Does interaction mean covariance, not correlation?

35 Views Asked by At

For example let's say that we want to predict math scores by income and history score (non standardized, raw score), and their interaction

y = b1(history) + b2(income) + b3(history*income)

the meaning of b3 is significant: income and history scores moderate each other on math score

In that case, does that mean that history and income covary?

(not correlate. history and income can have no -correlation and simultaneously have high covariance. income and history are IV, and let's assume that history and income show no correlation here. )