Hi Daniel, Great video, your papers and videos have been a huge help! I have a question about the metafor package and your subsequent scripts you have on your gitHub to run a meta-analysis in R. Does metafor/your script give weight to effect sizes, so studies with larger sample sizes carry more “weight” in the analyses than smaller studies?
CeeTee7 correct, studies with larger sample sizes (or to be more precise, studies with less variance) are given more weight in the summary effect size estimates when using these scripts
@@dsquintana Great, thanks for your help and for your content Daniel! May I also ask if the robust variation estimation (RVE) is suitable for large-scale meta-analyses, when accounting for multiple effect sizes from the same sample? I have seen in some papers that to account for this they pool these effect sizes to create one effect size per study/sample before the analysis, whereas the RVE seems to not require this manual pooling beforehand, instead clustering them and giving an output which you call the point estimate. Would this point estimate output also therefore be the main result you report for your analysis, as the point estimate RVE result is of independent samples, whereas the previous results are not of independent samples?
CeeTee7 correct, RVE is a good approach that I’ve used in the past to account for dependent effect sizes, such as this paper www.cambridge.org/core/journals/psychological-medicine/article/treatment-outcomes-for-anorexia-nervosa-a-systematic-review-and-metaanalysis-of-randomized-controlled-trials/AB05A8856C66EA5AD99AD9B2DF1C1F80
Hi Daniel! Thanks for your videos! I have a quick question that might just be silly.. But I have multiple calculated effect sizes (from different data points) per study/author and therefore have the same author and study listed multiple times. How do I go around this when creating a forest plot so that I don't list the same study multiple times? Do I consider each effect size a separate study (even though it is from the same study+author?) Thanks!
More give us more Daniel you are the hero of meta-analysts !!!
great Job! Dani
Thanks for the comment!
Hi Daniel,
Great video, your papers and videos have been a huge help! I have a question about the metafor package and your subsequent scripts you have on your gitHub to run a meta-analysis in R. Does metafor/your script give weight to effect sizes, so studies with larger sample sizes carry more “weight” in the analyses than smaller studies?
CeeTee7 correct, studies with larger sample sizes (or to be more precise, studies with less variance) are given more weight in the summary effect size estimates when using these scripts
@@dsquintana Great, thanks for your help and for your content Daniel!
May I also ask if the robust variation estimation (RVE) is suitable for large-scale meta-analyses, when accounting for multiple effect sizes from the same sample? I have seen in some papers that to account for this they pool these effect sizes to create one effect size per study/sample before the analysis, whereas the RVE seems to not require this manual pooling beforehand, instead clustering them and giving an output which you call the point estimate. Would this point estimate output also therefore be the main result you report for your analysis, as the point estimate RVE result is of independent samples, whereas the previous results are not of independent samples?
CeeTee7 correct, RVE is a good approach that I’ve used in the past to account for dependent effect sizes, such as this paper www.cambridge.org/core/journals/psychological-medicine/article/treatment-outcomes-for-anorexia-nervosa-a-systematic-review-and-metaanalysis-of-randomized-controlled-trials/AB05A8856C66EA5AD99AD9B2DF1C1F80
@@dsquintana Thanks very much for your help Daniel, it's greatly appreciated!
Hi Daniel! Thanks for your videos! I have a quick question that might just be silly.. But I have multiple calculated effect sizes (from different data points) per study/author and therefore have the same author and study listed multiple times. How do I go around this when creating a forest plot so that I don't list the same study multiple times? Do I consider each effect size a separate study (even though it is from the same study+author?) Thanks!