multi-level analysis

Discuss the jamovi platform, possible improvements, etc.
Post Reply
Hark
Posts: 9
Joined: Mon Jun 12, 2023 5:45 am

multi-level analysis

Post by Hark »

Research often encounters situations where data needs to be escalated to the team and analyzed using team-level data. To totalize, most studies report three data: rwg, ICC(1), and ICC(2). That is, to add up the totals, three conditions have to be met, rwg>0.7, ICC(1)>0.05, ICC(2)>0.5.

Does jamovi enable multi-level analysis? And are these metrics exportable ?
User avatar
mcfanda@gmail.com
Posts: 448
Joined: Thu Mar 23, 2017 9:24 pm

Re: multi-level analysis

Post by mcfanda@gmail.com »

Hi,
Multi level analysis can be done with GAMLj3 mixed model. There you get also the ICC.
Hark
Posts: 9
Joined: Mon Jun 12, 2023 5:45 am

Re: multi-level analysis

Post by Hark »

mcfanda@gmail.com wrote: Wed Nov 29, 2023 10:13 am Hi,
Multi level analysis can be done with GAMLj3 mixed model. There you get also the ICC.
I hope you can add the stats for RWG and ICC2, thank you very much.
User avatar
mcfanda@gmail.com
Posts: 448
Joined: Thu Mar 23, 2017 9:24 pm

Re: multi-level analysis

Post by mcfanda@gmail.com »

As regards the ICC2, it is simply given by the spearman-brown formula applied to ICC, so V_i /(V_i+V_e/k), where V_i is the variance of the intercepts, V_e is the residual variance, and k is the number of raters or cases within clusters. R_wg is not a model-based indices (it requires inventing a distribution that assumes a different model), thus it will not be implemented because it does not pertain the estimated mixed model.
Post Reply