kruskal wallis test

[lmyers2]

Hello,

I've been struggling with data that is overlapping (many defendants correlating to one claim). I'm trying to get a pvalue using the Kruskal Wallis test comparing medians of 3 continuous variables (paymentsum1, paymentsum2, paymentsum3). Below is the structure of the data. I can't combine paymentsum1-3 into long format because the 1, 2, 3 correlate to different groups and that grouping cannot be lost. I'm hoping this is obvious!

case_id  paymentsum1 paymentsum2 paymentsum3

1            $500                $1000             $20

2            $100                .                      .

3            $200                $500   

4           $1000              

Thanks

Laura

[PGStats · Accepted solution]

Why would this data structure be inappropriate?

claim_id group payment
1 1 $500
2 1 $100
3 1 $200
4 1 $1000
1 2 $1000
3 2 $500
1 3 $20

no information is lost.

[VDD]

use proc transpose to convert wide data to long data.

https://stats.idre.ucla.edu/sas/modules/how-to-reshape-data-long-to-wide-using-proc-transpose/