Significance test of proportions from weighted data

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Significance test of proportions from weighted data

Jignesh Sutar
This post was updated on .
Apologies for cross posting (with dev central) but wonder if members of this list have thoughts regarding the below.


Is it possible in SPSS to carry out Z-test for proportions of independent samples on weighted data, taking into account effective samples size (not simply the the weighted sample size)
 
See attached picture for desired calculation formula where e1 and e2 are effective sample sizes.
 
I've read the algorithm documentation for CTABLES "Compare proportions" and it seems the weighted sample size is taken and no toggle options for inputting the effective sample size? Are there any other procedures or modules able to carry this out? Also are there any options of using pooled vs. unpooled variance?
 
 



Many thanks in advance,
Jignesh