A study was performed to assess a potential relationship bet

A study was performed to assess a potential relationship between BMI (continuous) and two predictors: watching TV categorized as: 0 = not at all, 1 = rarely, 2 = 2 often) and gender (female, male). The estimated means across margins turned out to be all the same: Could there be an interaction between gender and watching TV? Briefly explain why or why not.

Solution

There could be an interaction between gender and watching TV because even if the averages are equal, its not responsible to tell about the interaction between the two variables. We need to have data (frequency form) for this problem to an association between the two variables or we can have ANOVA to see if the interaction between gender and watching TV is significant.