Highway Welcome Center Surveys: Problems with Nonresponse Bias

Abstract

The ability of survey research to accurately represent the characteristics of highway welcome center visitors has become an important issue. This article reports on a study that attempted to evaluate the nonresponse bias associated with Texas Welcome Center visitors. Two data sets are compared: (1) a study that sampled visitors without follow-up on nonrespondents (response rate was 3%); and (2) a study that sampled visitors with follow-up (response rate was 78%). Results indicated that welcome center visitors who responded to sampling without follow-up were associated with a lower proportion residing in the adjacent state and a higher proportion residing in the nonadjacent state compared to respondents of the sample with follow-up. Besides the above consistent bias, different nonresponse bias effects were found across two travel seasons and across two welcome center locations on age, trip length, expenditure pattern, trip planning horizon, and influence of welcome center information.