Tversky and Kahneman (1981) discussed how multiattribute choices may b
e framed in terms of alternative psychological accounts established fo
r specific problems. They identified three levels of account: minimal,
topical and comprehensive. The role of these mental accounts is inves
tigated in two studies of Tversky and Kahneman's Jacket and Calculator
problem. In the first, systematic variation of jacket and calculator
prices showed that the latter had a large effect on choice, and the fo
rmer a small but significant effect. It is concluded that: (1) the eff
ect of the calculator price supports Tversky and Kahneman's view that
people frame simple choices in terms of specific, topical accounts; an
d (2) the jacket price effect supports the view that more comprehensiv
e mental accounts also influence evaluation and choice. In the second
study, think aloud evidence is presented which supports these conclusi
ons and indicates that the level of account adopted may be contingent
on characteristics of the choice problem.