+ Reply to Thread
Results 1 to 4 of 4

Thread: Independent t-test - interpreting significance if sizes of samples differ greatly

  1. #1
    Points: 367, Level: 7
    Level completed: 34%, Points required for next Level: 33

    Posts
    16
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Independent t-test - interpreting significance if sizes of samples differ greatly




    Hello,

    I ran an independent t-test on two samples (composite scores of a health assessment for two different groups of youth by age category).

    Group 1: N=1,296, Mean=17.47, STDEV=3.49
    Group 2: N=628, Mean=19.49, STDEV=3.93

    Levene's Test significant at p=.000. Therefore, equality of means NOT assumed.
    Reading that line (equality of means not assumed), I see: T=-10.92, 2-tailed sig @ p =.000. Mean diff=-2.01, 95% CI = -2.38 to -1.65.

    Can I use these results or are they invalid because my sample sizes are vastly different (one is twice the size of the other)?

    Thank you,

  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: Independent t-test - interpreting significance if sizes of samples differ greatly

    Can I use these results or are they invalid because my sample sizes are vastly different (one is twice the size of the other)?
    If unequal variances are associated with unequal sample sizes, then problems can arise. The software has already corrected the calculations, though (Welch test). Your results can be trusted, AFAICS.

    With kind regards

    K.

  3. #3
    TS Contributor
    Points: 6,786, Level: 54
    Level completed: 18%, Points required for next Level: 164

    Location
    Sweden
    Posts
    524
    Thanks
    44
    Thanked 112 Times in 100 Posts

    Re: Independent t-test - interpreting significance if sizes of samples differ greatly

    Quote Originally Posted by Karabiner View Post
    The software has already corrected the calculations, though (Welch test).
    Are you sure about that? In Stata, for example, you actively need to choose to run a Welch test if that is what you want.

  4. #4
    TS Contributor
    Points: 18,889, Level: 87
    Level completed: 8%, Points required for next Level: 461
    CowboyBear's Avatar
    Location
    New Zealand
    Posts
    2,062
    Thanks
    121
    Thanked 427 Times in 328 Posts

    Re: Independent t-test - interpreting significance if sizes of samples differ greatly


    This bit:

    Reading that line (equality of means not assumed), I see: T=-10.92, 2-tailed sig @ p =.000. Mean diff=-2.01, 95% CI = -2.38 to -1.65.
    Suggests that it is a Welch test. In SPSS by default you get both conventional and Welch tests, in a table with rows labelled as equality of variance (not means!) assumed and equality of variances assumed. That's presumably where the OP is getting this from. (In R, Welch is the default, incidentally).

    Anyway, Karabiner is right, the unbalanced sample size doesn't change the interpretation of the result

  5. The Following User Says Thank You to CowboyBear For This Useful Post:

    Englund (04-19-2016)

+ 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