Chapter 5 Introduction to Factorial Designs

Similar documents
Design of Engineering Experiments Chapter 5 Introduction to Factorials

Factorial designs. Experiments

Design & Analysis of Experiments 7E 2009 Montgomery

Addition of Center Points to a 2 k Designs Section 6-6 page 271

Lec 5: Factorial Experiment

3. Factorial Experiments (Ch.5. Factorial Experiments)

Chapter 5 Introduction to Factorial Designs Solutions

Contents. TAMS38 - Lecture 6 Factorial design, Latin Square Design. Lecturer: Zhenxia Liu. Factorial design 3. Complete three factor design 4

Lecture 9: Factorial Design Montgomery: chapter 5

Design and Analysis of Experiments

Lecture 10. Factorial experiments (2-way ANOVA etc)

3.4. A computer ANOVA output is shown below. Fill in the blanks. You may give bounds on the P-value.

Contents. TAMS38 - Lecture 8 2 k p fractional factorial design. Lecturer: Zhenxia Liu. Example 0 - continued 4. Example 0 - Glazing ceramic 3

Chapter 13 Experiments with Random Factors Solutions

Factorial Treatment Structure: Part I. Lukas Meier, Seminar für Statistik

DESAIN EKSPERIMEN BLOCKING FACTORS. Semester Genap 2017/2018 Jurusan Teknik Industri Universitas Brawijaya

Stat 217 Final Exam. Name: May 1, 2002

Analysis of Variance and Design of Experiments-I

These are multifactor experiments that have

Design and Analysis of

Orthogonal contrasts for a 2x2 factorial design Example p130

Two-Way Factorial Designs

Lecture 10: Experiments with Random Effects

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008

The 2 k Factorial Design. Dr. Mohammad Abuhaiba 1

MEMORIAL UNIVERSITY OF NEWFOUNDLAND DEPARTMENT OF MATHEMATICS AND STATISTICS FINAL EXAM - STATISTICS FALL 1999

Contents. 2 2 factorial design 4

Chapter 6 The 2 k Factorial Design Solutions

Definitions of terms and examples. Experimental Design. Sampling versus experiments. For each experimental unit, measures of the variables of

STATISTICS 368 AN EXPERIMENT IN AIRCRAFT PRODUCTION Christopher Wiens & Douglas Wiens

Solution to Final Exam

STAT22200 Spring 2014 Chapter 8A

sociology 362 regression

Design of Engineering Experiments Part 5 The 2 k Factorial Design

Lecture 11: Nested and Split-Plot Designs

Design and Analysis of Multi-Factored Experiments

G. Nested Designs. 1 Introduction. 2 Two-Way Nested Designs (Balanced Cases) 1.1 Definition (Nested Factors) 1.2 Notation. 1.3 Example. 2.

What If There Are More Than. Two Factor Levels?

Chapter 4 Randomized Blocks, Latin Squares, and Related Designs Solutions

2-way analysis of variance

Analysis of Variance

. Example: For 3 factors, sse = (y ijkt. " y ijk

SMA 6304 / MIT / MIT Manufacturing Systems. Lecture 10: Data and Regression Analysis. Lecturer: Prof. Duane S. Boning

Chapter 4 Experiments with Blocking Factors

Chapter 4: Randomized Blocks and Latin Squares

Part 5 Introduction to Factorials

Lecture 10: Factorial Designs with Random Factors

Ch 3: Multiple Linear Regression

Blocks are formed by grouping EUs in what way? How are experimental units randomized to treatments?

Chapter 6 The 2 k Factorial Design Solutions

2.830 Homework #6. April 2, 2009

Contents. TAMS38 - Lecture 10 Response surface. Lecturer: Jolanta Pielaszkiewicz. Response surface 3. Response surface, cont. 4

Reference: Chapter 6 of Montgomery(8e) Maghsoodloo

3. Design Experiments and Variance Analysis

Outline Topic 21 - Two Factor ANOVA

Note: The problem numbering below may not reflect actual numbering in DGE.

Stat 6640 Solution to Midterm #2

STAT Final Practice Problems

RESPONSE SURFACE MODELLING, RSM

FACTORIAL DESIGNS and NESTED DESIGNS

Topic 29: Three-Way ANOVA

6 Designs with Split Plots

Power & Sample Size Calculation

Confidence Interval for the mean response

Reference: Chapter 13 of Montgomery (8e)

DESAIN EKSPERIMEN Analysis of Variances (ANOVA) Semester Genap 2017/2018 Jurusan Teknik Industri Universitas Brawijaya

y = µj n + β 1 b β b b b + α 1 t α a t a + e

Regression Analysis. Simple Regression Multivariate Regression Stepwise Regression Replication and Prediction Error EE290H F05

Analysis of Variance. Read Chapter 14 and Sections to review one-way ANOVA.

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Fractional Factorial Designs

Stat 579: Generalized Linear Models and Extensions

2 k, 2 k r and 2 k-p Factorial Designs

Chapter 4 Randomized Blocks, Latin Squares, and Related Designs Solutions

Two-factor studies. STAT 525 Chapter 19 and 20. Professor Olga Vitek

SIZE = Vehicle size: 1 small, 2 medium, 3 large. SIDE : 1 right side of car, 2 left side of car

Ch 2: Simple Linear Regression

Lecture 20: Linear model, the LSE, and UMVUE

Process/product optimization using design of experiments and response surface methodology

iron retention (log) high Fe2+ medium Fe2+ high Fe3+ medium Fe3+ low Fe2+ low Fe3+ 2 Two-way ANOVA

Unit 6: Fractional Factorial Experiments at Three Levels

Third European DOE User Meeting, Luzern 2010

Factorial and Unbalanced Analysis of Variance

Y it = µ + τ i + ε it ; ε it ~ N(0, σ 2 ) (1)

Institutionen för matematik och matematisk statistik Umeå universitet November 7, Inlämningsuppgift 3. Mariam Shirdel

Multiple comparisons - subsequent inferences for two-way ANOVA

In a one-way ANOVA, the total sums of squares among observations is partitioned into two components: Sums of squares represent:

One-Way Analysis of Variance. With regression, we related two quantitative, typically continuous variables.

Increasing precision by partitioning the error sum of squares: Blocking: SSE (CRD) à SSB + SSE (RCBD) Contrasts: SST à (t 1) orthogonal contrasts

EXST 7015 Fall 2014 Lab 11: Randomized Block Design and Nested Design

Table 1: Fish Biomass data set on 26 streams

Factorial BG ANOVA. Psy 420 Ainsworth

Unbalanced Designs & Quasi F-Ratios

Chap The McGraw-Hill Companies, Inc. All rights reserved.

Reference: Chapter 14 of Montgomery (8e)

Confidence Intervals, Testing and ANOVA Summary

Chapter 9 Other Topics on Factorial and Fractional Factorial Designs

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008

Mixed-effect model analysis of ISTA GMO Proficiency Tests

Multi-Factor Experiments

Transcription:

Chapter 5 Introduction to Factorial Designs

5. Basic Definitions and Principles Stud the effects of two or more factors. Factorial designs Crossed: factors are arranged in a factorial design Main effect: the change in response produced b a change in the level of the factor 2

Definition of a factor effect: The change in the mean response when the factor is changed from low to high 40+ 52 20+ 30 A + 2 A A 2 2 30+ 52 20+ 40 B + B B 2 2 52+ 20 30+ 40 AB 2 2 3

50+ 2 20+ 40 A + A A 2 2 40+ 2 20+ 50 B + 9 B B 2 2 2+ 20 40+ 50 AB 29 2 2 4

Regression Model & The Associated Response Surface β + β x + β x 0 2 2 + β x x + ε 2 2 The least squares fit is ˆ 35.5+ 0.5x + 5.5x + 0.5x x 2 2 35.5+ 0.5x + 5.5x 2 5

The Effect of Interaction on the Response Surface Suppose that we add an interaction term to the model: ˆ 35.5+ 0.5x + 5.5x + 8x x 2 2 Interaction is actuall a form of curvature 6

When an interaction is large, the corresponding main effects have little practical meaning. A significant interaction will often mask the significance of main effects. 7

5.2 The Advantage of Factorials One-factor-at-a-time desgin Compute the main effects of factors A: A + B - - A - B - B: A - B - - A - B + Total number of experiments: 6 Interaction effects A + B -, A - B + > A - B - > A + B + is better??? 8

5.3 The Two-Factor Factorial Design 5.3. An Example a levels for factor A, b levels for factor B and n replicates Design a batter: the plate materials (3 levels) v.s. temperatures (3 levels), and n 4: 3 2 factorial design Two questions: What effects do material tpe and temperature have on the life of the batter? Is there a choice of material that would give uniforml long life regardless of temperature? 9

The data for the Batter Design: 0

Completel randomized design: a levels of factor A, b levels of factor B, n replicates

Statistical (effects) model: i,2,..., a ijk µ + τi + β j+ ( τβ ) ij + εijk j, 2,..., b k,2,..., n µ is an overall mean, τ i is the effect of the ith level of the row factor A, β j is the effect of the jth column of column factor B and (τ β) ij is the interaction between τ i and β j. Testing hpotheses: H 0 : τ L τ a 0 v.s. H : at least oneτ i 0 H 0 : β L β b 0 v.s. H : at least oneβ j 0 H 0 : ( τβ ) ij 0 i, j v.s. H : at least one ( τβ) ij 2 0

5.3.2 Statistical Analsis of the Fixed Effects Model i... j. ij.... b j k i k n k a a n b n ijk n ijk ijk i j k ijk i... j. ij.... i.. bn ij. n. j. an... abn 3

a b n a b 2 2 2 ( ijk... ) bn ( i..... ) + an (. j.... ) i j k i j a b a b n 2 2 ( ij. i... j....) ( ijk ij. ) i j i j k + n + + SS SS + SS + SS + SS T A B AB E df breakdown: abn a + b + ( a )( b ) + ab( n ) 4

5 Mean squares 2 2 2 2 2 2 2 ) ) ( ( ) ( ) )( ( ) ( ) ) )( ( ( ) ( )) /( ( ) ( )) /( ( ) ( σ τβ σ β σ τ σ + + + n ab SS E MS E b a n b a SS E MS E b an b SS E MS E a bn a SS E MS E E E a i b j ij AB AB b j j B B a i i A A

The ANOVA table: 6

7

Example 5. Response: Life ANOVA for Selected Factorial Model Analsis of variance table [Partial sum of squares] Sum of Mean F Source Squares DF Square Value Prob > F Model 5946.22 8 7427.03.00 < 0.000 A 0683.72 2 534.86 7.9 0.0020 B 398.72 2 9559.36 28.97 < 0.000 AB 963.78 4 2403.44 3.56 0.086 Pure E 8230.75 27 675.2 C Total 77646.97 35 Std. Dev. 25.98 R-Squared 0.7652 Mean 05.53 Adj R-Squared 0.6956 C.V. 24.62 Pred R-Squared 0.5826 PRESS 3240.22 Adeq Precision 8.78 8

DESIGN-EXPERT Plot Life X B: T em perature Y A: M ateri al 88 Interactio n Graph A: Material A A A2 A2 A3 A3 46 Life 04 2 62 2 2 20 5 70 25 B: Tem perature 9

Multiple Comparisons: Use the methods in Chapter 3. Since the interaction is significant, fix the factor B at a specific level and appl Turke s test to the means of factor A at this level. See Page 74 Compare all ab cells means to determine which one differ significantl 20

5.3.3 Model Adequac Checking Residual analsis: e ijk ˆ ijk ijk ijk ij DE SIG N-EX P ERT Plo t Life Normal plot of residuals DE SIGN-EXPER T Plo t L ife Res iduals vs. P redicted 45.25 99 N orm al % probabilit 95 90 80 70 50 30 20 0 5 Residuals 8.75-7.75-34.25-60.75-60.75-34.25-7.75 8.75 45.25 49.50 76.06 02.62 29.9 55.75 Res idual Pre dicte d 2

DESIG N-E XPERT Pl ot L ife Residuals vs. Run 45.25 8.75 Re sid uals - 7.75-34.25-60.75 6 6 2 26 3 36 R un N um ber 22

DE SIG N-EX P ERT Plo t Life Residuals vs. Material DES IGN-E XPE RT P l ot L i fe Residuals vs. Temp erature 45.25 45.25 8.75 8.75 R es iduals -7.75 R esid ua ls -7.75-34.25-34.25-60.75-60.75 2 3 2 3 Material Tem pera tu re 23

5.3.4 Estimating the Model Parameters The model is µ + τ + β + ( τβ) + ε ijk The normal equations: µ : abnµ + bn Constraints: i j ij a i i τ : bnµ + bnτ + β : anµ + n ( τβ ) i a : nµ + nτ a i τ i τ i + i i n + + b j nβ b an β j j j anβ + j j b + + β n n j b + j a i n( τβ) ij ij n a ( τβ) ij ijk i j ij ( τβ) ij b ( τβ) i j ij ( τβ) ( τβ) 0 τ 0, 0, i β j ij j a i b j ij 24

Estimations: ˆ µ ˆ τ i ˆ β j i j ( τβ) ij ij i j + The fitted value: ˆ ijk ( τβ) ij ˆ µ + ˆ τ i + ˆ β j + ij Choice of sample size: Use OC curves to choose the proper sample size. 25

Consider a two-factor model without interaction: Table 5.8 The fitted values: ˆijk i + j 26

One observation per cell: The error variance is not estimable because the two-factor interaction and the error can not be separated. Assume no interaction. (Table 5.9) Tuke (949): assume (τβ) ij rτ i β j (Page 83) Example 5.2 27

5.4 The General Factorial Design More than two factors: a levels of factor A, b levels of factor B, c levels of factor C,, and n replicates. Total abc n observations. For a fixed effects model, test statistics for each main effect and interaction ma be constructed b dividing the corresponding mean square for effect or interaction b the mean square error. 28

Degree of freedom: Main effect: # of levels Interaction: the product of the # of degrees of freedom associated with the individual components of the interaction. The three factor analsis of variance model: µ + τ + β + γ + ( τβ) ijkl + ( ) ik + ( ) jk + ( ) The ANOVA table (see Table 5.2) Computing formulas for the sums of squares (see Page 86) Example 5.3 i τγ j k βγ ij τβγ ijk + ε ijkl 29

30

Example 5.3: Three factors: the percent carbonation (A), the operating pressure (B); the line speed (C) 3

32

5.5 Fitting Response Curves and Surfaces An equation relates the response () to the factor (x). Useful for interpolation. Linear regression methods Example 5.4 Stud how temperatures affects the batter life Hierarch principle 33

Involve both quantitative and qualitative factors This can be accounted for in the analsis to produce regression models for the quantitative factors at each level (or combination of levels) of the qualitative factors A Material tpe B Linear effect of Temperature B 2 Quadratic effect of Temperature AB Material tpe Temp Linear AB 2 Material tpe - Temp Quad B 3 Cubic effect of Temperature (Aliased) 34

35

36

37

5.6 Blocking in a Factorial Design A nuisance factor: blocking A single replicate of a complete factorial experiment is run within each block. Model: ijk µ + τ + β + ( τβ ) + δ + ε No interaction between blocks and treatments ANOVA table (Table 5.20) i j ij k ijk 38

39

Example 5.6: Two factors: ground clutter and filter tpe Nuisance factor: operator 40

Two randomization restrictions: Latin square design An example in Page 200. Model: ijkl µ + α + τ + β + ( τβ) + δ + ε Tables 5.23 and 5.24 i j k jk l ijkl 4