| Roundtable: A Model of Total Survey Error: Examining the Inferential Value in Survey and Questionnaire Data and its Implications on Evaluation Findings |
| Roundtable Presentation 324 to be held in Exec. Board Room on Thursday, Nov 3, 11:40 AM to 12:25 PM |
| Sponsored by the Quantitative Methods: Theory and Design TIG |
| Presenter(s): |
| Michelle Bakerson, Indiana University South Bend, mmbakerson@yahoo.com |
| Abstract: Charged with making judgments about the quality, merit or worth of a program many evaluators use survey or questionnaires to gather data, which is one of the most important and uniquely informative data collection tools available. The significance placed on this data collection tool is examined as limitations and biases inherently exist. Virtually all surveys will contain some form of error which will harm the inferential value of the data. A model of total survey error is examined including sampling error and non-sampling error with a major focus on non-response error. Practical attention to survey methodology in particular techniques for reducing and correcting for non-response will facilitate evaluators' knowledge and ability to make informed decisions regarding their data and in turn their ability to inform stakeholders. |