| Session Title: What Am I Supposed to Do With Three-Way Crosstabs? An Introduction to Log Linear Models |
| Demonstration Session 866 to be held in Lone Star E on Saturday, Nov 13, 2:50 PM to 4:20 PM |
| Sponsored by the Quantitative Methods: Theory and Design TIG |
| Presenter(s): |
| Eric Canen, University of Wyoming, ecanen@uwyo.edu |
| Nanette Nelson, University of Wyoming, nnelso13@uwyo.edu |
| Abstract: A common analysis method for cross-tabulated data is a Chi-Square test for independence. When considering only two factors, the choice of Chi-Square is straightforward; however, with three or more factors, the analysis of independence is more complicated. For instance, you can test whether all the factors are independent or whether one factor is independent of two other factors. Chi-Square methods can be used in these analyses however, such analysis is not easy completed using common statistical software. Log-linear models offer an easily understood and easily implemented alternative. This demonstration will provide step-by-step guidance on how to implement this analysis technique. The presenters will teach by example using data from an investigation of differences in smoking-related behaviors and attitudes pre- and post-implementation of smoke free ordinances. Participants will have time to ask questions and will receive a handout to guide them through their own analyses. |