Author Topic: Are two hedges g effect sizes significantly different?  (Read 9611 times)

0 Members and 1 Guest are viewing this topic.

Offline Dev

  • Administrator
  • Sr. Member
  • *****
  • Posts: 63
  • Karma: 2
  • Forum Administrator
    • ClinTools
Are two hedges g effect sizes significantly different?
« Reply #1 on: 14 February, 2007, 06:01:07 pm »
We recently had the below question:

>On the extra statistics tab for effect size generator
>I'm trying to use the function "are two hedges g effect sizes significantly
>different?"

>However, it is asking for the t-test statistic for group a and group b, but
>I do not understand where to get this t-statistic

>I am enjoying your software

>Thank you very much.

Thanks for the feedback.

Each effect size (group a's and group b's) should come from two means and standard deviations. From these it is possible to compute a t-statistic. For example, suppose you had an effect size (for group a) of 0.96518, and this came from Mean1 = 25, SD1 = 4, Mean2 = 20, SD2 = 6 and you had 25 people in each group. Putting this into the effect size generator's main window and computing the effect size (g=0.96518) and then clicking on the statistics tab tells us that the t-value associated with this effect size (and n=25  in group1 and 25 in group2) is 3.46688. Opening the procedure for "Are two hedges g effect sizes significantly different?", this figure would then be put in the text box as the t-statitic associated with a hedges' g of 0.96518 and na=25 and nb=25. Then you would do the same for the other effect size (group b) and hit the 'do it' button. Does this explain things ok?  ???
Best
Dev
« Last Edit: 14 February, 2007, 06:06:47 pm by Dev »
I am logged on, therefore I am.