Methods for Forecasting from Intentions Data

Abstract

Intentions are measures of individuals’ plans, goals, or expectations about what they will do in the future and are often used to forecast what people will do in the future. I develop nine principles from past studies of the predictive accuracy of intentions data that concern how to measure intentions, how to best use intentions to forecast, and reasons why forecasters should be cautious when using intentions. The principles relating to intentions measurement state that intentions should be measured using probability scales and that respondents should be instructed to focus on their own individual characteristics when responding to intentions questions. The principles relating to using intentions to forecast behavior state that intentions need to be adjusted to remove biases, that respondents should be segmented prior to adjusting intentions, and that intentions can be used to develop best- and worst-case forecasts. Further, one principle states that more reliance should be placed on predictions from intentions for behaviors in which respondents have previously participated. Finally, the principles relating to why researchers should be cautious when using intentions data state that researchers should be aware that measuring intentions can change behavior and that respondents who recall the time of their last purchase inaccurately may make biased predictions of their future purchases.