Measures of Nonresponse Bias Associated with the 2008 Administration of the University of California Undergraduate Experience Survey by Steve Chatman. SERU Project Technical Report (2009)

PDF document (351 kB)

Abstract: The current study examining the results of the 2008 University of California Undergraduate Experience Survey (UCUES) is a replication of the nonresponse bias analysis of the 2006 SERU/UCUES survey. Both traditional and response resistance techniques were applied to the academic factor score values of the 2008 SERU/UCUES survey so as to determine the practical significance, based on effect size statistics, of factors that might indicate nonresponse bias. While there were participation rate differences, especially by campus and grade point average, these and other student characteristics were either not associated with group differences or were found to have differences that did not exceed a small effect size threshold. Similarly, the response resistance analyses found examples of trending in mean responses from early to late responders at one campus, but the trends were not replicated across campuses. In addition, an examination of cumulative mean values suggests that a response rate as low as 20% would have adequately represented campus statistics. Taken together, these results suggest that differential weighting of survey responses to mirror population characteristics is not necessary and the observed response rates were adequate.