ABSTRACT
Louviere and Woodworth (1990) and Finn and Louviere (1992) developed a discrete choice task in which a person is asked to select both the best and the worst option in an available (sub)set of choice alternatives; they also presented and evaluated a probabilistic model of their data. There is growing interest in such best-worst scaling, with three recent empirical applications receiving “best paper at conference” awards (Chrzan & Golovashkina, 2006; Cohen, 2003; Cohen & Neira, 2004). There is also an increasing body of evidence showing a number of advantages of best-worst scaling over more traditional choice methods, such as those that ask a person to choose the best option; choose the worst option; rank the options; or rate the options. For instance: (1) a single pair of best-worst choices contains a great deal of information about the person’s ranking of options (e.g., if there are 3 items in a set, one obtains the entire ranking of that set; if there are 4 items in a set, one obtains information on the implied best option in 9 of the 11 possible non-empty, non-singleton subsets;1 and if there are 5 items in a set, one obtains information on the implied best option in 18 of the 26 possible non-empty, non-singleton subsets); (2) best-worst tasks take advantage of a person’s propensity to identify and respond more consistently to extreme options; (3) best-worst tasks seem to be easy for people; (4) best-worst data show improved discrimination between items relative to rating scales; and (5) the values (utilities) of individual decision-makers can be measured by this method. Detailed material and references on these points can be found in Chrzan and Golovashkina (2006), Jaeger, Jorgensen, Aaslyng, and Bredie (2008), and Louviere et al. (2008).
