+ Reply to Thread
Results 1 to 6 of 6

Thread: 22 enough for multiple regression?

  1. #1

    22 enough for multiple regression?




    Hello guys, i am doing a thesis on a given industry having only 22 enterprises. As the population size is small i have decided to go for census. But my worry is that can i conduct multiple regression analysis with only 22 participants? I have 7 independent variables and one dependent variable...

  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: 22 enough for multiple regression?

    So you have about 3 partcipants for each estimated parameter.
    I guess that this is very few.
    Some ideas can be found here and here

    With kind regards

    K.

  3. #3
    Points: 21, Level: 1
    Level completed: 41%, Points required for next Level: 29

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: 22 enough for multiple regression?

    Thanks K. I understand your point. But my problem is that 22 is all what i have. I just dont know what else i can do with this number...

  4. #4
    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: 22 enough for multiple regression?

    Choose 2 or 3 from your independent variables
    to include them into the analysis. Leave the rest
    out. You can choose the IVs on theoretical grounds,
    or because they are relevant for practical purposes.

    Resist the possible urge to play around with 7 IVs
    until you have identified those 2 which give the
    niciest results for those just 22 data points, since
    the practice of playing around almost inevitably
    gives mislesading and non-generalizable results.

    With kind regards

    K.

  5. #5
    TS Contributor
    Points: 13,936, Level: 76
    Level completed: 72%, Points required for next Level: 114
    jpkelley's Avatar
    Location
    Vancouver, BC, Canada
    Posts
    440
    Thanks
    17
    Thanked 90 Times in 84 Posts

    Re: 22 enough for multiple regression?

    I agree. Use a maximum of three predictors in your models. Also, don't try to include any interactions in your model. Stick with the three additive effects.

    One option might be Principal Components Analysis, but I think the sample size may be prohibitive as well. But I would first take Karabiner's advice and trim down your analysis to three predictors by looking at all bivariate associations. Start by identifying the strongest positive associations and omit one of those terms.

    Feel free to post your correlations and ask more advice about the process...

  6. #6
    Points: 21, Level: 1
    Level completed: 41%, Points required for next Level: 29

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: 22 enough for multiple regression?


    Thank you so much guys. You have given me valuable comments. Let me think about it and i will definitely come back to you.

+ 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