Hi, I have a question about matched-pair analysis. Can anybody give me a clue how to solve the following or similar problems? Assume we have found 1000 pairs of such kind of twins, one of the twins is very shy and another is very open. We would like to find out which factors or variables (among the 100 variables listed below) are significantly responsible for this character difference between the twins. variable 1: Education ( 1 means have college education and 0 means not). variable 2: Health ( 1 means good and 0 means bad). variable 3: Grow up with parents or not (1 means with and 0 means not) variable 4. ... variable 5. ... . . . variable 99. ... variable 100. ... In a more complicated case, all these variables may take not only dichotomous value, e.g. for education we may value college as 3, high school as 2, elmentary school as 1 and 0 means without any education. Does anybody know some good tools to solve this problem? Thanks a lot in advance! Xianghong Zhou Email: [log in to unmask] Tel: 0041-1-6326353 Fax: 0041-1-6321172 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%