Modeling Single Individuals: The Journey from Psych Lab to the App Store

Edward Elgar
Publication Type:
Choice Modelling: The State of the Art and the State of Practice, 2013, 352, pp. 1 - 47
Issue Date:
Filename Description Size
Thumbnail2012000758OK.pdf5.19 MB
Adobe PDF
Full metadata record
The purpose of this chapter is to develop the ideas discussed in my keynote address at the ICMC in Leeds in 2010 in more detail. In particular, I will review the choice-based measurement theory and measurement methods known as Best-Worst Scaling (BWS). I will introduce BWS Cases 1 and 2, and extend the ideas from Cases 1 and 2 to Case 3 (aka DECs). En route I will briefly touch on error variance differences and stability of preferences over time. The chapter ends with some conclusions and potential future research directions. Individual level models (ILMs) are not new. Indeed, psychologists have been interested in ILMs since Thurstone (1927). However, what is surprising is that unlike marketing, transport and applied economics, there historically was little interest in the formal application of sophisticated statistical design theory and methods to designing choice experiments. Instead, psychologists tended to rely on paired comparisons and/or ad hoc sets of triples, quadruples, etc. This is particularly surprising in so far as psychologists are much more interested in the underlying processes involved in choice behaviours than other disciplines; hence, one would think that many would have seen the obvious connection between experimental design and ones ability to both test and draw inferences from studies of choice behaviour.
Please use this identifier to cite or link to this item: