Dr. Michael Borenstein, one of the authors of Introduction to Meta-Analysis, is widely recognized for his ability to make statistical concepts accessible to researchers as well as to statisticians. He has lectured widely on meta-analysis, including at the NIH, CDC, and FDA.
➣ Take a meta-analysis course: meta-analaysis-workshops.com/
Comprehensive Meta-Analysis is a powerful computer program for meta-analysis. The program combines ease of use with a wide array of computational options and sophisticated graphics.
➣ Meta-analysis software - download your free trial now: meta-analysis.com/
Some elements of this channel are available under a creative commons 4.0 license, including some icons: fontawesome.com/license
Thank you for the video , professor. I would like to know whether the "y-axis" of the prediction interval stands for something just as the x-axis shows event rate, risk ratio etc.
Thanks for your interest. If you wish to take the course through Columbia University this summer, visit bit.ly/3piGP7j and search for "meta" at the top. The Meta-Analysis Workshops site provides just the basic course right now. The advanced should be available in 2024.
Thank you very much for such a handy application and your effort in educating the meta-analysis. You mention that the effect size should be chosen by the fact that which direction is our favourable. However, there seems to be a mistake. When you choose a single effect direction (for example positive), all the SMDs become positive which clearly is wrong. By the way, I'm using CMA v3.
It seems that the narrator is reading a written script, because I can hear he's turning the pages. This makes the flow a little fast. I really need 100% focus.
if Tau-squared is 0, do we get any prediction interval? when I run a check on the software, it just shows that studies share a common effect size.. does it mean that there will be no PI available? thanks for all the help
Hi Sir, if Hausman test indicates that fixed model is more appropriate than random effect model, and if in that case, in data time period (T) > cross section units (N), which FEM is to be chosen: time (T) FEM or Cross section (N) FEM?
As an MD and clinical researcher for 5 years, i observed that some MDs/physicians really have no/little idea on how to run statistical tests for research (which is understandable since statistics is really not part of the MD curriculum) so they are wholly/partially dependent on their statisticians. There are times also when the statisticians have the right idea / executed the tests correctly but the clinician interpreted the methods and/or results differently. 😅😅😅
Let me add three more to the list: eyeballing funnel plots to determine whether they're asymmetric, treating Egger's test as though it detected publication bias, using trim/fill or PET PEESE on highly heterogeneous studies
这个视频不断的让我想到了某国的大部分教授们，总是在不断地强调结果是不是显著，怎么巧妙的处理就可以让结果看着更好看……这种模式下学了很多甚至变得“很厉害”成为了“大佬”一头雾水的开始分享怎么做一篇“优秀”文章的经验，sucks。Anyway，这真的是一篇perfect lecture，highest respect for this vedio and the professor！
So why do the majority of studies justify their review by claiming that because I² is low, effect size variations are low? We can further use prediction intervals solely for that reason and avoid making any misleading assumptions about effect size variations. Just read "Basics of meta-analysis: I² is not an absolute measure of heterogeneity., (Borenstein et al, 2017)". Thanks for the video about it too, it was really helpful.
Hi @jamessakeah9791 , did you find any answers to your question? I am also interesting in comparing different results for females and males and any information about how to run the meta-analysis would be really helpful!
Hi, I am struggling to run a metanalysis. It is my first time. I have continuous data with mean, sample size, and standard deviation for each study (without means for control and exposure groups.) Any ideas on the appropriate argument suitable for this type of data?
Thus great indeed!, One is able to validate based on number of studies, but how can you get access to unpublished studies with all required parameters?
Thank you so much for this much-needed video.. many of the concepts are crystal clear now.. But I have a doubt...how should we interpret very high values of Isquare like more than 90 or sometimes 100? Is there any chance of such results in meta-analysis...
