Okay, In this and the next several videos, we're going to be addressing this question. What should we do when our data fail to meet the assumptions of a model, of a statistical model. I'm going to be focusing in these videos on the assumptions of a general linear models. And there are four main assumptions that we need to concern ourselves with. And they're not all of equal importance. The most important is that we obtained our samples through random sampling. The reason for that is really straightforward. We want to make sure that our samples are an unbiased representation of the population that we're interested in. If our samples do not read properly represent the population they're interested in, then any conclusions that we might make from an experiment or an observational study are not necessarily going to be applicable to the population that you're actually interested in. The second D is that the data that you're dealing with are independent. These first two assumptions relate to study design. And we're not going to focus on them here. I'm just going to say one thing about non-independence and that is I want to refer you to a previous video. I dealt with pseudo replication for more discussion of what it means for data to be non-independent. There are two general solutions to avoid a non-independence. The first is to simply design an experiment that can avoid non-independence in the first place. The second is if your experiment does have the possibility for non-independent data, you want to use an appropriate analysis to account for this. For example, mixed effects models can be a useful tool and we discussed those in later videos. The last two assumptions involve the residuals from the general linear models. Our third assumption is that the variances are equal among our groups. If we're, if we have an experiment with different groups or, or that the variances are equal among along a range of a covariate if we have a covariate in a general linear model. And finally, there's the issue of normality. These two assumptions of equal variances, normality, they relate to the data. So these are assumptions that we can test while we're doing our statistical analysis. And we test these assumptions by examining the residuals from our general linear model. This raises the obvious question of, well, what do nice residuals or worrisome residuals actually look like? And that's what we're going to deal with in the next video. Thank you.