FloSchuberth/cSEM

calculation of effects in calculateEffects

Closed this issue · 1 comments

Currnelty, we apply a rounding in the calculateEffects function which can have negative side effetcs, namely that effects are not calculated if they are very close to zero. A solution would be to either find a more accurate way of calculating the inverse of a matrix or to develop indicator matrices that show whether a certain variable has a direct, indirect and/or total effect on another variable.

This is issue is partly solved for recursive models. For non-recursive models it might still be a problem