The Juster 11-Point Probability Scale: A predictive measure of future intentions
In 1966, Professor F. Thomas Juster argued that since verbal intentions are simply disguised probability statements, then why not directly capture the probabilities themselves as measured by the respondents.
Justers’ 11 point probability scale (see inset) can be used to produce estimates of the average probability that a population will do something by a future time. Since what is being measured is a probability, the mean response estimates the proportion of the population that will perform the behaviour at issue.
An example is given by the question, “On a scale of 0 – 10 where 0 indicates no chance and 10 indicates certainty, what is the chance that you will change your primary bank in the next 12 months?” If then the average response is 3.2, this translates to 32% of the population intend to switch banks.
The Juster scale in its many applications has been found to be superior as a predictive measure of future purchase behaviour than other intentions scales. The distribution of responses, however, has been found to affect the predictive accuracy of the scale. Not surprisingly, the greater the variation in responses, the less accurate the predictions.
Studies have shown that purchase probabilities can be over or under estimated by the Juster scale, but on average, it is the most consistent in accurately predicting actual purchase rates. There are important issues to be considered in the administration of the Juster Scale that have been found to contribute to variation in its effectiveness. These include unfamiliarity of the respondent with new products, training of the administrator and differences in age and education level of respondents.
The Juster scale has also been successfully used to predict respondent behaviour outside the typical consumption behaviour realm, which includes being applied to telephone surveys, fast moving consumer goods, self-completion questionnaires, services, brands and customer loyalty. One example of such an extension involved predicting the percentage of a given population of adults currently at home looking after children, who will take up paid employment in the next year. At an aggregate level in this example, the Juster Scale mean was 1.9 indicating that a predicted 19% of respondents would find paid work in the next year. When actual behaviour was measured in the following year, it was found that indeed, 19% of these respondents were in paid employment.
Juster’s 11-point probability scale
