Subsamples adjusted for age

[Niko Kohls]

Dear List,

I'm currently working on a problem with drawing comparable subsamples
from a survey study and seem to be stuck. I already browsed the
archives but haven`t exactly found what I was looking for. Any hints
would be really appreciated.
Problem description: I have data from a study with two different
groups (n1= 350; n2 = 300). Mean age in group 1 = 45 years and in
group 2 = 34 years. I want to control for age by applying a match-case
approach. This should not be too difficult at all if the subsamples,
which to be drawn, are small (e.g. n= 30). However I want the two
subsamples to be comparable (i.e. controlled for age), but I also want
to have the solution with the biggest sample size (e.g.). In other
words, is there a way of computing that solution in SPSS?
Thanks for your help,
Best Niko

[Maguin, Eugene]

Niko,

Unless I misunderstand what you are doing, your final sample size is
controlled by your requirement that there be a one-to-one match on age. To
see what that is going to be do a crosstabs of age by sample. Within each
row the smaller of the two numbers will be the number of matches for that
age.

I'll assume you are familiar with syntax as my procedure is going to be
descriptive.

Open your datafile and save group 1 to a new file and then save group 2 to a
different new file.

Open the group 1 file and sort by age. Next, you need to number the cases
with each age group. Like this.

Compute rec=1.
If (idvar eq lag(idvar)) rec=lag(rec)+1.

Repeat for the group 2 file.

Lastly, do a match files with the group 1 and group 2 files using idvar and
rec as the 'by' variables.

Gene Maguin