Software

This chapter describes and compares suitable software for the analysis of basic and advanced discrete choice models. Software packages are classified into proprietary and non-proprietary, according to the operating system required and modelling capabilities. Abilities of both selected commercial (Stata, SAS and Latent Gold, e.g.) and open-source packages (Biogeme and R-libraries) are considered. Finally, some user-written estimation packages for Gauss, Matlab, R and Stata are presented.

Collecting the Data

This chapter discusses aspects related to data collection. It focuses, firstly, on sampling issues and, secondly, on the survey mode. Sampling issues include sample size and the type of sampling that enable precise estimates to be obtained. Regarding the survey mode, discrete choice experiments can be implemented by mail, telephone, face-to-face or web surveys. Each of these survey modes has its advantages and shortcomings. They are described and compared in the course of this chapter, addressing an important decision in the planning process of a discrete choice experiment.

Experimental Design

This chapter covers various issues related to the experimental design, a statistical technique at the core of a discrete choice experiment. Specifically, it focuses on the dimensionality of a choice experiment and the statistical techniques used to allocate attribute levels to choice tasks. Among others, the pros and cons of orthogonal designs, optimal orthogonal in the differences designs as well as efficient designs are addressed. The last section shows how a simulation exercise can help to test the appropriateness of the experimental design.

Theoretical Background

This chapter starts by briefly presenting the theoretical background of welfare economics and introducing key aspects such as the indirect utility function, the expenditure function, or the concepts of compensating surplus or equivalent surplus. Next, it draws attention to willingness to pay and willingness to accept, essential measures in environmental valuation. Finally, the chapter summarises the basic mathematical notation of the random utility maximisation models used throughout the book.

Econometric Modelling: Basics

This chapter addresses basic topics related to choice data analysis. It starts by describing the coding of attribute levels and choosing the functional form of the attributes in the utility function. Next, it focuses on econometric models with special attention devoted to the random parameter mixed logit model. In this context, the chapter compares different coefficient distributions to be used, addresses specifics of the cost attribute coefficient and it pays attention to potential correlations between random coefficients. Finally, topics related to the estimation procedure such as assuring its convergence or random draws are discussed.

Developing the Questionnaire

This chapter outlines the essential topics for developing and testing a questionnaire for a discrete choice experiment survey. It addresses issues such as the description of the environmental good, pretesting of the survey, incentive compatibility, consequentiality or mitigation of hypothetical bias. For the latter, cheap talk scripts, opt-out reminders or an oath script are discussed. Moreover, the use of instructional choice sets, the identification of protest responses and strategic bidders are considered. Finally, issues related to the payment vehicle and the cost vector design are the subject of this section.

Calculating Marginal and Non-marginal Welfare Measures

This chapter focuses on the calculation of marginal and non-marginal welfare measures. It outlines how the calculation of welfare measures is related to the specified model and the assumptions underlying that model. It further describes how the calculation of these measures is affected by the inclusion of preference heterogeneity, including the incorporation of interaction terms to capture observed preference heterogeneity or random parameters to capture unobserved preference heterogeneity. Finally, it discusses how these measures can be aggregated and compared.

Validity and Reliability

This chapter concerns different aspects of validity and reliability of a discrete choice experiment. Firstly, it focuses on three essential concepts for assessing the validity of the welfare estimates obtained in the choice experiment, namely content, construct and criterion validity. Secondly, it discusses how the reliability of the recorded choices can be assessed. It then discusses issues related to model comparison and selection. Finally, it addresses prediction in discrete choice models as a way to assess the quality of a model.

Econometric Modelling: Extensions

This chapter is devoted to advanced issues of econometric modelling. The topics covered are, among others, models in willingness to pay space, the meaning of scale heterogeneity in discrete choice models and the application of various information processing rules such as random regret minimisation or attribute non-attendance. Other topics are anchoring and learning effects when respondents move through a sequence of choice tasks as well as different information processing strategies such as lexicographic preferences or choices based on elimination-by-aspects.