If You Provide It, Will They Read It ? Response Time Effects in a Choice Experiment

Substantial effort is expended in the design of surveys, including the amount and type of information they contain. However, we often do not know how involved respondents are in reading and processing the informational content of a survey and making choices, and whether different levels of involvement result in systematic differences in estimated models. To address this issue, we recorded response times for each respondent of an internet-based choice experiment for stream restoration. Response times per survey section and for the entire survey were used as proxies for the amount of involvement in reading information provided or answering choice questions. Response times per survey section fell rapidly, possibly signaling learning, use of heuristics, or attempts to quickly dispel with the survey. Response times were found to be independent of demographics and attitudes. Log-likelihood ratio tests failed to reject the null hypotheses of equal coefficients and scale parameters across response time-partitioned data. However, there exists an association between response times and the increasing learning curve or difficult choice trade-offs, suggesting a heuristic response. Additional research on response time effects and survey design is needed, especially with the rise in electronic surveying media.