Move Heterogenous Calculated with airSlate SignNow

Move heterogenous payment

hello I'm Blair Johnson at the University of Connecticut and I'd like to talk a little bit about making sense of heterogeneity and meta analysis eventually I'll explain why heterogeneity appears in quotation marks so the goals of meta analytic statistics are to examine mean tendencies in the in the data that are gathered in a meta analysis we have to look at model fit statistics and we might want to look at moderators or sometimes they're called effect modifiers or subgroup analyses we'll have tests of significance for those moderators and we'll have model fit statistics and we might want to think a little bit about outliers but I think in this talk I'll do less with that I'll mention it in passing a few times so let's let's see how we can manage the subject of testing models and meta-analysis and now you see the point there so one of the necessary statistical ingredients you want to affect sizes and you're seeing I'm defining effect size as a T the common way of thinking about it so what you define them in one form not mixing with others the standardized mean difference log odds ratio Fisher Z transform of R or with metic means etc you can do a lot of different things with the Flex sizes put them in one form and then you're gonna have to have a variance calculated for each effect size and loops and Wilson provide a really tidy table of these equations those define the weights for analyses so you have it here's a representation of that table a lot of equations but essentially and in the second column first column you've got effect size type and the second column the statistic define in a simple way and then you've got two columns that are essentially related to the way it's the standard error and the inverse variance and a meta-analysis we use the inverse variance routinely although some some equations out there will use the standard error so it's good to have your hands on both of those I'll show you a couple of examples in the case of pearson product-moment R there's your effect size is the R and you can transform it to the Fisher's form of it using that equation and that's the fixed size stat and on the standard error and the inverse variance look like this very simple you can see how standard errors get smaller as you get bigger and bigger ends so what we're doing is valuing the studies well I'll show you that in a second but we're gonna value the studies more that have greater ends in other words they're more precise and the inverse variance for the are the inverse the Fisher transform is as you can see there's just n minus 3 so that's one example of the conventional statistics that are used for one effect size here's the standardized mean difference it's a little more complicated the effect size...