Hello, my question is triggered by an actual model I am running, but I will pose it as a very general question with a hypothetical example.
Take the following regression model: I have a binomial dependent variable "Happiness", whose two values are 0 (=unhappy) and 1 (=happy). My two independent continuous variables are "Income" and "Children". Imagine that "Income" has no significant effect, and that "Children" has a significant positive effect (more children give more happiness). Now I am interested in the interaction between "Income" and "Children", i.e. 'Income : Children'. Say that the model finds a non-significant negative coefficient. How do I interpret that? If I understand it correctly, the model is asking "When the numerical values in "Income" and "Children" are both increasing, does it significantly affect the dependent variable?". But what if the interaction between them is inverse - such that a decreasing value in "Income" paired with an increasing value in "Children" will significantly affect the dependent variable "Happiness". Will the model not be able to capture that connection without doing some additional tweaking? Thank you Sverre ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
