+ Reply to Thread
Results 1 to 2 of 2

Thread: Regression with two categorical predictors and 1 control/covariate?

  1. #1
    Points: 167, Level: 3
    Level completed: 34%, Points required for next Level: 33

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Regression with two categorical predictors and 1 control/covariate?




    Hello,

    I'm in need of help as these analyses I describe are for my thesis so any input would be greatly appreciated. By the way, I am using IBM Statistics SPSS 19.

    I have pretest and posttest measures on the dependent variable (mood) and I am interested in determining whether mood has changed significantly in various experimental groups. There were two experimental manipulations. Participants either interacted with a negative or neutral actor (actor mood condition) and were given an independence prime, interdependent prime or no prime at all (prime condition). Time 2 mood was measured after an experimental manipulation. In the design, there are two between-subject factors (actor mood and prime) and the first time I analyzed the data, I considered time (pre vs. post) as a within-subjects factor and analyzed the data using a split plot/mixed design ANOVA. Some of the assumptions were violated when I did this (e.g., Box's Test of Equality of Covariance matrices). I also tried analyzing the data using an ANCOVA and inputted the pretest mood measure as a covariate but I think there are problems with doing this because it looks as though the covariate is not independent of experimental treatment groups (on one of the between subject factors). I think the split plot/mixed design ANOVA method is preferred but given that some of the assumptions were violated, I was wondering if I could analyze the data in terms of a hierarchical linear regression (with two categorical variables i.e., the between subjects/treatment factors and one control variable i.e., the pre-test mood measure)? Does it make sense to do it this way? Would this add anything to using the ANCOVA given that I've read the ANCOVA has many similarities to doing a regression?


    Thank you!! Any input would be helpful! Please let me know if anything needs to be clarified.

  2. #2
    TS Contributor
    Points: 17,749, Level: 84
    Level completed: 80%, Points required for next Level: 101
    Karabiner's Avatar
    Location
    FC Schalke 04, Germany
    Posts
    2,540
    Thanks
    56
    Thanked 640 Times in 602 Posts

    Re: Regression with two categorical predictors and 1 control/covariate?


    Some of the assumptions were violated when I did this (e.g., Box's Test of Equality of Covariance matrices).
    That is not necessarily a problem:
    https://en.wikiversity.org/wiki/Box%27s_M

    Kind regards

    K.

+ 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