1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

T-test and effect size

  1. Nov 21, 2008 #1
    Hello! I hope it's okay to ask stats questions...
    I think the calculations are correct, but I would appreciate it if someone could check for me!
    These are the results from the exp. using repeated measures. My hypothesis is that the reuslts in condition 2 should be less than condition 1.
    Cond 1
    Cond 2

    I know working out the standard deviation is tedious, but I checked with excel and it matched (I used n-1, so I got the population, not just the sample).
    So, for condition 1:
    s.d.= 10.36895132
    standard error= 2.99325842
    and condition 2:
    mean= 40
    s.d.= 4.51260859854
    standard error= 1.30267789
    These (I think) are correct, its the t-test, p value and effect size I would greatly appreciate some feedback for.
    For the t-test I used the realted t-test, since its repeated measures (same participant in both):
    I got t= 3.346118674
    are the units sd?
    and DoF would be 11, and since I specified a direction, I got p=0.003261139 (one-tailed); which is significant? since P is less than 0.05?
    and for effect size:
    I got (d)= 1.567935986
    Can somone verify the t-test, effect size and p value for me? Please! I used fractions at some points where I could quickly get the data (I wrote it down), in ideal circumstances should I use fractions for all the 'inputs' (e.g. stand. dev of difference, etc.), if possible, since its more 'accurate'?
    Thanks! I now it's a lot to ask, but I would geratly appreciate it! I realise the formuale are slightly different depending on the narue of the experiment, I think I chose the right ones!
  2. jcsd
  3. Nov 22, 2008 #2


    Staff: Mentor

    It's been a while since I've done anything with statistics, but I think you might not be using the right form for t. The form I've included is for a two-sample t test, which seems to me the right one for your problem.

    Instead of the standard error, I think you might need to use a pooled standard deviation.

    [tex]t = \frac{X - Y}{S_p\sqrt{\frac{1}{12} + \frac{1}{12}}}[/tex]
    where X and Y are the sample means for the two conditions, and [tex]S_p[/tex] is the pooled s.d., which is given as
    [tex]\sqrt{\frac{(n - 1)s_x^2 + (m - 1)s_y^2}{n + m - 2}}[/tex]
    For your problem, n = m = 12.

    Your use of fractions and decimal values probably won't affect your values, since you apparently used quite a few decimal places in your calculations.

    The t distribution is close to a normal distribution with mean 0 and s.d. 1. For n around 30 there's no difference between the two distributions. That should answer your question about units of s.d.s.

    Hope this is helpful.
  4. Nov 25, 2008 #3
    Thanks for the response!
    I've got to do some reading... I haven't come across (or I don't think I have) the pooled s.d. yet.
    I kow the related and independent t-tests. The latter being the (mean 1 - mean 2)/sqrt{(variance 1/n1)+(variance 2/n2)}.
    Last edited: Nov 25, 2008
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Similar Threads - test effect size Date
Applying hypothesis test data collected (Statistics) Dec 1, 2017
Two path test Sep 29, 2017
Two sample t tests statistics Mar 12, 2017
A statistical question Feb 17, 2017
Help me derive the Doppler effect ratio Aug 31, 2016