Weighting

[Alkmim]

Analysis for universal research or surveys without weights, Jamoni is perfect. For surveys whit pos-stratification no. I hope for a solution soon.

[chris.cambron]

I would like to add another vote for this. I am considering switching my doc student stats classes to jamovi but without standard complex sample options like weights and stratification I am unable to do that. Thanks so much.

[jonathon]

yeah, it's tricky. we use R for analyses under the hood ... but there's no way in R to "throw a switch", and weight things accordingly.

the R packages which consume weights, tend to be quite narrow and tailored to a specific problem ... and they often don't play nicely with the other packages we use to do other things. so it's non-trivial. having said that, we have been thinking more seriously about this lately.

jonathon

[shabde]

I can totally see the problem with making weighting available to inferential statistic functions in JAMOVI. Indeed, you have problems finding packages that support this in e.g. mixed models or factor analysis. But what about confining weighting to functions provided in Exploration,t-Tests,ANOVA and Regression menus of JAMOVI?
Exploration should be covered by srvyr package fully, and for t-Tests, ANOVA and Regression, packages with weighting support do exist.

[jonathon]

yeah, stay tuned, we should release weights in the next month or so.

jonathon

[bij_nz]

yeah, stay tuned, we should release weights in the next month or so.

jonathon

Excellent news!
A lot of the analyses I do use decimal weights to correct for non-random or stratified sample selection (e.g. over-sampling minority populations) or implementing post-stratification weights.
Brendan

[Ana_GRL]

Thank you very much for implementing stratification weights.
I have seen that it can be applied with t-Tests, ANOVA, crosstabulations and contingency tables...
Is it planned to be implemented in other analyses such as factor analysis? In sociological, psicological surveys, weighting is necessary
Again THANK YOU VERY MUCH

[FinnK]

Hi there,
As I understand from the post in Mar 11, weights should be released. Is that into the latest version or is it a module?
Is there a way to see how to apply the weights to the data prior to running the analyses?
Thanks

[jonathon]

in the 2.4 series. you'll find a button for weights under the 'data' tab.

jonathon