Multiple informants

[PsychResearch]

---

Members don't see this ad.

.

 

[Ollie123]

What is the IOR for the informants, assuming they are on the same metric? How "different" are the findings across them? This alone shouldn't doom an otherwise strong paper. Is there one informant who may be more "trustworthy" than the others? You could just discuss the differences, etc.

Technically, I think you are misusing covariates unless this was a group they were randomized to (though pretty much everyone does it whether its appropriate or not), so I also think you are justified in dropping it and analyzing them separately and just discussing accordingly.

 

[PsychResearch]

What is the IOR for the informants, assuming they are on the same metric? How "different" are the findings across them? This alone shouldn't doom an otherwise strong paper. Is there one informant who may be more "trustworthy" than the others? You could just discuss the differences, etc.

Technically, I think you are misusing covariates unless this was a group they were randomized to (though pretty much everyone does it whether its appropriate or not), so I also think you are justified in dropping it and analyzing them separately and just discussing accordingly.

I decided to summarize the findings and discuss the discrepancies within the context of using multiple informants. The idea is to turn this into a strength rather than making it a weakness.

I can't drop the covariates as it makes theoretical sense to have them in. Just as a example, suppose the focus was anxiety and y outcome and the covariate is depression. Since depression and y have been previously linked to each other, one has to ensure that whatever findings they get is actually due to anxiety rather than depression. Without getting insto specifics, the rationale for using covariates in my paper is similar.

Thanks for your reply. It helps to have a second opinion...or triple etc.