+ Reply to Thread
Results 1 to 3 of 3

Thread: Interaction effect: Do the single variables both need to be significant?

  1. #1
    Points: 277, Level: 5
    Level completed: 54%, Points required for next Level: 23

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Interaction effect: Do the single variables both need to be significant?




    Dear all,

    If you have a regression model Y=b0+b1*A+b2*B+b3*A*B, and the dependent variable A and the interaction effect A*B have significant results, is this okay? Or do both A and B need to be significant in order to be able to interpretate the significant effect of A*B?

    Thanks...

  2. #2
    Probably A Mammal
    Points: 31,087, Level: 100
    Level completed: 0%, Points required for next Level: 0
    bryangoodrich's Avatar
    Location
    Sacramento, California, United States
    Posts
    2,564
    Thanks
    398
    Thanked 618 Times in 551 Posts

    Re: Interaction effect: Do the single variables both need to be significant?

    First, the A and B are independent variables. Y is the dependent variable as indicated in the model: Y depends on the explanatory variables A and B, which ideally are independent of each other but correlated with Y.

    Second, what is the correlation between A and B? (Multicollinearity may be present).

    In any case, you shouldn't worry too much. You may try centering your independent variables (i.e., use "X-mean(X)" for X). This explains why. It also reduces multicollinearity. You want to look at main effects, which require you to look at the conditional effects (the A or B) in the presence of their interaction (A:B). You cannot consider them alone when there is interaction in your model.

  3. #3
    Devorador de queso
    Points: 95,866, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,936
    Thanks
    307
    Thanked 2,630 Times in 2,246 Posts

    Re: Interaction effect: Do the single variables both need to be significant?


    You definitely don't need the main effects to be significant to interpret the interaction. Its not too difficult to imagine a situation where both main effects are 0 but the interaction is highly significant.
    I don't have emotions and sometimes that makes me very sad.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats