 Welcome everybody. My name is Kyle Hamilton. I'm doing my presentation today on JMOVI as a platform for running a meta-analysis. A little bit about myself. I'm a fifth year PhD candidate in the health psychology program at UC Merced. My advisor is Dr. Linda Cameron. I'm affiliated with my school's Health Sciences Research Institute and the Nicotine Cannabis Policy Center. My substantive research focuses on adverse childhood experiences, so child abuse, child neglect, electronic cigarettes, health communications, and I put a lot of work into evidence synthesis and meta-analysis, and I do a lot of work in the R programming language. I should be graduating pretty soon with my PhD. COVID kind of mucked it up a bit, but I'm looking for a postdoc. I'm really great with R. I love evidence synthesis. I have great teaching evaluations. I'll go anywhere. So if you're interested or you know something, feel free to send me a message either on my email, my DMs are open on Twitter, or you can see my CV and learn more about me on my website at KyleHamilton.com. So real quick, what is JMOVI? JMOVI is a free and open source alternative to SPSS. It was first released in late 2016, and it's used in over 100 universities undergraduate programs for teaching statistics. And in the background, JMOVI uses R to perform all of its analyses. And here are all the actual developers of JMOVI, and they're really great. They helped me a lot when I was developing major and figuring out how to make everything work. As a quick aside, Datalab.cc is a really great introductory level resource for if you want to learn how to use JMOVI. It has over four hours of instruction with a bunch of creative, common videos. They're really great. Most of the stuff you do in JMOVI is pretty intuitive, but if you're looking for resources, they're available, including this great book by Daniel Navarro and David Foxcroft. It's a great textbook. It's a free PDF. If you're looking for something with a little bit more statistics background, take a look. It's really great. I enjoyed it. So JMOVI is based, sorry. Major is based on the R package metaphor by Wolfgang Bischbauer. It allows for fixed effect, random effect and mixed effect models. It allows you to produce reproducible results, both JMOVI and R, and it's usable as an R package in the upcoming release. So a JMOVI module is just like an R package. In this case, you can load a load major into a native session of R, and you'll be able to reproduce the exact same results you were able to do in JMOVI. As a quick look over the packages and the underlying features of major, for MetAnalytic models, we see that most of the work here is done by Metafor for both univariate and multivariate models. Bayes Meta is used for some of the univariate Bayesian models that are available. A big addition to this last iteration of major was the addition of a bunch of publication bias assessment tools, including P-uniform and P-curve, as well as toaster from Daniel Lakin. Metafor, again, produces a lot of additional publication bias features, including Eggers Regression, Trim and Fill. In the upcoming version of Metafor and the currently developers edition, selection models, including the Vivei Hedges weight function model and other step function models are available. As for graphics, Metafor also plays a big part of this in making the funnel plots, force plots, as well as influence and likelihood plots. And the data handling between the JMOVI platform and the R engine in the background is handled by the JMV core package, which is produced by the people that have made the rest of the JMOVI ecosystem. So here we are now in JMOVI. If you go up here to this top ribbon, you can see a bunch of different statistical tests and options for you. In our case, we are going to be doing a meta-analysis of mean differences from Norman 1999. Going into major, going down to mean differences. We see that it gives us a warning saying that we need to have a sample size, the mean, standard deviation study labels in order to actually run the analysis. That's not too difficult. We have our sample size from our first group, our first group's mean, our first group's standard deviation. Then we have the second group sample size, second mean, and the second group standard deviation. And if we go to source, put that in the label, and we get our results. So we get our intercept for our random effects model up here in K. We can see that we have how many studies were included in the sample. We have our P values. We have whatever we use for our tau squared estimator, heterogeneity statistics. And also from the reporter function in metaphor, now we have an output that tells us, explains what the model means to us and some of the other verbiage that we would need. Scrolling down, we have a forest plot, our publication bias assessment, our funnel plot. And if we go here to the left, we have different options. We can change the model estimator from restricted maximum likelihood to maximum likelihood. If we come to the top, we see that that changes. We also can include a moderator if we wanted to. If we look at the plots and come down here to the forest plot, we can change things like the prediction interval, or we can change how big we want the actual plot to come out. Same with funnel plots down here. We can either do a contour enhanced funnel plot, or we can change the predictor instead of standard error to be in sample size or sampling variance. Looking at publication bias assessment, we can look at adding the P-curve output or the P-uniform output. So if we scroll down, here's our P-curve output and the test results for our P-uniform, which this one didn't converge or was an error. But that's okay. We also can look at our equivalence test. So if we want to look at what the toaster package produces, it will give us a small explanation as well as a plot. And we also can put in diagnostic plots if we wanted to. So if we want to see if there is a particular study that the external standardized residuals say that there is a problem with or cooks distance, etc. So at the end of all of this, if we want to take this and actually put it into like a Word document, all we have to do is right click, copy, go to Word, and just paste it. And that's all there for us. The tables are editable as actual tables. The graphics are all there in their correct form. And we can change some of the words around if we need to, including any of the actual tables we need for publication. And yeah, so that's a brief overview of major. This is the developer's edition, so there will be a little bit more work that has to get done before it's finished. But hopefully they'll be taken care of by the end of this semester. Well, thanks a lot.