jamovi

Is there a possibility to include the options for weightings?

like a 'weight cases' ala spss?

i've been meaning to do it, but haven't got to it yet.

jonathon

Yes, that would be great. Thanks Jonathon.

Hello,

I would like to reiterate the desire. When I discovered jamovi I was flashed, I saw beauty, structure and pure joy, the angels singing from heaven and God writing "jamovi" in the clouds. Unfortunately, I can't use it because without a weighting function for sociology, it's a one-way street.

In contrast to the natural sciences and psychology, we rarely conduct experiments in sociology. In experimental design, the sample can be constructed so that it corresponds to reality. In surveys, however, many respondents drop out for various reasons, so that the sample is no longer correct and weighting is necessary.

I understand that jamovi wants to be puritanical, I love that. Weighting seems like a different category of "puritanical" to me, though. I think when the ordinary sociologist is looking for a puritan tool, he says, hey, I don't need all that high level stuff. I just need simple analytics. But without weighting, he gets a tool that can't do even simple analysis.

Greetings

Small note : Data sets from population surveys conducted by large survey institutes, e.g. European Working Conditions Surveys, provide their data sets with weighting factors. Weighting is not an insignificant matter. https://www.eurofound.europa.eu/surveys/european-working-conditions-surveys/fifth-european-working-conditions-survey-2010/ewcs-2010-methodology/ewcs-2010-weighting

I would like to make one comment. A weighting like in bluesky, which adds or takes away cases from the data set, would indeed be only a small gain. This also works with database tools before import. The problem is that weights are usually decimal numbers though. But I can't prepare my dataset so that all women are both 1.33 times and still nominal scale.

I dug through all the free software and found only in PSPP a true weighting function that I can use to evaluate the major national and international surveys.

The one question is yes, should Jamovi be able to evaluate all kinds of analyses. Surely not. Another question is, should Jamovi be able to evaluate all types of data. Because in principle the missing function means: Guys, Jamovi should not be able to analyze survey data.

wait ... i've only been thinking of adding 'counts' as weights ... you're saying there's a market for weights with decimal places?

jonathon

I think there is a market for this. Not all the time, but in some kinds of work that I do (experiments), I might have effects that depend on demographic factors. There's increasingly a use for being able to deferentially weight various demographic variables to correspond to the precise demographic makeup of different locations / settings / etc.

I agree with you guys. If I am to use and recommend Jamovi as a viable alternative to SPSS, weights need to be implemented. Most sociological surveys, at least in Spain, use this type of weighting scheme and Jamovi needs to have a tool to deal with this issue.

Maybe we can help, Jonathon. I have a few ideas. Could we get in touch in private?

sure. i've sent you a slack invite.

cheers