quick stats question (Kuder Richardson 21/KR-21)

[bmedclinic]

---

Members don't see this ad.

Needing to run a KR-21 for one part of my dissertation.
I know SPSS can do a Chronbach's alpha, but I've never seen an option for the KR-21. Is there a way to do it in SPSS, or should I bet looking towards a different application?

Thanks!

Bmedclinic

 

[irish80122]

I am not aware of a way to do it in SPSS, but then again I am a SAS guy, so take that with a grain of salt. You likely have found this, but here is a brief explanation of how to do it semi-manually.

https://www-304.ibm.com/support/docview.wss?uid=swg21476088

 

[bmedclinic]

I am not aware of a way to do it in SPSS, but then again I am a SAS guy, so take that with a grain of salt. You likely have found this, but here is a brief explanation of how to do it semi-manually.

https://www-304.ibm.com/support/docview.wss?uid=swg21476088

I actually hadnt found that. All I found was some way to do it in excel, which i had considered a last resort. Thanks.

And btw, I was a SAS guy. Then I left my masters program and came to my doctorate where everyone used SPSS and SAS was more or less rejected. I prefer the command line.🙂

 

[aequitasveritas]

I think you can find some syntax commands for KR20 but not KR21

 

[Ollie123]

Unless I'm mistaken KR-20 should actually be identical to Cronbach's alpha (it was a case specific to dichotomous variables before Cronbach developed the more general one). KR-21 is a different story.

I just glanced through and didn't see anything. Good news is that the math for it seems pretty straightforward so it should be easy to replicate with syntax/excel.

 

[thewesternsky]

Yep, KR-20 is identical to Cronbach's alpha and SPSS can do it using the Reliability procedure. (Just use the value it gives you for Alpha; they're the same). Not sure about KR-21.