Digit preference, recall bias, and nonresponse bias in self reports of angling participation

Abstract

Explored in this study is the role of digit preference as a response style that helps explain the problems of recall and nonresponse bias in self‐report surveys of recreation participation. Digit preference is a tendency to guess participation by rounding estimates to values that end in zero or five. Recall bias was assessed by comparing three separate recall periods: extended, intermediate, and short. Nonresponse bias was assessed by conducting telephone interviews with subjects who had not responded to the initial requests. The dependent variable was the total number of days fished per month from April to June 1989. Analysis of variance and chi‐Square were used to explore the relationships among the three sources of bias. A significant Recall × Nonresponse × Digit Preference interaction was found. Results suggest that bias traditionally attributable to recall and nonresponse may be a function of digit preference. Ways of controlling for the effects of the three sources of bias on recreation participation estimates are discussed.

Keywords:

• digit preference
• recall bias
• nonresponse bias
• participation rates