Data Quality in an Internet-Based College Satisfaction Survey

This paper reports on the results of an experiment that investigated the response rate and data quality effects of two alternative methods for making initial contact with potential respondents in a sample survey. The study was conducted in a university setting and selected students were initially contacted about their participation in the survey either thru an e-mail message or thru their own customised university portal. Survey findings revealed that the e-mail alternative resulted in a significantly higher overall response rate. However, the findings also revealed that there were significant differences in the responses of the two groups with respect to the evaluations given by students regarding their academic and social well-being at the university. Specifically, sample members in the portal group gave significantly more positive evaluations regarding their academic and social well-being than did members of the e-mail group. However, there were no significant differences between the responses of the two groups in terms of their evaluations of either the university‟s food court or the campus recreational centre. Additionally, despite efforts within both groups to achieve a high response rate and a representative sample, neither alternative was able to produce a sample that had a high response rate or a sample that was representative of the population in terms of either students ‟ gender or grade point average. Implications of these findings for conducting surveys, especially those of student populations, are also discussed