Adds three-level meta-analysis by specifying study_ids argument in RoBMA.
However, note that this is (1) an experimental feature and (2) the computational expense of fitting selection models with clustering is extreme. As of now, it is almost impossible to have more than 2-3 estimates clustered within a single study).
Further computational improvements are coming in the future and this is not the final version of the multivariate functionality. However, merging this "sub-ideal" functionality now allows me to incorporate additional functionalities without creating additional conflicts.
Adds three-level meta-analysis by specifying
study_idsargument inRoBMA.However, note that this is (1) an experimental feature and (2) the computational expense of fitting selection models with clustering is extreme. As of now, it is almost impossible to have more than 2-3 estimates clustered within a single study).
Further computational improvements are coming in the future and this is not the final version of the multivariate functionality. However, merging this "sub-ideal" functionality now allows me to incorporate additional functionalities without creating additional conflicts.