 This playlist contains a set of videos related to our article on ignoring the random effects assumption in multi-level models, review, critique and recommendations. The first video is a general video made by John Antonakis and myself and we go over the entire paper here, talk about why we wrote this paper and explain the most important concepts. So you should probably start by looking at that video. The remaining videos are more focused on specific topics and also give you some background information on some concepts in multi-level modeling. So the first video is the one made by John and I. The second one is this introductory video and then we have 14 other videos that explain different concepts. The next three videos are generally about multi-level data. So if you are new to multi-level data, if you are new to levels of variation, if you don't know what interclass correlation is, or if the concept of unobserted or gen 80 is not familiar to you, then these three videos are probably helpful when you try to understand what our paper is about. The next set of videos, these two videos are about what is the impact of non-independence of observations on regression analysis and how do we deal with the non-independence of observations by using cluster robust or just robust standard errors. The key here is to understand that if you are just interested in estimating the population average effect which the regression gives you, you don't necessarily need to do any fancy stuff like multi-level modeling or panel data regression using GLS. So this is the first set that explains why you should care about non-independence of observations in regression analysis and what it means that regression assumes that the observations are independent. The video about heteroscedasticism is consistent and cluster robust standard errors a bit technical, but I hope that those technical details will help you to understand the issue on a more fundamental level. Then I'll talk about how we can model different effects from the data using normal regression analysis. In multi-level data we can estimate four different kinds of effects. Population average effect is whatever regression gives you if you ignore clustering, then you have the within effect which is estimated typically by eliminating a cluster effect. Then we have the contextual effect which refers to an effect of that others in the same context have on an individual level outcome. The between effect is the effect of cluster mean of one variable on cluster mean of another variable. I'll talk about how you estimate these effects in just normal regression context. In most cases you are actually going to be okay by just using normal regression analysis and these centering techniques and cluster robust standard errors if required. The next set of videos are about more advanced techniques so I'll cover what is a mixed effects model, what's a random part, what's a fixed part, what's a fixed effect, what's a random effect. Then I'll talk about GLS fixed effects and GLS random effects estimators. This is something that if you for example use data, if you run XD-reg this is what the command does. I'll explain a bit about the idea of the GLS estimation and why one would like to use that kind of techniques and when. Then finally I'll talk about correlated random effects models which is an application of these cluster means in a normal regression analysis or a GLS or multi-level model. Finally we have two videos about how to test the random effects assumption. So we have generally a set of three tests that we talk about in the paper. We have the Hausmann test, we have Wald test and then we have a likelihood ratio test. I have a separate video about a general Wald test and a general likelihood ratio test but those are general tests so they are not specific to this testing for this random effect assumption. However what is fairly specific or the most well-known test for the random effect assumption is the Hausmann test. If you look at management journals they almost exclusively use Hausmann test for the purpose of testing random effects assumption. Therefore it's a good idea to understand the logic behind that test because its application is more specific. And finally I have one video that gives you a summary of the main takeaway of the paper. So how do you choose an analysis technique when you work with multi-level data and you want to estimate the within effect and possibly the contextual effect or the between effect of a variable. I hope that you enjoyed the paper and if you find these videos or the paper useful please cite that identify for the paper and you can always find the most recent version from here.