• Hi Robert! It'll be too long for a reply so if its not too much to ask, see here: rpubs.com/jmgirard/sr5 , Questions 5H1, 5H2 . At the end there's an explanation : but these effects get cancelled out in the bivariate regressions because territory area and group size are positively related. Which I can't seem to grasp.. – Yarden Gur Aug 8 '20 at 9:04