+ Reply to Thread
Results 1 to 3 of 3

Thread: Reporting Multiple Regression with Categorical Variable(s)

  1. #1
    Points: 220, Level: 4
    Level completed: 40%, Points required for next Level: 30

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Reporting Multiple Regression with Categorical Variable(s)




    Hi everyone,

    I'm overthinking one of my analyses and need a community sanity check.

    For simplicity's sake, let's say that my basic multiple regression model I am using for the first iteration of my analysis currently contains the following variables:

    Step 1: Centered continuous variable (IV 1)
    Step 2: Categorical variable (conditions coded 1, 2, and 3), (IV 2)
    Step 3: Interaction between IV 1 and IV 2

    In a second iteration of this model, IV 2 would be replaced by two dummy codes and then basically the same analyses would be conducted with those dummy codes. So, it would be:

    Step 1: Centered continuous variable (IV 1)
    Step 2: Two dummy codes in place of original categorical variable (Dummy1, Dummy2)
    Step 3: Interactions between IV 1 and each of the dummy codes

    I am being asked to analyze and report both iterations of this model.

    My question is: What is the purpose/utility of reporting the first iteration of this analysis, and how does that data inform how you proceed from there?

    Does iteration #1 provide main effects information for the categorical variable, while iteration #2 provides information on the contrasts between conditions? If so, what happens when the "main effect" is not significant but Dummy1 is (not Dummy2)? Do I not report the dummy code results if the "overall" categorical variable from the first iteration is not significant--even though the effects of the dummy codes (one sig, one not sig) may be "cancelling each other out"? Or am I not thinking about this correctly?

    Thanks!

    tbf
    Last edited by tophbeifong; 08-16-2016 at 11:22 AM.

  2. #2
    Points: 2,462, Level: 30
    Level completed: 8%, Points required for next Level: 138

    Posts
    200
    Thanks
    20
    Thanked 48 Times in 43 Posts

    Re: Reporting Multiple Regression with Categorical Variable(s)

    Hi, I don't understand what " IV 2 would be replaced by two dummy codes" means. Since IV 2 is already a categorical variable with 3 categories, these categories would be already represented as a dummy variable in the context of a regression analysis, represented by 3 numbers. So how exactly is the coding in the second model for this variable?

  3. #3
    Points: 220, Level: 4
    Level completed: 40%, Points required for next Level: 30

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Reporting Multiple Regression with Categorical Variable(s)


    Hi mmercker,

    In the first iteration of this analysis, the categorical variable contains the values 1, 2, and 3 (random numbers associated with experimental conditions 1, 2, and 3 that participants had been randomly assigned to for the study).

    In the second iteration of this analysis, I created two dummy codes so that I could directly examine comparisons between my experimental conditions and my control condition. Conditions 1 and 2 are my experimental conditions, and condition 3 is my control condition. Thus, I created Dummy1 (comparing condition 1--recoded as 1--with my control condition, coded as 0). Then I created Dummy2 (comparing condition 2--recoded as 1--with my control condition, coded as 0).

    Does that make more sense? I can provide more details if needed.

+ 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