Hello...I have a problem which one side of my brain should be simple - but that side of my brain can't seem to make by conscious brain understand.

I have longitudinal data where the respondents are age 14 to 18 in spell 1, age 15 to 19 in spell 2,..... and 28 to 32 in spell 11. (11 spells). My dependent variable is if the person ever killed anyone (dummy variable coded 0 if no and 1 if yes murder). What I want to be able to say is what is if anyone ever did murder for age 18, age 19, age 20...and so forth.

I can do this if I calculate each value for every separate spell in the data (I have variables age1, age2, age3, age4....age11 constructed).

the murder variable is a dummy variable which is also coded at each spell if they ever did murder (murder1, murder2...murder11)

So I could do this:

temp. select if age1=18. desc murder1. exec.

temp. select if age2=18. desc murder2. exec.

etc.....

Then I could sum the mean values and divide by the number of values I have to get the average percent who did murder at a given age.

This seems like a very cumbersome way of getting my results. I'd much rather be able to have SPSS figure it out with less comands.

Okay - what am I missing?

If this is not explained well please let me know.

Thanks

Joshua Masnick Kim Visiting Assistant Professor West Virginia University - Department of Sociology and Anthropology Morgantown WV 26506-6326 PO Box 6326 phone (304) 293-5801 ext. 1633 fax (304) 293-5994