Recall Salience: Concept, Use, and Estimation

Abstract

A 1998 review of the 1994, 1996, and 1997 Canadian Travel Surveys (CTS) provided evidence that a large decline in estimated travel was related to a change in respondents' efficiency in recalling trips, in other words, to a change in trip recall salience (TRS) between the years. Because the CTS data are collected on all trips that respondents take in a month, one can examine the order in which different categories of trips are reported. Research reviewed in this article shows how the statistical significance of TRS and the estimation of a TRS scale can occur. Scale estimation is critical to work cited as making estimates of the consequence of changes in survey methodology. This research pursues the systematic estimation of TRS scales using regression. Topics covered include avoiding bias, estimation of a TRS scale using regression, and estimating bias in the CTS using a TRS scale. Because numerous surveys collect data on occurrences recalled for a given period of time, it follows that some analyses where salience is relevant will be based on data sets large enough that the ideas presented and the methodology developed will be of benefit. For other studies a caution is discussed about the impact of salience, even if a scale cannot be estimated and recall bias evaluated.