| Roundtable: Addressing Complex Sampling Issues in Evaluation Analyses |
| Roundtable Presentation 688 to be held in the Sandstone Boardroom on Friday, Nov 7, 4:30 PM to 6:00 PM |
| Sponsored by the Quantitative Methods: Theory and Design TIG |
| Presenter(s): |
| Cindy Kronauge, Weld County Department of Public Health and Environment, ckronauge@co.weld.co.us |
| Susan Hutchinson, University of Northern Colorado, susan.hutchinson@unco.edu |
| Abstract: Evaluators are often asked to analyze probability surveys. It is important that survey data be analyzed in the best possible way. Complex sampling designs are often more economical and more accurate than simple random samples. However, many evaluators do not analyze probability survey data according to its underlying sampling design. The purpose of the proposed session is to demonstrate through use of several real datasets, step-by-step procedures for calculating sample weights, incorporating weights into several different popular statistical software packages (e.g., SPSS, LISREL), and conducting both descriptive and inferential statistical procedures. Examples will be shown to illustrate results that might occur when an evaluator conducts various analyses with and without the appropriate weights. Participants will gain a better understanding of the importance of incorporating a survey sample design into their analyses and will receive handouts to consult when they encounter a real life dataset they need to analyze. |