0

This question might be bit naive question. But is it possible to identify which variable are suppressor and which are confounder before running regression just by reviewing the correlation values among independent variables and dependent variable.

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Beta
  • 5,784
  • 9
  • 33
  • 44
  • 1
    Among answers about suppression, here is my http://stats.stackexchange.com/a/73876/3277. There is both theoretical and technical definitions of a suppressor (in linear regression context). Hope it will help. It says you need to do a regression, a concrete model, to be able to detect a supressor, if any. However, often a suppressor is supposed beforehand by prior knowledge/theory and entered on purpose - and often it justifies itself so. – ttnphns Mar 29 '16 at 18:10
  • @ttnphns: Thanks a lot for the suggestion. I'll let read this and modify my question if required. – Beta Mar 29 '16 at 18:14

0 Answers0