Abstract
The concepts of reference dependence, loss aversion, diminishing sensitivity, mental accounting, and response mode compatibility were used to manipulate task and contextual features predicted to influence the construction of preferences. Three experiments used riskless matching tasks to elicit from undergraduate business students information about their relative preferences for two attributes: affordability and convenience. Subjects were presented with two options that differed on these attributes. Depending on the experiment and treatment condition, they supplied either a missing price (affordability) or a missing driving time (convenience) that would equate the options. Invariance violations predicted by the joint effect of reference dependence and loss aversion, or by response mode compatibility, were observed in all three experiments and confirmed by replications. The different but formally equivalent descriptions and procedures yielded estimates on the affordability versus convenience trade-off that ranged from $10.50 to $61.28 per hour. The experiments also tested for the joint effect of mental accounting and diminishing sensitivity. Contrary to predictions, diminishing sensitivity was observed only for price matching. Possible explanations for the failure to find diminishing sensitivity in time matching and its implications are discussed.