Meaning/interpretation of interaction (term in regression)

Question

I want to study how some certain parameters affect sensors of different manufacturers. To study any possible interactions, e.g., in lm in R, it is possible to study these interaction effects via

lm(Response ~ (rH + T)^2, data = df)

with the Response to the predictors relative humidity and temperature. I'm in the weird situation now, that relative humidity rH and temperature T have a strong influence on both sensors while the interaction term has also a strong influence on the one sensor and on the other not at all.

How do I interpret this? Even in general, how to treat interaction terms? When, in my case, humidity and temperature have already an influence, is there any (physical) meaning for their interaction?

Answer 1

You didn't tell us anything about the nature of the sensors ... you might get better answers if you do so! But, you say

When, in my case, humidity and temperature have already an influence, is there any (physical) meaning for their interaction?

Certainly there can be! Interaction between humidity and temperature measures the degree to which the effect of humidity depends on temperature, or, equivalently, the degree to which the effect of temperature varies with humidity. One example could be rusting, the speed of which would be influenced by both temperature and humidity, and I would guess the effect of humidity on rusting could well be dependent on temperature ...