Weights

Discuss statistics related things
Post Reply
tahidus
Posts: 11
Joined: Wed Aug 31, 2022 5:44 am

Weights

Post by tahidus »

jamovi 2.4.0 has support for weights. Please, any ideas on how to use the weights feature? The weight icon is under the data tab
Thanks to the jamovi hard working team for adding this feature
Alex
Posts: 2
Joined: Tue Aug 01, 2023 7:03 am

Re: Weights

Post by Alex »

tahidus wrote: Mon Jul 10, 2023 11:22 pm jamovi 2.4.0 has support for weights. Please, any ideas on how to use the weights feature? The weight icon is under the data tab
Thanks to the jamovi hard working team for adding this feature
+1 !
User avatar
Dietmar
Posts: 2
Joined: Sun May 22, 2022 11:31 am
Location: Austria
Contact:

Re: Weights

Post by Dietmar »

I really am very happy about the weights function - thank you very much for implementing it! :grinning:
We use it, if the sample structure slightly differs from the structure of the population at some predefined variables (e.g. age, sex, education, place of residence). This is intentional with disproportionate samples. However, the deviations can also occur unintentionally due to various biases in the sampling process.
And within academic lectures on market research / data analysis it is very useful to show the effects of structural nonconformity.
Shabd
Posts: 3
Joined: Thu Dec 13, 2018 9:17 pm

Re: Weights

Post by Shabd »

Its absolutely great to see that JAMOVI supports case weights now. However, to see that only integer weights are possible is really a drawback, because this means that only very rough reweighting is possible. In survey statistics, design or redressment weights are usually floating point values (since you try to keep the reweighting factor as close to 1 as possible) and I look forward to see this working in JAMOVI too. Is there any planning or perspective to add this functionality?
User avatar
reason180
Posts: 278
Joined: Mon Jul 24, 2017 4:56 pm

Re: Weights

Post by reason180 »

If you use large integers (e.g., weights: 666666666, 333333333) won't that suffice?
Post Reply