I am running an analysis looking at individual voter turnout where it is a 1 for voting and 0 for not voting. I have some information on individuals including age, gender, and several years of past voting history. I can also geocode the information and add all sort of other grouped or clustered data including school district level, county level, township level, city/village level, and block group and census block level data. As I have several prior years of voter history data, I would like to add year specific data such as polling data, unemployment rates, personal income, etc.
My question is two fold, is GLMM the best solution to this problem? If so, I am having trouble determining which level data (geographic level or year/time level) should be on the Data Structure subject line, and which data (if any) should be in the repeated measures line. I believe (and correct me if I am wrong) that the geographic level data should be in the subject section and the year/time should be in the repeated measures section. Any Thoughts?