pre Requisites
Hypothesis testing - Statistics of interest : Variance Data Type Continuous
Test for two or more variance Distribution Normal
Determine whether the variance or the standard deviation of two or more groups
are different
Business Scenario: A vegetable oil company procure a “Chemical -x” one of the
important ingredients from two different suppliers. As Quality manager you are
tasked to find if the pH value of the chemical of both the suppliers have same level
of variation

Minitab exercise
Refer : Test for Two variance

44
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Statistics of interest : Variance Pre Requisites

Data Type Continuous
Test for two or more variance Distribution Normal
Determine whether the variance of two or more groups are different
Null hypothesis: All variances are equal
Alternative hypothesis: At least one variance is different

95% Bonferroni Confidence Intervals for Standard Deviations

Sample N StDev CI
PH Value from Suppler 1 100 0.196350 (0.169943, 0.232061)
PH Value from Suppler 2 100 0.467028 (0.392291, 0.568752)
CI clearly
don’t over C
Individual confidence level = 97.5%
lap
Tests
Test
Method Statistic P-Value
Multiple comparisons 34.49 0.000
Levene 37.08 0.000
P-Value < 0.05 go with Alternative hypothesis
P-Value > 0.05 go with Null hypothesis

45
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Statistics of interest : Variance Pre Requisites

Data Type Continuous
Test for two or more variance Distribution Normal
Determine whether the variance of two or more groups are different
Null hypothesis: All variances are equal
Alternative hypothesis: At least one variance is different

95% Bonferroni Confidence Intervals for Standard Deviations

Sample N StDev CI
PH Value from Suppler 2 100 0.467028 (0.392291, 0.568752)
PH Value from Suppler 3 100 0.422790 (0.372456, 0.490930)
Individual confidence level = 97.5%
Tests CI clearly
Test over lap
Method Statistic P-Value
Multiple comparisons 0.91 0.341
Levene 0.07 0.796
P-Value < 0.05 go with Alternative hypothesis
P-Value > 0.05 go with Null hypothesis

46
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE
Pre Requisites
Hypothesis testing - Statistics of interest : Variance Data Type Continuous
One variance test Distribution Normal
Determine whether the variance or the standard deviation of a sample is different
from a specified value
Business Scenario: An Automobile manufacturing company procure paint from
leading MNC paint company. As per the contract the acceptable level of Delta – E
variation is 0.04 with a mean of 1.2, As procurement manager you are tasked to
check if the variation is with in the 0.04 level as per contract

Minitab exercise
Refer : One variance test

47
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE
Pre Requisites
Hypothesis testing - Statistics of interest : Variance Data Type Continuous
One variance test Distribution Normal
Determine whether the variance or the standard deviation of a sample is different
from a specified value
Null hypothesis σ = 0.04
Alternative hypothesis σ > 0.04

The chi-square method is only for the normal distribution. The Bonett method is for any continuous distribution.

Statistics
Variable N StDev Variance
Delta_E 60 0.0923 0.00851
95% Confidence Intervals
Variable Method CI for StDev CI for Variance
Delta_E Chi-Square (0.0782, 0.1125) (0.00612, 0.01266)
Bonett (0.0803, 0.1096) (0.00645, 0.01200)
Tests
Test
Variable Method Statistic DF P-Value
Delta_E Chi-Square 200.92 59 0.000 P-Value < 0.05 go with Alternative hypothesis
Bonett — — 0.000 P-Value > 0.05 go with Null hypothesis

48
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE
Pre Requisites
Hypothesis testing - Statistics of interest : Variance Data Type Continuous
One variance test Distribution Normal
Determine whether the variance or the standard deviation of a sample is different
from a specified value

49
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
1 Sample Z-test Equal Variance (More than 1 sample data set)
Determine whether the mean of a sample differs significantly from a specified
value when the population standard deviation is known
Business Scenario: As training manager you are asked check if the average call
handling time (AHT) of two training teams team a and team b is not greater than 12
min (Current AHT of the operations tem). from operations team we also know the
standard deviation of the populations is 1min

Minitab exercise
Refer : 1 Sample Z-test

50
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
1 Sample Z-test Equal Variance (More than 1 sample data set)

Null hypothesis μ of Team A = 12 Min

Alternative hypothesis μ of Team A > 12 Min

Variable N Mean StDev SE Mean 95% CI Z P

AHT - Team A 100 15.021 1.170 0.100 (14.825, 15.217) 30.21 0.000

P-Value < 0.05 go with Alternative hypothesis

P-Value > 0.05 go with Null hypothesis

51
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
1 Sample Z-test Equal Variance (More than 1 sample data set)

Null hypothesis μ of Team B = 12 Min

Alternative hypothesis μ of Team B > 12 Min

Variable N Mean StDev SE Mean 95% CI Z P

AHT - Team B 100 11.970 1.017 0.100 (11.774, 12.166) -0.30 0.767

P-Value < 0.05 go with Alternative hypothesis

P-Value > 0.05 go with Null hypothesis

52
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
1 Sample t-test Equal Variance (More than 1 sample data set)
Determine whether the mean of a sample differs significantly from a specified value.
Business Scenario: A Semiconductor making company decided to evaluate a new
supplier for Mercury . As a R&D team member you are tasked to check the average
viscosity of the liquid metal supplied in different samples meets the requirement of
1.526 cP (centipoise)

Minitab exercise
Refer : 1 Sample Z-test

53
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
1 Sample t-test Equal Variance (More than 1 sample data set)

Null hypothesis μ velocity of Mercury supplied by new supplier = 1.526 cP

Alternative hypothesis μ velocity of Mercury supplied by new supplier ≠ 1.526 cP

Variable N Mean StDev SE Mean 95% CI T P

Viscosity Of mercury 30 1.52751 0.01072 0.00196 1.52351, 1.53151) 0.77 0.446

P-Value < 0.05 go with Alternative hypothesis

P-Value > 0.05 go with Null hypothesis

54
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
2 Sample t-test Equal Variance (More than 1 sample data set)
Determine whether the mean differs significantly between two groups
Business Scenario: Operations manager of a contact centre need to identify between
team managed by Team leader Mike and Team managed by Team leader David which
team have relatively less AHT

Minitab exercise
Refer : 2 Sample T-test

55
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous

2 Sample t-test Distribution Normal

Equal Variance (More than 1 sample data set)

Two-Sample T-Test and CI: Team - Mike, Team - David

Two-sample T for Team - Mike vs Team – David

Null hypothesis μ AHT of Team - Mike = μ AHT of Team - David

Alternative hypothesis μ AHT of Team - Mike ≠ μ AHT of Team - David

N Mean StDev SE Mean

Team - Mike 100 15.02 1.17 0.12
Team - David 100 11.97 1.02 0.10

Difference = μ (Team - Mike) - μ (Team - David)

Estimate for difference: 3.051
95% CI for difference: (2.745, 3.357)
T-Test of difference = 0 (vs ≠):

T-Value = 19.68 P-Value = 0.000 DF = 194

P-Value < 0.05 go with Alternative hypothesis
P-Value > 0.05 go with Null hypothesis
56
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
2 Sample t-test Equal Variance (More than 1 sample data set)

57
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
Paired t-test Equal Variance (More than 1 sample data set)
Used to compare measurements that are made on the same item under different
conditions
Business Scenario: An F1 teams R&D department want to check if response time of
moving from 0-100 km differ based on Air or Nitrogen in the tyres

Minitab exercise
Refer : Paired t-test

58
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
Paired t-test Equal Variance (More than 1 sample data set)
Paired T for Air - Nitrogen

Null hypothesis μ Response time of Air = μ Response time of Nitrogen

Alternative hypothesis μ Response time of Air ≠ μ Response time of Nitrogen

N Mean StDev SE Mean

Air 60 11.9693 0.7450 0.0962
Nitrogen 60 9.0957 0.7070 0.0913
Difference 60 2.874 1.011 0.130

95% CI for mean difference: (2.613, 3.135)

T-Test of mean difference = 0 (vs ≠ 0):
T-Value = 22.02
P-Value = 0.000

P-Value < 0.05 go with Alternative hypothesis

P-Value > 0.05 go with Null hypothesis

59
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
ANOVA Equal Variance (More than 1 sample data set)
Determine whether the means of two or more groups differs significantly
Business Scenario: Machine shop manager want to check the tensile strength of
10mm steal rods procured form three different sources, to see if all the roads are of
same strength or is there any difference (Acceptable level is 400 to 415 PSI)

Minitab exercise
Refer : ANOVA

60
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
ANOVA Equal Variance (More than 1 sample data set)

Null hypothesis All means are equal

Alternative hypothesis At least one mean is
different
Significance level α = 0.05

Means

Source DF Adj SS Adj MS F-Value P-Value

Factor 2 12051.5 6025.74 643.69 0.000

Factor N Mean StDev 95% CI

Vendor 1 30 408.463 3.231 (407.352, 409.573)
Vendor 2 30 385.667 3.230 (384.556, 386.777)
Vendor 3 30 411.654 2.686 (410.543, 412.764)

Pooled StDev = 3.05961

P-Value < 0.05 go with Alternative hypothesis
P-Value > 0.05 go with Null hypothesis
61
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - MEAN Data Type

Pre Requisites
Continuous
Distribution Normal
ANOVA Equal Variance (More than 1 sample data set)

62
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
1 Proportion test
Determine whether the Proportion of a event observed in a sample differs from a
specified value
Business Scenario: As a Sales Manager you need to statistically prove to potential
customer that your machine will produce less defects compared to current defect
rate of 6% in the customer shop floor

Minitab exercise
Refer : 1 proportion test

63
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
1 Proportion test
Null hypothesis : Proportion defective of new machine A = 6%
Alternative hypothesis : Proportion defective of new machine A < 6%

Test of p = 0.06 vs p ≠ 0.06

Exact
Sample X N Sample p 95% CI P-Value
1 6 283 0.021201 (0.007819, 0.045573) 0.003

P-Value < 0.05 go with Alternative hypothesis

P-Value > 0.05 go with Null hypothesis

64
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
1 Proportion test

65
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
1 Proportion test

©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
2 Proportion test
Determine whether the sample Proportion of a event for two groups differs
significantly
Business Scenario: As a procurement manager you have to chose between two
suppliers (“A” and “B”) of spark plugs, comparing their defective %. Requirement is
spark plug should ignite the engine in the first attempt.

Minitab exercise
Refer : 2 proportion test

©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
2 Proportion test

©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
Chi - Square Proportion test
Determine whether the sample Proportion of a event for more two groups differs
significantly
Business Scenario: As a procurement manager you have to chose between two
suppliers (“A” , “B” and “c”) of spark plug comparing their defective %. Requirement
is spark plug should ignite the engine in the first attempt.

Minitab exercise
Refer : 1 proportion test

©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
 LEAN SIX SIGMA ANALYZE

Hypothesis testing - Proportion Data Type

Pre Requisites
Discrete
Distribution Binomial (defective)
Chi - Square Proportion test

©2020 ISEL® All rights reserved. Six Sigma Green Belt Course
1
©2020 ISEL® All rights reserved. Six Sigma Green Belt Course