I'm doing a meta-analysis of intervention research on argumentative writing. So far I have calculated the effect size of each stud and the unweighted and weighted means per treatment. I have also calculated the confidence intervals of the unweighted means per treatment. However, is there a way to calculate the confidence intervals that takes into account the weight of each effect size?
It is unclear as to what you are doing, What are unweighted and weighted means per treatment? How are they related to the calculation of your effect size? What is your effect size? What is being compared to what?
Thank you for your answer Gordon L Warren . To calculate unweighted mean effect size per treatment, I calculate the mean of all effect sizes that have the same treatment. To calculate the weighted mean effect size per treatment, I factor in the number of participants each study has. Therefore, a study that has a bigger sample size will have a bigger impact when calculating the weighted mean than one with a smaller sample size. Related to your last question, the effect sizes of each treatment are compared to each other to see which treatment has been more effective. I hope this information clarifies a bit more what I was asking.
- Badges
- Science topic
- Similar topics
- Correction