sdm-help-list > Correlative and group analyses
Showing 1-4 of 4 posts
Display:
Results per page:
Mar 17, 2015  04:03 AM | Nobody
Correlative and group analyses
Is it possible to conduct a VBM meta-analysis where some studies test group differences and other studies are correlational? Or is there something fundamentally wrong with such an analysis? Of the 11 studies I'm using 7 use cut off values to categorize participants then contrast the two groups, whereas 4 studies correlate grey matter with a continuous scale. Any help would be much appreciated.

Thanks!
Casey
Mar 22, 2015  03:03 PM | Joaquim Radua
RE: Correlative and group analyses
Dear Casey,

I think that it is possible, and indeed SDM has been already modified for combining one- and two-sample studies (see http://onlinelibrary.wiley.com/doi/10.1111/pcn.12185/full ). Of course two-sample studies and correlation studies are not the same, but this will probably be only one of the many methodological differences between studies (others being the scanning device used, the specific steps applied in the pre-processing, and etcetera).

Best wishes,

Joaquim
Oct 14, 2020  07:10 AM | Bing Chen
Question about correlative and group analyses
Originally posted by Nobody:
Is it possible to conduct a VBM meta-analysis where some studies test group differences and other studies are correlational? Or is there something fundamentally wrong with such an analysis? Of the 11 studies I'm using 7 use cut off values to categorize participants then contrast the two groups, whereas 4 studies correlate grey matter with a continuous scale. Any help would be much appreciated.

Thanks!
Casey


Dear Joaquim,
I'm a new user of SDM, I also have the same question. I would like to ask that n1, n2 must be entered in the two-sample t-test in the SDM table; but the correlation study has only one sample (n1?), should we directly put the two types of researches in one SDM table?

Thank you very much in advance!
With best wishes,
Bing
Oct 14, 2020  02:10 PM | Bing Chen
RE: Correlative and group analyses
Originally posted by Joaquim Radua:
Dear Casey,

I think that it is possible, and indeed SDM has been already modified for combining one- and two-sample studies (see http://onlinelibrary.wiley.com/doi/10.1111/pcn.12185/full ). Of course two-sample studies and correlation studies are not the same, but this will probably be only one of the many methodological differences between studies (others being the scanning device used, the specific steps applied in the pre-processing, and etcetera).

Best wishes,

Joaquim

Dear Joaquim,

I'm a new user of SDM, I also have the same question. I would like to ask that n1, n2 must be entered in the two-sample t-test in the SDM table; but the correlation study has only one sample (n1?), should we directly put the two types of researches in one SDM table?

Thank you very much in advance!

With best wishes,
Bing