UNIANOVA with Post-hoc Tukey using weighted variable for mean testing

#1
Hi everyone,
I am trying to analyze in SPSS if there a significant difference in the proportion of a gene mutation among different age groups (20-29, 30-39, etc).

I codified the age groups from the continuous var (age) into a nominal variable with values 1,2,3,etc for each age group and the gene mutation with 0 for negative and 1 for positive. In total, I have 6990 rows (cases) with valid entries for both age group and gene mutation to be analyzed.
I was thinking that a UNIANOVA procedure could be useful to test whether the gene is more mutated in specific age groups compared to others. This is my input obtained using the gene mutation as dependent var and the age groups as fixed factors:

UNIANOVA ASXL1 BY agegroup_no0_8
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/POSTHOC=agegroup_no0_8(TUKEY)
/CRITERIA=ALPHA(0.05)
/DESIGN=agegroup_no0_8.

and here´s the output: (see attachment)

Do you think it is conducted appropriately in this way? If it´s good, I would need the post hoc to keep into account the weighted age groups to calculate the mean differences and not the 1,2,3 values. i tried using the WLS weight field but I am not sure I am using that correctly. Do you recommend any workaround?
Should I instead use another entirely different procedure?

Thanks in advance :))
Luca
 

Attachments