Hello,
In relation to another recent, as yet unanswered, post (https://www.researchgate.net/post/How_to_calculate_comparable_effect_sizes_from_both_t-test_and_ANOVA_statistics), I am wondering how I can calculate the sample variance of an effect size, in order for me to then infer the confidence intervals.
So far, I have been calculating 'Cohen's d' effect sizes from studies' experiments, by using the t-value of two-sample t-tests and the sample size (n) for each group. I then convert the Cohen's d into an unbiased 'Hedge's g' effect size.
I understand that, normally, in order to calculate sample variance, I would also need to know the means of the two groups and their standard deviation. However, these are not reported in most studies when a t-test is calculated. Is there any way I can calculate the sample variance of my Hedge's g effect sizes without this information?
Many thanks,
Alex
You can calculate proportion of explained variance by squaring the t-statistic and dividing it by the same number plus the degrees of freedom. But for Cohens and variance, you do need mu and sigma
- Badges
- Science topic
- Similar topics
- Biological Science
- Neuroscience