+ Reply to Thread
Results 1 to 2 of 2

Thread: Regression Misconceptions

  1. #1
    Points: 57, Level: 1
    Level completed: 14%, Points required for next Level: 43

    Posts
    5
    Thanks
    3
    Thanked 0 Times in 0 Posts

    Regression Misconceptions




    Dear all,

    Just wanted to clear some misconceptions I have about regression analysis. Your help would be highly appreciated

    1) One of the assumptions for a regression analysis is the normal distribution of residuals at every value of the predictor value.

    This does not mean that ALL the predictor values have to be standardized for regression analysis, right? It just means that I should check via P-P plot that the residuals for each variable are normally distributed - am I right?

    I got confused by the following information:

    "Regression assumes that variables have normal distributions. Non-normally distributed variables (highly skewed or kurtotic variables, or variables with substantial outliers) can distort relationships and significance tests. There are several pieces of information that are useful to the researcher in testing this assumption: visual inspection of data plots, skew, kurtosis, and P-P plots give researchers information about normality, and Kolmogorov-Smirnov tests provide inferential statistics on normality. Outliers can be identified either through visual inspection of histograms or frequency distributions, or by converting data to z-scores. " - http://pareonline.net/getvn.asp?n=2&v=8

    2) Am I right in assuming that conducting regression analysis (normal moderation analysis) on PROCESS should yield the same results than conducting a regression analysis on SPSS without the PROCESS plugin?

    Thank you!

  2. #2
    Cookie Scientist
    Points: 13,431, Level: 75
    Level completed: 46%, Points required for next Level: 219
    Jake's Avatar
    Location
    Austin, TX
    Posts
    1,293
    Thanks
    66
    Thanked 584 Times in 438 Posts

    Re: Regression Misconceptions


    Quote Originally Posted by majortom88 View Post
    This does not mean that ALL the predictor values have to be standardized for regression analysis, right? It just means that I should check via P-P plot that the residuals for each variable are normally distributed - am I right?
    This has nothing to do with standardization, it is about whether certain quantities are normally distributed. The assumption can be stated more simply as: the errors are normally distributed. Also note that there is not a separate set of residuals for each variable, as you suggested, rather there is only one set of residuals (for the outcome variable).

    For a more detailed refutation of the quote you posted, see here:
    http://www.talkstats.com/showthread....-TalkStatters!

    Quote Originally Posted by majortom88 View Post
    2) Am I right in assuming that conducting regression analysis (normal moderation analysis) on PROCESS should yield the same results than conducting a regression analysis on SPSS without the PROCESS plugin?
    In theory yes, but I have no idea how PROCESS is written or what exactly it does under the hood.
    “In God we trust. All others must bring data.”
    ~W. Edwards Deming

  3. The Following 2 Users Say Thank You to Jake For This Useful Post:

    CowboyBear (04-19-2016), majortom88 (04-21-2016)

+ 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