I'm working with survey data dealing with reproductive health among women with HIV. I would like to explore possible factors which may contribute to pregnancy amoung this population. I have survey data from HIV positive women who are pregnant and those not pregnant. I have 487 surveys with over 150 variables. I've been told I should do cluster analysis and then multi-variate regression. I've been working on the cluster analysis, but it seems this method is mostly for clustering cases, not variables. Any ideas on how best to go about exploring potential factors which may contribute to pregnancy among this data set? Maybe cluster analysis is not the way to go?? Many thanks to anyone who can offer suggestions!