INSEAD Library

Includes bibliographical references and index

Digitized

The Contruction of Optima Stated Choice Experiments Theory and Methods Contents List of Tables Preface 1 Typical Stated Choice Experiments 1.1 Definitions 1.2 Binary Response Experiments 1.3 Forced Choice Experiments 1.4 The "None" Option 1.5 A Common Base Option 1.6 Avoiding Particular Level Combinations 1.6.1 Unrealistic Treatment Combinations 1.6.2 Dominating Options 1.7 Other Issues 1.7.1 Other Designs 1.7.2 Non-mathematical Issues for Stated Preference Choice Experiments 1.7.3 Publ shed Studies 1.8 Concluding Remarks 11 12 13 xi xvii 1 2 3 5 7 8 9 9 10 11 11 2 Factoriel Designs 2.1 Complete Factorial Designs 2.1.1 2k Designs 2.1.2 3k Designs 15 16 16 19 24 25 27 27 33 37 39 41 41 43 44 46 52 53 55 56 56 57 58 58 59 60 61 62 65 67 70 79 79 79 82 82 83 83 84 88 2.1.3 Asymmetric Designs 2.1.4 Exercises 2.2 Regular Fractional Factorial Designs 2.2.1 Two-Level Fractions 2.2.2 Three-Level Fractions 2.2.3 A Brief Introduction to Finite Fields 2.2.4 Fractions for Prime-Power Levels 2.2.5 Exercises 2.3 Irregular Fractions 2.3.1 Two Constructions for Symmetric OAs 2.3.2 Constructing OA[2k; 2k1, 4k2; 4] 2.3.3 Obtaining New Arrays from OId 2.3.4 Exercises 2.4 Other Useful Designs 2.5 Tables of Fractional Factorial Designs and Orthogonal Arrays 2.5.1 Exercises 2.6 References and Comments 3 The MNL Model and Comparing Designs 3.1 Utility and Choice Probabilities 3.1.1 Utility 3.1.2 Choice Probabilities 3.2 The Bradley­Terry Model 3.2.1 The Likelihood Function 3.2.2 Maximum Likelihood Estimation 3.2.3 Convergence 3.2.4 Properties of the MLEs 3.2.5 Representing Options Using k Attributes 3.2.6 Exercises 3.3 The MNL Model for Choice Sets of Any Size 3.3.1 Choice Sets of Any Size 3.3.2 Representing Options Using k Attributes 3.3.3 The Assumption of Independence from Irrelevant Alternatives 3.3.4 Exercises 3.4 Comparing Designs 3.4.1 Using Variance Properties to Compare Designs 3.4.2 Structural Properties 3.4.3 Exercises 3.5 References and Comments 4 Paired Comparison Designs for Binary Attributes 4.1 Optimal Pairs from the Complete Factorial 4.1.1 The Derivation of the A Matrix 4.1.2 Calculation of the Relevant Contrast Matrices 4.1.3 The Mode! for Main Effects Only 4.1.4 The Model for Main Effects and Two-factor Interactions 4.1.5 Exercises 4.2 Small Optimal and Near-optimal Designs for Pairs 4.2.1 The Derivation of the A Matrix 4.2.2 The Model for Main Effects Only 4.2.3 The Model for Main Effects and Two-Factor Interactions 4.2.4 Dominating Options 4.2.5 Exercises 4.3 References and Comments 5 Larger Choice Set Sizes for Binary Attributes 5.1 Optimal Designs from the Complete Factorial 5.1.1 Difference Vectors 5.1.2 The Derivation of the A Matrix 5.1.3 The Model for Main Effects Only 5.1.4 The Model for Main Effects and Two-Factor Interactions 5.1.5 Exercises 5.2 Small Optimal and Near-Optimal Designs for Larger Choice Set Sizes 5.2.1 The Model for Main Effects Only 5.2.2 The Mode! for Main Effects and Two-Factor Interactions 5.2.3 Dominating Options 5.2.4 Exercises 5.3 References and Comments 6 Designs for Asymmetric Attributes 6.1 Difference Vectors 6.1.1 Exercises 6.2 The Derivation of the Information Matrix A 6.2.1 Exercises 6.3 The Model for Main Effects Only 6.3.1 Exercises 6.4 Constructing Optimal Designs for Main Effects Only 6.4.1 Exercises 91 92 95 95 97 99 100 105 117 118 118 119 121 133 134 134 137 138 138 143 147 152 159 159 160 163 164 165 165 167 169 173 174 180 180 189 189 197 6.5 The Model for Main Effects and Two-Factor Interactions 6.5.1 Exercises 6.6 References and Comments Appendix 6. A.1 Optimal Designs for m = 2 and k = 2 6. A.2 Optimal Designs for m = 2 and k = 3 6. A.3 Optimal Designs for m = 2 and k = 4 6. A.4 Optimal Designs for m = 2 and k = 5 6. A.5 Optimal Designs for m = 3 and k = 2 6. A.6 Optimal Designs for m = 4 and k = 2 6. A.7 Optimal Designs for Symmetric Attributes for m = 2 7 Various Topics 7.1 Optimal Stated Choice Experiments when All Choice Sets Contain a Specific Option 7.1.1 Choice Experiments with a None Option 7.1.2 Optimal Binary Response Experiments 7.1.3 Common Base Option 7.1.4 Common Base and None Option 7.2 Optimal Choice Set Size 7.2.1 Main Effects Only for Asymmetric Attributes 7.2.2 Main Effects and Two-Factor Interactions for Binary Attributes 7.2.3 Choice Experiments with Choice Sets of Varions Sizes 7.2.4 Concluding Comments on Choice Set Size 7.3 Partial Profiles 7.4 Choice Experiments Using Prior Point Estimates 7.5 References and Comments 8 Practical Techniques for Constructing Choice Experiments 8.1 Small Near-Optimal Designs for Main Effects Only 8.1.1 Smaller Designs for Examples in Section 6.4 8.1.2 Getting a Starting Design 8.1.3 More on Choosing Generators 8.2 Small Near-Optimal Designs for Main Effects Plus Two-Factor Interactions 8.2.1 Getting a Starting Design 8.2.2 Designs for Two-Level Attributes 8.2.3 Designs for Attributes with More than Two Levels 8.2.4 Designs for Main Effects plus Some Two-Factor Interactions 8.3 Other Strategies for Constructing Choice Experiments 197 209 210 211 212 213 217 220 221 223 225 227 228 228 233 234 236 237 237 240 242 243 243 245 246 249 251 251 256 264 269 269 271 272 276 279 8.4 Comparison of Strategies 8.5 References and Comments Appendix 8. A.I Near-Optimal Choice Sets for l1 = l2 = 2, l3 = l4 = l5 = l6 = 4, l7 = 8, l8 = 36, and m = 3 for Main Effects Only 8. A.2 Near-Optimal Choice Sets for l1 = l2 = l3 = l4 = l5 = 4, l6 = 2, l7 = l8 = 8, l9 = 24, and m. = 3 for Main Effects Only Bibliography Index 291 293 294 294 296 301 309