Simple preference intensity comparisons

Abstract

We propose a general model of simple preference intensity comparisons. The model encompasses those that belong to the utility-difference class, has transparent behavioural underpinnings and features purely ordinal uniqueness properties. Its empirical content is characterized by an easily testable condition on dual behavioural data that feature choices and additional observables with intensity-revealing potential that are often elicited in experimental/empirical work, such as survey ratings, response times or willingness to pay. Compared to the benchmark model of pseudo-cardinal utility differences, the proposed model provides a significantly better fit of choice and intensity-ratings data that were collected in a lab experiment and analysed in a novel non-parametric way.

Stochastic Choice and Imperfect Judgments of Line Lengths: What Is Hiding in the Noise?

“…economics literature contains research on the implications of the di¤erences in the utility-of strength of preferences-on stochastic choice (Mosteller and Nogee, 1951;Debreu, 1958). Further, there appears to be renewed interest in the topic (Ballinger and Wilcox, 1997;Blavatskyy and Pogrebna, 2010;Gerasimou, 2021;Alós-Ferrer and Garagnani, 2021,2022a,2022b. The use of line lengths as a proxy for utility has the advantage that we can make small changes in the di¤erences in the lines and observe the e¤ect on optimal choice.…”

Intensity-Efficient Allocations

“…This can be thought of as a binary relation over pairs in X × X or as a quaternary relation over elements of X. We assume that ˙ i is representable by a preference intensity function (Gerasimou, 2020), i.e. a mapping s…”

“…for some function u i : X → R + . This special case allows one to write s i (a, b) Gerasimou (2020) and references therein). As is well-known, when such a utility-difference representation exists it is unique neither up to a positive affine transformation nor up to an arbitrary strictly increasing transformation.…”