Quality

Improvement
Chapter 13- Experimental
Design

PowerPoint presentation to accompany

Besterfield, Quality Improvement, 9e
 Homework Assignments

• HW #13 - Due (For Practice Only)

• Pages 209-210: 6, 8 (Significance Test only), 9, 10, and 12

2
 Outline
1. Introduction
2. Basic Statistics
3. Hypotheses
4. One factor at a time
5. Orthogonal Design
6. Two Factors
7. Full Factorials
8. Fractional Factorials
9. Examples
10.Final Considerations
3
 Learning Objectives
When you complete this chapter, you are expected to:

❑ Know the applicable terminology;

❑ Understand the concept of hypothesis testing;
❑ Be able to determine significant factors and their levels using the F
test;
❑ Understand the concept of Full and fraction factorials.

.
.
.

4
 Introduction
• Experimental design is a methodology of varying a number of
input factors simultaneously in a carefully planned manner, such
that their individual and combined effects on the output can be
identified.
• Goals are to determine :
• The variable(s) and their magnitude that influences the
response.
• How to manipulate these variables to control the response.
• For example, it may be desirable to understand the effect of
temperature and pressure on the strength of a glue bond.
- Does change in temperature or pressure affect the strength of
a glue bond

5
 Definitions
• Factor– Controllable variable (time, temperature, operator, etc)
• Level—value assigned to a factor (50°F for temp)
• Treatment Condition (TC)—set of conditions (factor/levels)
• Replicate—repeat of TC with change in set up
• Repetition—multiple results of a TC
• Random—TC are run in a random order
• Orthogonal Array—TC are put together so the design is
balanced and factor/levels can be analyzed singly or in
combination.
• Interaction—two or more factors that together produce results
different than their separate effects

6
Basic Statistics

7
 Hypotheses
• Hypotheses (Ho) testing is a statistical decision where inferences are made
about the population from one sample or populations of two samples

• Does or doesn’t the mean of one sample from a populations differ from a value
μ0 (One Sample):
Null hypothesis.

Ho: μ = μ0

• Does or doesn’t the means of two samples from two identical populations differ.
Null hypothesis (Two Samples).

Ho: μ1 = μ2 OR μ1 - μ2 = 0

• Because we are dealing with samples there is the risk of Type I/II Errors:
- Type I error: Null hypothesis is rejected when it is actually true
- Type II error: Null hypothesis is accepted when it is actually false

8
 Hypotheses (Continued)
• Therefore, we use alternate hypothesis (One Sample):
• μ ǂ μ0 Both tails of distribution
• μ > μ0 Right tail of distribution
• μ < μ0 Left tail of distribution

• Therefore, we use alternate hypothesis (Two Samples):

• μ1 ǂ μ2 Both tails of distribution
• μ1 > μ2 Right tail of distribution
• μ1 < μ2 Left tail of distribution

9
 Analysis of Variance (ANOVA)
• One factor (Multiple Levels)
• Mathematical model of an experiment

10
 ANOVA (Continued)
• Initial calculation equations for a spreadsheet are:

• N = Total number of observation

• Ai = Total of Observations under ith factor
• T = Total of all observations under all factor levels
• SST = Total variability in data
• SSF = Sum of squares due to the factor
• SSE = Sum of squares due to the error

11
ANOVA (Continued)

12
 Experimental Design (One Factor)

Once you calculate the SS values the remaining

calculations easy. Critical F value is obtained from
Appendix C and if it is smaller than the F value the test is
significant
13
 ANOVA (Example 13-4)
For the following data, determine if there is a significance difference in the data (factor level
means are different). use significance level = 0.05:
Level Data
1 26 25 32 29 30
2 34 31 26
3 15 16 25 18 13
4 12 15 13 11
Hypotheses:
H0: µ1 = µ2 = µ3 = µ4 and H1: At least two of the means are not equal

ANOVA Table:
Sum of Degress Mean F F Significant
Source Squares Freedom Squares (Calc) (Table) (yes/no)
Factor 860.65 3 286.88 23.33 3.41 Yes
Error 159.85 13 12.3
Total 1020.5

DECISION: Since F (Calculated) > F (Table), Reject H0

➔ The factor level means are different

14
 One Factor at a Time Design
• Multiple factors (Two levels for each factor)
• Is not balanced (only one 2 in each column)
• Effects (e)

15
One Factor at a Time Design (Example 13-5)
• Objective: Maximize Response Variable Factors Level 1 Level 2
Speed (m/s) 25 35
• Three factors at two levels Pressure (Pa) 50 75
Time (min) 5 10
FACTORS/LEVLES
Treatment Response
Condition Spped Pressure Time (Results)
1 1 (25 M/S) 1 (50 Pa) 1 (5 min) 2.8
2 2 (35 M/S) 1 (50 Pa) 1 (5 min) 3.4
3 1 (25 M/S) 2 (75 Pa) 1 (5 min) 4.6
4 1 (25 M/S) 1 (50 Pa) 2 (10 min) 3.7
eS = yS - y0 = 3.4 – 2.8 = 0.6
ep = yp - y0 = 4.6 – 2.8 = 1.8
et = yt - y0 = 3.7 – 2.8 = 0.9
• It appears that speed, time, and perhaps pressure have a strong positive
affect on the process. Increasing any of the factors increases the response.
16
 Orthogonal Design
Orthogonal means balanced (Levels – Same number of 1’s and 2’s in each column

𝐴1 = Results of TC’s (1 + 3 + 5 + 7 ) /4
𝐴2 = Results of TC’s (2 + 4 + 6 + 8 ) /4
𝐵1 = Results of TC’s (1 + 2 + 5 + 6 ) /4
𝐵2 = Results of TC’s (3 + 4 + 7 + 8 ) /4
𝐶1 = Results of TC’s (1 + 2 + 3 + 4 ) /4
𝐶2 = Results of TC’s (5 + 6 + 7 + 8 ) /4

𝑒𝐴 = 𝐴2 − 𝐴1
𝑒𝐵 = 𝐵2 − 𝐵1
𝑒𝐶 = 𝐶2 − 𝐶1 17
Orthogonal Design (Example 13.6)
FACTOR/LEVELS
Treatment Response
Condition Speed Pressure Time (Results)
1 1 (25 M/S) 1 (50 Pa) 1 (5 Min) 2.8
2 2 (35 M/S) 1 (50 Pa) 1 (5 Min) 3.4
3 1 (25 M/S) 2 (75 Pa) 1 (5 Min) 4.6
4 2 (35 M/S) 2 (75 Pa) 1 (5 Min) 3.8
5 1 (25 M/S) 1 (50 Pa) 2 (10 Min) 3.7
6 2 (35 M/S) 1 (50 Pa) 2 (10 Min) 2.7
7 1 (25 M/S) 2 (75 Pa) 2 (10 Min) 3.1
8 2 (35 M/S) 2 (75 Pa) 2 (10 Min) 4.4

2.8+4.6+3.7+3.1 3.4+3.8+2.7+4.4
𝑆1 = 4
= 3.550 𝑆2 = 4
= 3.575
2.8+3.4+3.7+2.7 4.6+3.8+3.1+4.4
𝑃1 = 4
= 3.150 𝑃2 = 4
= 3.975
2.8+3.4+4.6+3.8 3.7+2.7+3.1+4.4
𝑇1 = 4
= 3.650 𝑇2 = 4
= 3.475

18
 Orthogonal Design (Example 13.6)

𝑒𝑠 = 𝑆2 − 𝑆1 = 3.575 – 3.550 = 0.025

𝑒𝑃 = 𝑃2 − 𝑃1 = 3.975 – 3.150 = 0.825
𝑒𝑇 = 𝑇2 − 𝑇1 = 3.475 – 3.650 = - 0.175
Compare results of one factor at a time to Orthogonal
Design
An increase in Pressure from (Level 1 to Level 2) has a
strong effect on the process, where small effects of
Speed and Time most likely due to natural variation in
the process.

19
 Two Factors
❑ A simple experimental design has two factors and can have many
levels as shown below; however, two levels are common and
sometimes three levels. Each factor/level can have replications.

20
 Two Factors (Continued)
ANOVA is the decision method and once the SS are
calculated, the process leads to a decision using the
F test.

Example 13-9 (Text Book)

21
 Two Factors (Example)
Consider an experiment on the volatility of a fluid with two factors,
concentration (factor A) and temperature (factor B). Determine the
significance of each factor for a level α = 0.05. Following are the results
(responses) of the experiment.

Factor B Levels (Temperature)

Factor A Levels (Concentration) 1 2 3
1 41, 38 43, 49 46, 47
2 42, 32 46, 45 48, 49

First, we need to calculate the sum of squares, mean

square, Degrees of Freedom, and F are calculated
(See Example 13.9 in Text Book)

22
 Two Factors (Example – Continued)
Degrees
Sum of of Mean Significant
Source Squares Frredom Square F (Yes/No)
A 6.75 1 6.75 1.6535 5.99 NO
B 90.5 2 42.25 11.0816 5.14 YES
AB 6.5 2 3.25 0.7959 5.14 NO
Error 24.5 6 4.0833
Total 128.25 11

Based on these calculation, the effect of temperature (Factor

B) on volatility of the fluid is significant; whereas the effects of
concentration (Factor A) and the interaction between
concentration and temperature are not significant at a level α
= 0.05.
23
 Full Factorials
A 3 factor/2 level, with a + for the high level and a – for the low level is shown
below. Signs for the 3 factors is as shown.

24
 Full Factorials (Continued)
• Signs for the 2 factor and 3 factor interactions are based on the
signs of the 3 factors in each TC.

• Two factor Interaction signs are (- x - = +)

(- x + = -) (+ x - = -) (+ x + = +)

• Three factor interaction signs are (- x – x - = -)

(- x + x - = +) (+ x + x + = +) (+ x - x + = -)

25
 Full Factorials (Continued)
• ANOVA is the decision method and once the SS are calculated, the process
leads to a decision using the F test.

26
 Full Factorials (Continued)
The number of TC is determined by

Thus for a 2 level design 2² = 4; 2³ = 8; 4-16; 5-32; etc

for a 3 level design 3² = 9; 3³ = 27; 4-81; etc

27
 Fractional Factorials
❑The number of TC’s can become quite large.
❑Use Engineering judgment to eliminate some
interactions and substitute a factor

28
 Examples
• Evans Clay
• John Deere
• Nabisco
• Wilkes-Barre Hospital
• Eastman Kodak
• Farmington High School
• K2 Corporation
• Hercules Corporation

29
 Implementation
• Set good factor/levels and response variable
• Replicate and random
• Block out known or unwanted sources of variation such as
environment
• Evaluate interactions and make adjustments
• Transfer lessons learned to other experiments
• Confirm results with another experiment

30
 Computer Program
• Microsoft EXCEL will calculate t distribution, F distribution,
confidence and SS.

31