Sample Representativeness in Surveys With Lower Response Rates

• In Event: Conducting and Validating Surveys

Abstract

Much of the literature has been focused on how to increase response rates; however, it may be more important to know whether lower response rates are leading to data that are not representative of the options of a survey population. Our research focused on comparing key student characteristics and the college attended from the admitted non-enrolling student surveys and through National Student Clearinghouse at the population level. Significant differences were found between the survey sample and population for gender in 2008 and 2011, and for all three SAT test scores in 2008 and mixed findings in 2011. The 2011 survey respondents were more representative of the population, which led to consistent findings from both data sources on the college attended.

Authors

• Hui-Ling Chen, Saint Anselm College

• Jere Turner, Manchester Community College