COURSE PROGRAM
Each week consists of a 3.5 hour live online session, plus a series of pre-recorded videos that explain key concepts/theories and demonstrate widely used software tools in choice modelling. Students are also given exercises to complete in their own time each week. It is strongly recommended to reserve half a day each week to watch these videos, practice with software, and work on exercises.
Delegates will receive installation instructions and temporary licences (where needed)
for the following software tools:
• ​Apollo for estimating choice models
• ​Ngene for generating choice experimental designs
• ​SurveyEngine for creating online questionnaires and choice experiments
WEEK 1: INTRODUCTION TO CHOICE MODELLING
-
Introduction to key concepts
-
Study plan: a guide
-
Case studies: study plan
-
Decision rules
-
The multinomial logit model
WEEK 2: DATA COLLECTION
-
Stated and revealed preference data
-
Data collection: a guide
-
Creating questionnaires
-
Case study: revealed preference data collection
-
Case study: stated preference data collection
WEEK 3: MODEL BUILDING, SPECIFICATION, ESTIMATION AND REFINEMENT
-
Model specification and estimation: a guide
-
Formulating utility functions
-
Parameter estimates
-
Model comparison
-
Case study: model specification and estimation
WEEK 4: DESIGNING CHOICE EXPERIMENTS
-
Experimental design: a guide
-
Orthogonal designs
-
Efficient designs
-
Advanced designs
-
Case study: experimental design
WEEK 5: INTERPRETATION OF OUTPUTS AND MODEL APPLICATION
-
Relative importance of attributes
-
Marginal rates of substitution, willingness-to-pay
-
Predictions and elasticities
-
Determining precision of derived measures of interest
-
Case study: model interpretation and application
WEEK 6: ACCOUNTING FOR RANDOM PREFERENCE HETEROGENEITY
-
Heterogeneity: an overview
-
Latent class analysis
-
Mixed logit
-
Interpreting model results
-
Case study: random preference heterogeneity
WEEK 7: OTHER CHOICE MODELLING CONSIDERATIONS
-
Behavioural artefacts in choice models
-
Choice models with flexible error structures
-
Heteroskedastic choice models
-
Hybrid choice models