Obtaining Purchase Probability Data via Telephone Surveys: A Preliminary Test of Two Techniques

Abstract

This paper examines two alternative methods of using the Juster Scale in telephone surveys, and the effects on response rate, respondent comprehension, and accuracy of prediction. A random sample of 80 people was selected from the Palmerston North telephone directory and randomly allocated to one of two treatment groups. Respondents in Group 1 (Juster Scale) were mailed a letter with the Juster Scale printed on the back. The letter informed them that they would be contacted by telephone within the next two days. There was no preliminary mail contact with Group 2 (Verbal Probability Scale). Two days after the mailout to Group 1, all respondents were surveyed by telephone. In the first telephone interview, respondents were asked to provide a probability of purchase for nine food items, using either the pre-mailed Juster Scale (Group 1) or a verbal eleven-point purchase probability scale (Group 2). A followup telephone call was made to each respondent one month after the first call, to determine actual purchases of the products involved in the study. Both the mailed Juster Scale and the Verbal Probability Scale produced acceptable estimates, although the Verbal Probability Scale consistently produced more accurate results. The results of this study suggest that the Verbal Probability Scale is a viable method for obtaining purchase probabilities in telephone surveys.