Hello,

I have calculated the user-written command "inteff" after a logit regression to check my interaction term consisting of two continuous variables with values between 0 and 1. The usage of the "inteff" command is given, so I have to use it. Basically I understand the outputs from the "inteff" command, only in the case of the following figure I don't understand the statement. It irritates me, that the axis "Interaction Effect (percentage points)" has negative values in connection with a predominantly increasing "predicted probability". Can someone kindly help me how to interpret this figure?

I understand the illustration as follows: The interaction effect is negative. A negative interaction effect means that the closer the both continuous variables get to their maximum value of 1, the more negative the interaction effect becomes. For me, however, it only makes logical sense that a negative interaction effect is accompanied by a decreasing predicted probability.

Totally, I don't get it.

Array

Many thanks for your help.