The Who and Why of Context Dependent Choice




The use of context to influence choice has been a subject of increasing interest to academics studying choice, and has been apparently useful to firms.  In this paper we focus on the underlying cognitive processes of the two most well known effects, the attraction effect and the compromise effect.

Process predictions of three classes of existing accounts were derived. The first class of existing accounts attribute context effects to relative valuation of product attributes (e.g, Relative Advantage Model, Tversky and Simonson, 1993). The second class of accounts emphasize the dynamics of the choice process and attributes context effects to lateral inhibition in a sequential sampling process (e.g., Decision Field Theory, Roe et al. 2001). The third class of accounts are dynamic models of attribute distortion, which include insights from predecisional distortion of information (e.g., Brownstein, 2003; Russo et al., 1998), and which we label Decision by Distortion.

We tested process predictions of these accounts with an online process tracing tool (  on data from 374 'ordinary' US citizens. The data not only demonstrate the impact of presentation order and individual differences on the strength of compromise and attraction effects, but also identifies part of the underlying cognitive processes showing that alternative-focused processing consistent with Decision by Distortion prevails in the context effects studied.

We present new methods for presenting these process data, and employ multi-level models to capture heterogeneity in the strategy used and choices made, and to portray how processes change over time.
Contact information:
Dr. S. Puntoni