sdm-help-list > Estimate interpretation
Showing 1-2 of 2 posts
Aug 7, 2018 08:08 AM | Arnaud Leroy
Estimate interpretation
Hi all,
I have a naive question about the interpretation of the meta-regression results.
I have significant regions for the contrast patients > controls (Z positive) and regions for the contrast controls > patients (Z negative).
How can I interprete it with my regressor ? Does it increase the effect of the contrast (if the score of my regressor increase the contrast increase) ? Is there another way to interprete it (I don't know how to interprete the negative estimate for deactivation: when the regressor increases, the deactivation increases ? or decreases ?)
In another way, how can I know the way of the interaction ?
Than you for your help
AL
I have a naive question about the interpretation of the meta-regression results.
I have significant regions for the contrast patients > controls (Z positive) and regions for the contrast controls > patients (Z negative).
How can I interprete it with my regressor ? Does it increase the effect of the contrast (if the score of my regressor increase the contrast increase) ? Is there another way to interprete it (I don't know how to interprete the negative estimate for deactivation: when the regressor increases, the deactivation increases ? or decreases ?)
In another way, how can I know the way of the interaction ?
Than you for your help
AL
Dec 5, 2018 08:12 AM | Joaquim Radua
RE: Estimate interpretation
Dear Arnaud,
The interpretation might be as follows:
- If a region shows patients > controls in the main analysis and the regressor is positive, we may conclude that the regressor increases the difference patients > controls.
- If a region shows patients > controls in the main analysis and the regressor is negative, we may conclude that the regressor decreases the difference patients > controls.
- If a region shows patients < controls in the main analysis and the regressor is negative, we may conclude that the regressor increases the difference patients < controls.
- If a region shows patients < controls in the main analysis and the regressor is positive, we may conclude that the regressor decreases the difference patients < controls.
I hope this helps.
Joaquim
The interpretation might be as follows:
- If a region shows patients > controls in the main analysis and the regressor is positive, we may conclude that the regressor increases the difference patients > controls.
- If a region shows patients > controls in the main analysis and the regressor is negative, we may conclude that the regressor decreases the difference patients > controls.
- If a region shows patients < controls in the main analysis and the regressor is negative, we may conclude that the regressor increases the difference patients < controls.
- If a region shows patients < controls in the main analysis and the regressor is positive, we may conclude that the regressor decreases the difference patients < controls.
I hope this helps.
Joaquim