Consumer Buying Intentions and Purchase Probability: An Experiment in Survey Design

Abstract Surveys of consumer intentions to buy are inefficient predictors of purchase rates because they do not provide accurate estimates of mean purchase probability. This is a consequence of the fact that intentions surveys cannot detect movements in mean probability among nonintenders, who account for the bulk of actual purchases and for most of the time-series variance in purchase rates. Comparison of predictions from alternative surveys, one of subjective purchase probabilities and the other of buying intentions, indicates that purchase probabilities explain about twice as much of the cross-section variance in automobile purchase rates as buying intentions. Similar but not quite so conclusive differences are obtained from analysis of selected household durables. The probability variable predicts more accurately than the intentions variable largely because it divides nonintenders, and those who report that they “don't know” about their buying intentions, into subgroups with systematically different purchase rates.