Closed floswald closed 4 years ago
in 4.4, interactions, it's confusing that the partila effect of avginc is negative. there should be another model
lm_inter = lm(formula = testscr ~ str + avginc + str*avginc, data = Caschool) u = update(lm_inter, .~.-avginc)
which removes the avginc effect. how to intuitively explain this?
in 4.4, interactions, it's confusing that the partila effect of avginc is negative. there should be another model
which removes the avginc effect. how to intuitively explain this?