The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:

Slides:



Advertisements
Similar presentations
Multiple Comparisons in Factorial Experiments
Advertisements

Chapter 9A Process Capability and Statistical Quality Control
II.2 Four Factors in Eight Runs Introduction Introduction Confounding Confounding –Confounding/Aliasing –Alias Structure Examples and Exercises Examples.
Experimental Design, Response Surface Analysis, and Optimization
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
II.2 Four Factors in Eight Runs Some Fiber Optics Examples and Exercises* Some DOE’s Some DOE’s Responses: Responses: –Shrinkage (S) –Excess.
14-1 Introduction An experiment is a test or series of tests. The design of an experiment plays a major role in the eventual solution of the problem.
Go to Table of ContentTable of Content Analysis of Variance: Randomized Blocks Farrokh Alemi Ph.D. Kashif Haqqi M.D.
T WO W AY ANOVA W ITH R EPLICATION  Also called a Factorial Experiment.  Factorial Experiment is used to evaluate 2 or more factors simultaneously. 
T WO WAY ANOVA WITH REPLICATION  Also called a Factorial Experiment.  Replication means an independent repeat of each factor combination.  The purpose.
1 14 Design of Experiments with Several Factors 14-1 Introduction 14-2 Factorial Experiments 14-3 Two-Factor Factorial Experiments Statistical analysis.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
Factorial Experiments
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
Design of Experiments Chapter 21.
QNT 531 Advanced Problems in Statistics and Research Methods
1 Least squares procedure Inference for least squares lines Simple Linear Regression.
6.1 What is Statistics? Definition: Statistics – science of collecting, analyzing, and interpreting data in such a way that the conclusions can be objectively.
Chapter 13Design & Analysis of Experiments 8E 2012 Montgomery 1.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
F OUNDATIONS OF S TATISTICAL I NFERENCE. D EFINITIONS Statistical inference is the process of reaching conclusions about characteristics of an entire.
The Essentials of 2-Level Design of Experiments Part II: The Essentials of Fractional Factorial Designs The Essentials of 2-Level Design of Experiments.
The Essentials of 2-Level Design of Experiments Part II: The Essentials of Fractional Factorial Designs The Essentials of 2-Level Design of Experiments.
Design of Engineering Experiments Part 5 – The 2k Factorial Design
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
IV.3 Designs to Minimize Variability Background Background An Example An Example –Design Steps –Transformations –The Analysis A Case Study A Case Study.
Measures of central tendency are statistics that express the most typical or average scores in a distribution These measures are: The Mode The Median.
III.7 Blocking Two-level Designs _ Blocking _ Example _ Four Blocks _ Exercise.
Design Of Experiments With Several Factors
1 Doing Statistics for Business Doing Statistics for Business Data, Inference, and Decision Making Marilyn K. Pelosi Theresa M. Sandifer Chapter 14 Experimental.
II.3 An Example and Analyzing Interactions Emergency Room Example Emergency Room Example Interaction Plots and Tables Interaction Plots and Tables.
Understanding Your Data Set Statistics are used to describe data sets Gives us a metric in place of a graph What are some types of statistics used to describe.
1.1 Statistical Analysis. Learning Goals: Basic Statistics Data is best demonstrated visually in a graph form with clearly labeled axes and a concise.
DOX 6E Montgomery1 Design of Engineering Experiments Part 2 – Basic Statistical Concepts Simple comparative experiments –The hypothesis testing framework.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
Copyright © 2013, 2009, and 2007, Pearson Education, Inc. Chapter 14 Comparing Groups: Analysis of Variance Methods Section 14.3 Two-Way ANOVA.
MEASURE : Measurement System Analysis
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I:
Statistics (cont.) Psych 231: Research Methods in Psychology.
Testing Differences in Means (t-tests) Dr. Richard Jackson © Mercer University 2005 All Rights Reserved.
Designs for Experiments with More Than One Factor When the experimenter is interested in the effect of multiple factors on a response a factorial design.
Engineering Statistics Design of Engineering Experiments.
Chapter 15 Analysis of Variance. The article “Could Mean Platelet Volume be a Predictive Marker for Acute Myocardial Infarction?” (Medical Science Monitor,
Inference for Linear Regression
BAE 5333 Applied Water Resources Statistics
Two way ANOVA with replication
CHAPTER 13 Design and Analysis of Single-Factor Experiments:
Two way ANOVA with replication
Determining the Best Measure of Center
Psych 231: Research Methods in Psychology
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
The Essentials of 2-Level Design of Experiments Part II: The Essentials of Fractional Factorial Designs Developed by Don Edwards, John Grego and James.
Some 24-1 DOE’s Responses: Shrinkage (S) Excess Length (L)
The Essentials of 2-Level Design of Experiments Part II: The Essentials of Fractional Factorial Designs Developed by Don Edwards, John Grego and James.
III.7 Blocking Two-level Designs
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
ENM 310 Design of Experiments and Regression Analysis Chapter 3
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
II.2 Four Factors in Eight Runs
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
IV.3 Designs to Minimize Variability
The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch.
14 Design of Experiments with Several Factors CHAPTER OUTLINE
Presentation transcript:

The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs The Essentials of 2-Level Design of Experiments Part I: The Essentials of Full Factorial Designs Developed by Don Edwards, John Grego and James Lynch Center for Reliability and Quality Sciences Department of Statistics University of South Carolina

Part I.3 The Essentials of 2-Cubed Designs Methodology Methodology –Cube Plots – Estimating Main Effects – Estimating Interactions (Interaction Tables and Graphs) Statistical Significance: When is an Effect “Real”? Statistical Significance: When is an Effect “Real”? An Example With Interactions An Example With Interactions A U-Do-It Case Study A U-Do-It Case Study Replication Replication Rope Pull Exercise Rope Pull Exercise

U-Do-It Exercise Rope Pull Study* with Replication Purpose of the Design Purpose of the Design –Test Hose to Determine the Effect of Several Factors on an Important Quality Hosiery Characteristic, Rope Pull –Response y = Upper Boot Rope Pull (in inches) Factors: Factors: –A: Vacuum level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200) Two Replicates of the Full 2 3 Were Performed * Empirical basis for this data was motivated by a much larger study performed by the developers at Sara Lee Hosiery Two Replicates of the Full 2 3 Were Performed * Empirical basis for this data was motivated by a much larger study performed by the developers at Sara Lee Hosiery Purpose of the Design Purpose of the Design –Test Hose to Determine the Effect of Several Factors on an Important Quality Hosiery Characteristic, Rope Pull –Response y = Upper Boot Rope Pull (in inches) Factors: Factors: –A: Vacuum level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200) Two Replicates of the Full 2 3 Were Performed * Empirical basis for this data was motivated by a much larger study performed by the developers at Sara Lee Hosiery Two Replicates of the Full 2 3 Were Performed * Empirical basis for this data was motivated by a much larger study performed by the developers at Sara Lee Hosiery

Replication Why? Average values have less variability as the number of things you average increases Average values have less variability as the number of things you average increases –Estimated effects will be reliably closer to true effects –More of the mid-sized and small effects will be distinguishable from error Data from replicated experiments can be used to estimate the amount of variability in the process (This allows more formal test for “real” effects—ANOVA). Data from replicated experiments can be used to estimate the amount of variability in the process (This allows more formal test for “real” effects—ANOVA). Data from replicated experiments can be used to determine not only which factors affect the mean of the process, but which factors affect the variability of the process. Data from replicated experiments can be used to determine not only which factors affect the mean of the process, but which factors affect the variability of the process.

Replication Analysis of a Replicated 2 3 Replication means repeating the entire set of 8 runs, but (for the analysis as described below), the entire collection of runs should be done in random order (be it 16, or 24, or 48, etc. runs); if you want to do them in complete sets of 8, you should analyze the results in blocks—explained later). Replication means repeating the entire set of 8 runs, but (for the analysis as described below), the entire collection of runs should be done in random order (be it 16, or 24, or 48, etc. runs); if you want to do them in complete sets of 8, you should analyze the results in blocks—explained later). For our analysis, you can reduce the data to averages over each of the 8 treatment combinations; use these averages as your “y’s” in the rest of the analysis. For our analysis, you can reduce the data to averages over each of the 8 treatment combinations; use these averages as your “y’s” in the rest of the analysis. –Discussion of shortcomings of this approach to follow Effects plot, interaction plots, and EMR calculations are done as before using these estimated effects. Effects plot, interaction plots, and EMR calculations are done as before using these estimated effects. Replication Example to Follow!

U-Do-It Exercise Rope Pull Study - Experimental Report Form

U-Do-It Exercise Rope Pull Study - The Analysis To do: Analyze the data. This should include... To do: Analyze the data. This should include... –Fill in the table on the next slide. –Analyze the averages in Minitab:  Create a 3-factor 2-level design, enter the averages as a response variable; compute factor effects and construct a normal probability plot of the effects.  If appropriate, graph interaction plots.  Compute EMR using only the significant terms To do: Analyze the data. This should include... To do: Analyze the data. This should include... –Fill in the table on the next slide. –Analyze the averages in Minitab:  Create a 3-factor 2-level design, enter the averages as a response variable; compute factor effects and construct a normal probability plot of the effects.  If appropriate, graph interaction plots.  Compute EMR using only the significant terms

U-Do-It Exercise Rope Pull Study - The Analysis

U-Do-It Exercise Solution Rope Pull Study The signs table, cube plot, effects normal probability plot and AC interaction table and graph are given on the next few pages. The signs table, cube plot, effects normal probability plot and AC interaction table and graph are given on the next few pages. –The cube plot leads us to expect a negative main effect for A (Vacuum level), and a positive main effect for C (upper boot speed). Note that the changes in the response for changes in A are much larger at Lo C than at Hi C, which suggests an AC interaction. Estimated effects from the response table and the normal probability plot of effects support this observation. –An AC interaction table and plot are therefore called for, and have been constructed. The signs table, cube plot, effects normal probability plot and AC interaction table and graph are given on the next few pages. The signs table, cube plot, effects normal probability plot and AC interaction table and graph are given on the next few pages. –The cube plot leads us to expect a negative main effect for A (Vacuum level), and a positive main effect for C (upper boot speed). Note that the changes in the response for changes in A are much larger at Lo C than at Hi C, which suggests an AC interaction. Estimated effects from the response table and the normal probability plot of effects support this observation. –An AC interaction table and plot are therefore called for, and have been constructed.

U-Do-It Exercise Solution Rope Pull Study - Completed Cube Plot and Signs Table Factors: Factors: –A: Vacuum Level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200) Response: Response: –Rope Pull (in inches) Factors: Factors: –A: Vacuum Level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200) Response: Response: –Rope Pull (in inches)

U-Do-It Exercise Solution Rope Pull Study -Completed Seven Effects Paper Factors: Factors: –A: Vacuum Level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200) Factors: Factors: –A: Vacuum Level (Lo, Hi) –B: Needle Type (EX, GB) –C: Upper Boot Speed (1000,1200)

U-Do-It Exercise Solution Rope Pull Study - Completed AC Interaction Table

U-Do-It Exercise Solution Rope Pull Study - AC Interaction Plot o Factors A: Vacuum Level (Lo, Hi) C: Upper Boot Speed (1000,1200)

U-Do-It Exercise Solution Rope Pull Study - Interpretation of the Experiment There is non-ignorable interaction between A = Vacuum level and C = Upper boot speed, so we should not interpret main effects for these factors individually. For example, a Hi Vacuum level greatly increases the effect of a change from 1000 to 1200 RPM in Upper boot speed. Judging from the interaction plot, There is non-ignorable interaction between A = Vacuum level and C = Upper boot speed, so we should not interpret main effects for these factors individually. For example, a Hi Vacuum level greatly increases the effect of a change from 1000 to 1200 RPM in Upper boot speed. Judging from the interaction plot, –At Lo Vacuum level, we expect a decrease of about 7” in rope pull when changing Upper boot speed from 1000 to 1200 RPM. –At Hi Vacuum level, we expect a decrease of about 14’’ in rope pull when changing Upper boot speed from 1000 to 1200 RPM. –At 1000 RPM Upper boot speed, we expect an increase of about 10” in rope pull when changing Vacuum level from Lo to Hi. – At 1200 RPM Upper boot speed, we expect an increase of about 3: in rope pull when changing Vacuum level from Lo to Hi. The above interpretations hold for both needle types (Factor B). There is no detectable difference in mean rope pull between the two needle types. The above interpretations hold for both needle types (Factor B). There is no detectable difference in mean rope pull between the two needle types.

Replication Why? It Allows You To See Things More Clearly! Below Are the Normal Probability Plots for the First and Second Replication Below Are the Normal Probability Plots for the First and Second Replication Notice How Hard it is to see that the AC Interaction is Significant Notice How Hard it is to see that the AC Interaction is Significant Below Are the Normal Probability Plots for the First and Second Replication Below Are the Normal Probability Plots for the First and Second Replication Notice How Hard it is to see that the AC Interaction is Significant Notice How Hard it is to see that the AC Interaction is Significant

Replication Why? It Allows You To See Things More Clearly! Below Are the Normal Probability Plots for the Individual Replicates and the one based on the Averages Below Are the Normal Probability Plots for the Individual Replicates and the one based on the Averages Notice How Replication Makes it Easier to see that the AC Interaction is Significant Notice How Replication Makes it Easier to see that the AC Interaction is Significant Below Are the Normal Probability Plots for the Individual Replicates and the one based on the Averages Below Are the Normal Probability Plots for the Individual Replicates and the one based on the Averages Notice How Replication Makes it Easier to see that the AC Interaction is Significant Notice How Replication Makes it Easier to see that the AC Interaction is Significant

Replication Why? It Allows You To Use ANOVA! o The Small p-Values in the ANOVA Table Indicate that there are Significant Main Effects and that the Interaction is Significant o The zero p-values in the Factor Effects Table Indicate that both A and B are Significant o The Small p-Values in the ANOVA Table Indicate that there are Significant Main Effects and that the Interaction is Significant o The zero p-values in the Factor Effects Table Indicate that both A and B are Significant