Preference, projection, and packing: Support theory models of judgments of others' preferences

People frequently need to predict the preferences of others. Such intuitive predictions often show social projection, in which one's own preference for an option increases its perceived popularity among others. We use support theory to model social projection in the prediction of preferences, and in particular interactions between social projection and description-dependence. Preferred options are predicted to have consistently high salience, and therefore should be less susceptible to description variations, such as unpacking, which normally affect option salience. This preference salience premise implies an interaction between social projection and option description, with reduced unpacking effects for hypotheses including preferred options, or equivalently, with reduced social projection when less-liked alternatives are unpacked. Support theory models accommodating different preference-dependent unpacking effects are tested. These models distinguish two substantial contributors to social projection effects: (a) greater evidence recruited for preferred options and (b) greater discounting of packed less-preferred options.