Weighting merged datasets

[Zuhura Anne]

Hello,

I have merged two data sets. One has weight x and the other weight y.

Can I just use one weight or the other or is there a way I can use both weights?

Thanks

[Carlo Lazzaro]

Zuhura:
it depends on what those weights mean (something that we cannot know).

[Phil Bromiley]

To expand on Carlo's answer, we can't begin to give you advice on this matter without a lot of information about the context, objectives, models, and kind of analysis. The reason for the weights also is critical.

I suspect you're merging these observations so you have one set of variables that you think should have weight a and are going to merge this with variables that you think should have weight b. It is not clear this makes sense. Normally, we weight observations rather than variables. For example, we might have aggregate data by state and weight by state population. To "weight" some variables and not others in a given observation sounds a lot like just rescaling.