+ Reply to Thread
Results 1 to 6 of 6

Thread: Interactions in regression analysis

  1. #1
    Points: 3,811, Level: 39
    Level completed: 8%, Points required for next Level: 139

    Posts
    8
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Interactions in regression analysis




    Is their anyone who knows how to identify interactions between independent variables in a regression analysis? How does one use the output diagnostics to determine that there is an interaction.

    Thanks

  2. #2
    Admin
    Points: 20,637, Level: 90
    Level completed: 58%, Points required for next Level: 213
    quark's Avatar
    Location
    Canada
    Posts
    479
    Thanks
    25
    Thanked 196 Times in 73 Posts
    You can look at pairwise correlations between the independent variables, if r values are greater than 0.80 (or smaller than -0.80) the variables are strongly correlated.

  3. #3
    Points: 3,811, Level: 39
    Level completed: 8%, Points required for next Level: 139

    Posts
    8
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hi Quark

    Thank you for your response. From what you say, I infer that if two variables correlate with one another then they will interact? I have not come across this in my statistical studies. If two variables are highly correlated I am inclined to consider removing one from the analysis rather than try to model an interaction between them.

    But I know that interactions are possible between variables that are not correlated; how would I detect interaction effects between such variables?

  4. #4
    Admin
    Points: 20,637, Level: 90
    Level completed: 58%, Points required for next Level: 213
    quark's Avatar
    Location
    Canada
    Posts
    479
    Thanks
    25
    Thanked 196 Times in 73 Posts
    Quote Originally Posted by firefly View Post
    But I know that interactions are possible between variables that are not correlated; how would I detect interaction effects between such variables?
    There's a VIF(Variance Inflation Factor), which measures the impact of collinearity among the X's in a regression model. Most statistics software will show it in the multiple regression output. You can google it to learn more.

  5. #5
    Points: 3,811, Level: 39
    Level completed: 8%, Points required for next Level: 139

    Posts
    8
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hi Quark

    Two points here.

    1) Does correlation imply interactions? your responses so far seem to imply so, but I have never heard of this before.

    2) Yes, VIF has a lot to do with collinearity, but how is it related to interaction effects, especially when the variables are not correlated as mentioned in your quote from my previous mail

  6. #6
    Admin
    Points: 20,637, Level: 90
    Level completed: 58%, Points required for next Level: 213
    quark's Avatar
    Location
    Canada
    Posts
    479
    Thanks
    25
    Thanked 196 Times in 73 Posts

    1) In multiple regression context, if two IVs are significantly correlated, then chances are there are significant interactions. There's no need for a broad statement like "correlation imply interactions".

    2) Multicollinearity is interaction between IVs. Typically a VIF value greater than 10 MAY suggest multicollinearity.

+ Reply to Thread

           




Similar Threads

  1. Poisson regression with interactions
    By bobban in forum Regression Analysis
    Replies: 0
    Last Post: 05-05-2009, 04:53 PM
  2. Transformed Interactions in Regression
    By mandy_a41 in forum Regression Analysis
    Replies: 3
    Last Post: 03-25-2009, 07:42 PM
  3. Three way interactions in regression
    By sifar16 in forum Regression Analysis
    Replies: 4
    Last Post: 03-20-2009, 11:49 PM
  4. Regression - 2 and 3 way Interactions
    By bkplussk in forum Statistics
    Replies: 2
    Last Post: 11-30-2008, 07:53 PM
  5. Within/ between subjects interactions using regression
    By scf in forum Psychology Statistics
    Replies: 1
    Last Post: 06-29-2008, 07:50 AM

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