Hello everyone
Does anyone know much about when it is appropriate to adjust for using a large number of regression models in one piece of work? I have around 80 models. They are all testing one overall hypothesis - association between socio-demographics and attitudes.
It seems to be that socio-demographics attitudinal research is by nature not a 'precise' affair if that is the right term, and so it doesn't make sense to make precise adjustments to p values, when the overall idea is to build a narrative based around a large number of models. But I want to be sure. Any justification or references would be very much appreciated.
Thanks for reading!
Note I am already using adj. r2 values as measures of association if that makes any difference.
|