Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

6-Sigma Green Belt > Analyze > Flashcards

Analyze Flashcards

1
Q

What question does the ANALYZE phase answer?

A

What causes the problem?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What methods are used to collect potential causes & develop hypotheses?

A

Fishbone, FEMA, 5-why, GEMBA

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are the parts of a Fishbone? Describe each part.

A

6 bones (‘M’s’) + Head

Machined
Measure
Man
Mother Nature
Method
Material

Head - Why don’t we fulfill CTC’s?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the steps of the fish-bone process?

A

Complete fishbone - brainstorm & cluster ideas (categories)
Rank ideas - Constant (c), Noise (n), Variables (x)
5-why

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the result of fishbone & 5-why?

A

Hypothesis & potential causes

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is used to prioritize potential causes?

A

Tool 3

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is Tool 3?

A

Table to prioritize potential caused determined from fishbone, 5-why, etc.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are the parts of Tool 3? What is the process for using it?

A

Table: Output measurements vs. potential causes

Prioritize each
- 9 High Impact
- 3 Med Impact
- 1 Low Impact
- 0 No Impact

Sum up each column.
Creating ranking based on sum

Handle the ‘verification’ process - this is the Process door & Data door.
Review results from process & data door.
Determine true root cause

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is GEMBA? What are the steps?

A

An opportunity to capture topics & concerns related to how effectively the process is performing. Purpose is to allow managers/leaders to observe the actual work process, engage with employees, gain knowledge about the work process, and explore opportunities for continuous improvement.

A concept that can be used in any part of the DMAIC cycle or anywhere that value is created.
1. Choose process.
2. Determine hypothesis.
3. Measurements & Data
4. Observations (often video taped)
5. Insights & conclusions.

Rules
1. Go & see
2. Ask why
3. Respect people

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are ways to ‘verify’ potential causes through process (process door)?

A

Process mapping
Value analysis
Time analysis
Capacity analysis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What are the ways to ‘verify’ potential causes through data (data door)?

A

Data Stratification
Hypothesis testing
Design of Experiment (DOE)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What are the different types of processing mapping?

A

Flow charts
Swimlanes
Spaghetti diagrams
Value stream diagrams

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What are the different types of value analysis? Describe each type.

A

Value adding - customer pay
Non-value adding - waste
Value enabling - company pay

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is the tool used to identify waste in a process? Describe the tool.

A

TIM WOODS

Transport
Inventory
Motion
Waiting
Over Production
Over Processing
Defects
Skills

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Describe the ‘Data Stratification.’

A

Find graphical evidence that a potential cause (hypothesis) may be valid by splitting the data into logical subcategories - what, when, how, where, who?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are the calculations for the ‘Time Analysis’ verification step (process door)?

A

Determine process lead time (PLT).
Determine exit rate (ER).
Determine work in process (WIP)
Calculate Little’s Law (WIP/ER)
Calculate process cycle efficiency (PE).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What is Process Lead Time? What is its symbol? How is it calculated?
(Note: Little’s Law is a separate question)

A

PLT = sum of all process time
E.g waiting, processing, setup/change-over time, transportation, inspection, etc.

2 options
1. Measure
2. Little’s Law

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What is the exit rate? How it is calculated? What is its symbol?

A

Exit rate calculates the number of pieces over time for that process.
ER = # pieces/time
Determined by the process step with the lowest capacity (aka the bottleneck)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is WIP?

A

Work in Progress - material waiting to be finished in the process.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is Little’s Law?

A

A way to calculate Process Lead Time (PLT).
PLT = WIP (item) / ER (item/time)

Concept - reduce either WIP or ER to improve process.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is the Basic Principle of Hypothesis Testing?

A

Compare the confidence intervals of different data sets to see if there is a significant difference.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

If you compare the confidence intervals of 2 different sets of data and the data sets overlap, what does that tell you? What if they do not overlap?

A

Overlap = no significant difference in the data sets
Gap = significant difference in the data sets

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

How is Process Cycle Efficiency calculated? What is its symbol?

A

PE = VA/PLT

VA = total processing time of all value-adding steps. To get this, evaluate the process steps using value analysis. Reflection of waste level.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What are the primary parts of Capacity Analysis (process door)?

A

Takt rate calculation
Takt time calculation
Understanding bottlenecks & constraints

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

What is Takt Rate? What is its symbol? How is it calculated?

A

TR = # units to be produced per customer need / available production time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is Takt Time? What is its symbol? How is it calculated?

A

TT = Available production time / # units to be produced per customer need

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

True/False: If the Exit Rate is bigger than the Takt Rate, the process can satisfy customer demand.

A

True

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

True/False: If the Exit Rate is smaller than the Takt Rate, the customer demand can be fulfilled?

A

False

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

What is the difference between bottlenecks & contraints?

A

Bottleneck: slowest process step; there is always a bottleneck
Constraints: if the bottleneck prevents fulfilling customer demand

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

What are the different types of hypothesis options for hypothesis testing? What do the mean?

A

Ho - Null Hypothesis - no significant difference
HA - Alternative Hypothesis - significant difference

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
31
Q

When selecting Ho or Ha during hypothesis testing, the decision can be accepted or rejected based on the p-value. How can one understand the risk of making an incorrect selection?

A

Alpha error: rejection of null hypothesis although it is correct; max accepted risk is 5%

Beta error: failure to reject the null hypothesis although it is wrong

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
32
Q

What are the stratification factors for ‘Data Stratification’?

A

Who?
When?
Where?
What?
How?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
33
Q

What is the purpose of hypothesis testing (data door)?

A

To tell if we are right or wrong about our hypotheses. To check if the differences in our data are statistically significant.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
34
Q

Describe the 2 types of ‘First Leads for Hypotheses.’

A

Comparison Hypotheses - e.g. Kaitlyn shoots shorter than Matt.

The… The… Hypotheses - e.g. The higher the tension, the further the distance.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
35
Q

How can we describe the Null Hypothesis?

A

Ho = Null Hypothesis
Everything is equal (comparison hypotheses)
Everything is free of influence (the… the… hypotheses)
Everything is free of significance (regression model)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
36
Q

True/False: You can prove equality or freedom of the Null Hypothesis?

A

False. You cannot prove equality or freedom, so instead you prove the alternative, aka ‘Reject the Null Hypothesis.’

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
37
Q

How can we describe the Alternative Hypothesis?

A

Ha = Alternative Hypothesis
It is not equal.
It is not free of influence.
It is not free of significance.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
38
Q

True/False: You can prove an alternative hypothesis not equal.

A

True. You can prove prove Ha. If Ha is valid, then Ho can be rejected.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
39
Q

In hypothesis testing, how would you describe the risk you are willing to take?

A

In statistics, accepted alpha-error is 5%, which is the max accepted risk to reject Ho wrongly.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
40
Q

What does it tell you if the p-value is smaller than the alpha-level?

A

Shows the actual risk of rejecting Ho wrongly.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
41
Q

Hypothesis Test Overview: What tests for the below?
Target Value (y) = discrete
Influencing Value (x) = discrete

A

1-proportion test
2-proportion test
chi^2 test

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
42
Q

Hypothesis Test Overview: What tests for the below?
Target Value (y) = continuous
Influencing Value (x) = discrete

A

1-sample t-test
2-sample t-test
1-way ANOVA
2-way ANOVA
2-variances test –> F-test or Levine’s test
Test for equal variances –> Bartlett’s or Levine’s test

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
43
Q

Hypothesis Test Overview: What tests for the below?
Target Value (y) = continuous
Influencing Value (x) = continuous

A

Correlation
Regression

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
44
Q

Hypothesis Test Overview: What tests for the below?
Target Value (y) = discrete
Influencing Value (x) = continuous

A

Logistic Regression

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
45
Q

What are the names of the 3 normality tests in Minitab?

A

Anderson-Darling
Ryan-Joiner
Kolmogorov-Smirnov

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
46
Q

How do you deal with non-normality? (4 options)

A
  1. Get more samples.
  2. Data stratification.
  3. Remove special causes & document
  4. Take data like it is.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
47
Q

What are the prerequisites for the 1-way ANOVA hypothesis test?

A

Sigma^2(1) = Sigma^2(2) = Sigma^2(3)…
which is the Ho for Test for Equal Variances…
which means Bartlett’s test or Levine’s test

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
48
Q

In hypothesis tests, what is ‘x’?

A

Influencing factor

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
49
Q

In hypothesis tests, what is ‘y’?

A

Output

50
Q

True/False: In hypothesis testing, you can have both discrete & continuous factors?

A

True.

51
Q

What is the model to check the Mu for the 1-way ANOVA test?

A

y = y-bar(total) + Tao(i) + Epsilon

Tao = Individual influence/effect of the shooter
Epsilon = ‘Unexplained rest’

52
Q

What are the 3 steps to evaluating the 1-way ANOVA test?

A
  1. Visualize data: box plot + main effects chart
  2. Check the Model: Ho: T(i) = 0
  3. Determination coefficient (in more detail on another slide)
53
Q

What is the determination coefficient?

A

R^2(adj) > 80%, which tells you how good your [influencing] factor affects the reality. At this point, you’ve found enough root causes to stop investigating.

54
Q

What is the difference between a 1-way ANOVA & a 2-way ANOVA?

A

1-way ANOVA = 1 factor
2-way ANOVA = 2 factors

55
Q

Describe a 2-way ANOVA.

A

It is a hypothesis test with 2 influencing factors.

X factors have a direct influence on y (output).
X factors have an indirect influence on each other.

Requirements: balanced design

Example:
y = distance (continuous)
x = Ball + Shooter (both discrete)

56
Q

What is the requirement of a 2-way ANOVA?

A

A balanced design!

57
Q

What are the 4 steps to evaluating the 2-way ANOVA test?

A
  1. Visualize data: box plot + main effects chart + Interaction plot
  2. Check the Model:
    Ho: T(i) = 0
    Ho: B(j) = 0
    Ho: (TB)ij = 0
  3. Determination coefficient (in more detail on another slide)
  4. Check Residuals (aka Epilson - unexplained rest)
58
Q

What is the model of a 2-way ANOVA?

A
59
Q

What are the 3 things you check for when checking ‘Residuals’ (aka Epsilon/unexplained rest) for 2-way ANOVA?

A
  1. Normal distribution
  2. Check for patterns over y
  3. Check for trends over time
60
Q

What are the 2 more specific tests for the 2-variances hypothesis test?

A

F-test - normal distribution
Levine’s test - non-normal distribution

61
Q

What are the 2 more specific tests for the Test for Equal Variances?

A

Bartlett’s test - normal distribution
Levine’s test - non-normal distribution

62
Q

What are the prerequisites for proportion based hypothesis tests?

A

n >/ 100 & at least 5 in each category

63
Q

What is the correlation hypothesis test checking for?

A

Checking the magnitude of the relation. Is there a significant connection between A & B, & how strong is the connection?

64
Q

What is the symbol for Pearson’s Correlation Coefficient?

A

r

65
Q

What is the r value for a Perfect Correlation? Describe the graph(s).

A

-1 = r = 1
Diagonal linear line going up or down.

66
Q

What is the r value when no correlation? Describe the graph(s).

A

r = 0
Scatter plot, dots all over
Horizontal line across

67
Q

What is the r value when there is no linear correlation? Describe the graph(s). How is the correlation ranked in this case?

A

r = 0
Various options, but a non-linear shape.

Rank by absolute value…
>/ 0.8 strong
>/ 0.5 medium
</ 0.5 weak

68
Q

What does FMEA stand for?

A

Failure Mode & Effect Analysis

69
Q

What are the 2 types of FMEA analysis?

A

Design FMEA: check design for potential risks
Design FMEA: check process for potential risks

70
Q

For FMEA, the Analyze phase looks at what? Improve phase?

A

Analyze - ‘As Is Process’
Improve - ‘To Be Process’

71
Q

In FMEA, how are identified risks categorized?

A

Severity, Occurrence, Detection

72
Q

In FMEA, how do you calculate the Risk Priority Number (RPN)? (new & old way)

A

RPN = S * O * D (old way)
Action Priority (AP) - High, Medium, Low

73
Q

What does Regression hypothesis testing tell you?

A

Describes the influence of X on Y statistically.

74
Q

In regression testing, x is called the…

A

Predictor

75
Q

In regression testing, y is called the…

A

Response

76
Q

In the Analyze Phase, regression testing is looking at?

A

Cause & Effect - how big is the impact on X & Y?

77
Q

In the Improve Phase, regression testing is looking at?

A

Prognosis - what should the X be like if I want Y to be… (target). Use the model to get the target value you want - interpolation.

78
Q

What is the difference between Interpolation & Extrapolation?

A

Interpolation: estimate a value within two known values

Extrapolation: estimate a value based on extending a known sequence of value

79
Q

How many Predictors (factors) in a Simple Linear Regression?

A

1 predictor/factor

80
Q

How many Predictors (factors) in a Multiple Linear Regression?

A

2+ predictors/factors

81
Q

What is the prerequisite of a multiple linear regression?

A

Predictors must be independent of each other (no correlation).

82
Q

What is the indicator that the predictors are independent of each other (no correlation)? How do you know if it’s good or bad?

A

Variation Inflation Factor (VIF)
VIF = 1 :)
VIF >/ 5 :(

83
Q

What do you do if your VIF value is too high?

A

Kick out factors to meet the prerequisites before running the test.

84
Q

How do you Analyze regression data in Minitab (4 items)?

A
  1. R^2: goal >/ 80%
  2. Is the model significant?
  3. Are the coefficients significant?
  4. Analyze Residuals
85
Q

How do you Analyze Residuals?

A
  1. Check for normal distribution
    (Ho = Normal Distribution, p-value)
  2. Check for Patterns over Y (cone)
  3. Check for trends over time
86
Q

Describe the OFAT Strategy. What are the pros & cons?

A

Change only 1 factor at a time to find the correlation & best result. It’s intuitive & the effect is clear, however, it requires a lot of setup and you cannot see the interactions between the factors.

87
Q

In 2^k Design, what does the 2 stand for?

A

Factor Levels (min/max of each factor)

88
Q

In 2^k Design, what does the k stand for?

A

Factors

89
Q

What assumption does the 2^k Design make?

A

Must be a linear model

90
Q

In the Process Model, what are the factors affecting the target value?

A

Influencing factors, Constant factors, Noise factors

91
Q

How do you deal with variation from unknown noise?

A

Repeats & Replicates

92
Q

Why do we want centerpoints?

A

Need to check the assumption of linearity & have more data to review.

93
Q

When calculating the Pearson’s Correlation Coefficient, what is the Ho?

A

Ho: A & B are ‘free of each other’ or ‘free of correlation’

94
Q

What is the formula for Multiple Linear Regression?

A

y = bo + b1x1… + bu xu + Epsilon

95
Q

In the regression model, what is Bo?

A

The intersection with the y-axis.

96
Q

In the regression model, what is B1?

A

The slope gradient of the line generated by the model

97
Q

In the regression model, what is x?

A

The x-axis of that particular factor.

98
Q

In the regression model, what is Epsilon?

A

The unexplained rest.

99
Q

In the regression model, how do you analyze the coefficients?

A

Ho: Bo = 0
(line goes through the origin; Bo is free of significance)

Ho: B1 =
(line is parallel to the x-axis, no slope; B1 is free of significance)

100
Q

When dealing with unknown noise, what are repeats?

A

Used to understand how to deal with short term unknown noise.

“Make more parts/tests”

101
Q

When dealing with unknown noise, what are replicates?

A

Used to understand how to deal with long term noise.

“Do the whole plan again”

102
Q

What do I do if I don’t have data, a clue, or a lot of money to spend?

A

Design of Experiment!

103
Q

What’s the intuitive approach in DOE?

A

OFAT

104
Q

In DOE, how do I know the effect of interactions between the factors?

A

Perform a Full Factorial Design

105
Q

How to calculate the number of setups in a Full Factorial Design?

A

2^k… k is the number of factors

106
Q

When you use DOE to understand the interactions between the factors, how can you tell if the interactions are significant?

A

To the right side of the significance line. You can also check the p-values.

107
Q

In DOE, what do you do if the interactions are not significant?

A

Remove non-significant factors 1 at a time. Then the model will recalculate. Always start with the smallest one.

108
Q

When removing non-significant interactions in DOE, be careful not to what?

A

Remove a factor where there is an interaction (letter combo).

109
Q

In a main effects plot, how can you tell if there is an effect from from the factor?

A

Diagonal line = effect
Horizontal line = no effect

110
Q

In an interaction plot, how can you tell if there is an interaction?

A

Interaction if crossing lines.

111
Q

What are the 5 things to check when reviewing analytical results in DOE?

A
  1. Check VIF
  2. Significance of main factors & interactions (p-value)
  3. Determination Coefficient (R^2 adj.) >80%
  4. Check the significance of the model (p-value)
  5. Check the model’s suitability - Residuals
112
Q

True or False: Centerpoints are included in the Residual Analysis?

A

False

113
Q

What is the difference between Full Factorial Design and Fractional Factorial Design?

A

Full Factorial Design looks at all possible combinations.

Fractional Factorial Design looks at less combinations to save money.

114
Q

How do you determine the number of samples/tests for a Fractional Factorial Design? What does each item mean?

A

2^k-q
k = # factors
q = reduction level
2 = # factorial levels (min/max)

115
Q

What is the risk of using Fractional Factorial Design?

A

Confounding effects! Must decide if we can accept the risk of these effects.

116
Q

With a resolution Type III, what is confounded?

A

1 main factors confounded, 2 factor interactions

117
Q

What can you find in your Lean toolbox?

A

5s
Push/Pull
Polka Yoke
Kanban
Little’s Law
PDCA
SMED
TIM WOODS
JIT
Kaizen

118
Q

Why was Lean developed?

A

MUDA (TIM WOOD)
MURI (overuse/underuse of resources - S from TIM WOODS)
MUSA (unbalanced)

119
Q

How can I find out PLT? Must have what to use this process?

A

Little’s Law = PLT = WIP/ER
Must have stable WIP.

120
Q

Why does confounding happen?

A

Confounding arises in fractional designs only and is caused by the signage of one column being the same as another. This means when calculating the effect of these factors/interactions, we will get the same value for the size of the effect.

6-Sigma Green Belt (7 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions