Choosing a Household Survey Method: Results of Dallas—Fort Worth Pretest

Abstract

Presented are the findings of a pretest of a travel behavior survey conducted in the Dallas—Fort Worth region. The pretest directly compared five methodological options: ( a) a 24-h versus 48-h diary period; ( b) a shorter versus longer series of questions about each activity recorded; ( c) three types of incentives to encourage response; ( d) a booklet versus log format for the diaries; and ( e) telephone versus mailback retrieval of information. A diary recording period of 24 h was 1.6 times more likely to result in a completed household survey response than a 48-h recording period ( p-value <0.01). Telephone retrieval was twice as likely to result in a completed household than through mailbacks ( p-value <0.01). There was a significant difference ( p <.01) in the mean number of trips per day per person for those retrieved by telephone and mailback. Households with a 24-h recording period responded with significantly more trips per day per person (4.7) than households with a 48-h period (3.9) ( p <.01). Considering separately the trip rate on each of the 2 d that made up a 48-h reporting interval, the mean number of trips per day per person showed a statistically significant trend ( p <.01) toward fewer trips on the second day (mean difference of 0.6 fewer trips compared with the first day).