effect size

Moyu

New Member
#1
Hello everyone!
I'm currently a psychological student and i'm working on a meta-analysis, but I have a problem: many meta analysis uses the term "effect size". I've did some research on what is an effect size: i've understood that it's very important in meta analysis to compare research that don't have the same sample's size, scale... but i don't deeply understand what it is. Can somebody explain to me please?
Here is a part of the study that i'm currently working on:
"We examined the long-term effects of psychotherapy compared to control groups across 12 comparisons from 11 studies (in one study two types of psychotherapy were compared to control group). The results indicated that psychotherapy outperformed control groups at 6 months or longer after the beginning of the treatment of older adults with depressive symptoms (g=0.27; 95%CI: 0.16∼0.37). Heterogeneity was zero (95%CI: 0∼58) while there was no indication for publication bias."

Thanks for your help! :)
 

hlsmith

Not a robit
#2
Is that hedge's g, and heterorgeneity is the difference between studies - usually a value 0.50 or higher is a sign that you need to use a random effects model.
 

gdaem

New Member
#3
An effect size is a quantitative measure of something, and in the context of psychological research they go with p-values in NHST. An effect size is based on the power of your statistical test, and in psychology they encourage reporting an effect size. The effect size basically says 'There is a large effect in this analysis because Cohen's d is .86'.

There are many different methods of reporting an effect size and they are all based on different statistical properties. In the study you mentioned above they provide g, which I am assuming in Hedges' g, and is a measure based on the standardized difference. Cohen's d on the other hand is based on difference between two means divided by the standard deviation of the data.

In the grand scheme of things, unless you are really getting into conducting a meta-analysis, I would focus more on the appropriate test and the accepted critical value for each. Each effect size calculation has a respective value that goes with it.

Although not an academic resource, Wikipedia has a good overview of effect sizes that is worth checking out. https://en.wikipedia.org/wiki/Effect_size
 

Moyu

New Member
#4
Thank a lot to both of you!
So if I've understood what you've said, in the study that I've mentioned, g results correspond to the mean of the effects sizes of each study? But if g=0.27 why they say that psychotherapy outperformed control group? ( 0.2 corresponds to a small effect size right?, so the different between control group and psychotherapy group is not very large?). I'm not sure that I've understood in the right way...