 I think we'll start this out with an example of what the Helmholtz chain is going to have to eat. Yeah, I wonder why I'm alive. You're grabbed by the function of a bullet lighter. Yeah. Well, in order to understand this, you don't have to do anything. Which is kind of lovely. So now I can no longer get to my bookers. Here is that you're probably wondering why I'm blowing on leaves in order to make a funny whistle. Well, naturally, parametric analysis. What else? I mean, you're like, what the flip does he talk about? So there seems to be some confusion in our field, at least confusion among how we talk about these things with regard to the types of analyses that exist within the field. And I'm going to leave this here because I'm going to come back to that. I promise you it is relevant to the video. So there's lots of different types of analyses. So we need to look at our experimental design first in order to understand what I'm really talking about. So the first things first, well, we're talking about experiments. Remember, we talked about experimental control. We talked about the type of design that you choose that will maximize the experimental control and also address the particular research question that you have, the experimental question, which I completely forgot to talk about in the video about experimental questions, which is just completely ironic. So the experimental question is the thing that you're trying to find out. What effect does sitting on a rock have on my bum? Especially given it's high noon, it's rather warm out here and my bum's getting kind of toasty because the rock is warm. So what effect does that have? I don't know, I'll tell you in about 10 minutes after the video is over. The experimental question then would be what effect does the rock have, right? The design type, I don't know what design this is. An AB design, I suppose I could get off the rock and come back on the rock and we'd have an AB, AB design. I mean, we could do that whole thing, but that's not what this video is about. This video is about the types of analyses that we do while we are in order to analyze our data. So when we're doing single subject research, keep this in mind, single subject research is about using a person as their own control. So when we talk about a person as their own control, we're not going to create a control group of 50 people and then compare that to an experimental group of 50 people. We're going to use one person or a couple of people, whatever, and we're going to look at their data, their behavior, and we're going to see what they are under one condition and we're going to see what they are under a different condition and what they are under a different condition and so on and so forth. That's within subjects type research. But we're doing it focusing on one person, so a single subject design. We do that usually using a small N type research. There's not very many people in our study because it's hard to analyze data that way. So if we were doing something different, though, if we had a different type of experimental question, if we were not in behavior analysis and if we were not talking about single subject analysis, if we were talking about group design, group research, then what we would be focusing on is something called parametric type designs, right? See, I'm getting closer back to this. So when we think about group research and we think about large experiments that may have two conditions, let's say, let's say a baseline. Oops, I'm so well trained in behavior analysis. That's all I can think about. A control group instead of a baseline. So we have a control group and then we would have an experimental group, right? So that experimental group is a large group of people and so is the control group. So we're going to average their data and we're going to average their data and we're going to make a comparison between those two conditions to see is there a significant difference based on the, is there a significant difference between the groups on whatever dependent variable it is we were measuring, right? So how do we do that? We don't do that by just looking at a graph. We do that with a very fancy set of statistical tools that are based on the field of parametric analysis. Parametric analysis is literally the types of statistics and the types of analytical work that you do using large groups of people assuming they're normally distributed. That's the key, right? So pretend that my hand is a normal distribution. I know it's not quite right, but you get the idea, right? So you get this little bell shaped curve. If we're in a psychology experiment that has large groups, we would measure whether or not they have, the control group has a normal distribution. Then we would compare whether or not the intervention or the intervention condition or the experimental condition has a normal distribution. If they both do, then we can do this. We can make a comparison and here's what parametric analysis does. Now, let's say I have them overlapped, right? So the assumption for any experiment is that the null hypothesis is the case. We're trying to falsify that the null hypothesis is the case. So with parametric analysis, what we do is we say, what's the mean for the control group and what's the mean for the experimental group? If they're the same, then there's no difference between their groups. But here's what really happens in an experiment. We drive them apart from each other, right? So that independent variable drives and separates the means, which is about where my middle finger is and I don't intend to flip you off. Anyway, so it drives those means apart from each other. If those means are far enough apart, if the data doesn't overlap too much, then we have what's called a statistically significant difference. If so, again, parametric analysis, large group designs, you need to have normal distributions. If they're not normal, you can use non-parametric analyses. But parametric and non-parametric analyses are literally used for large group research. There are occasions when we might use them in behavior analysis, but it is not common, right? There are reasons to use them in behavior analysis, but I'm not going to go over them right now. That's for like advanced level 400, 7000 stuff, but I'm not going to go there with this video. So I just want you to know the difference between parametric analysis and single-subject analysis, what we're focusing on, focusing on how to analyze the data. So you're probably wondering about the leaf? Well, it's simple. Grab the wrong kind of leaf, and you can't make that sound, right? So I've tested a lot of leaves over my life, right? And all of the ones that have that sort of shape go into one group. That would be our experimental group. And all the ones that have a different shape would go into a different group. And do they produce a different sound, or do they make the whistle that I was after? And the answer is yes. Let's try it here. Let's try a little bit of replication. Although, shh, don't tell anybody it was a single subject. Yep, we got it. So anyway, parametric analysis, we could have analyzed that with thousands of leaves and thousands of other leaves and made all these fun comparisons and all that. So parametric, again, just normal distributed. It's parametric. So that's really what parametric is all about. Don't overthink it, because we don't use it too often in our field, but there are appropriate times which we'll cover some other time. If you like our videos and you have yet to figure out how we've organized them, maybe you should go look at the playlist. It's kind of cool. They cover pretty much everything. In fact, we've got really probably a lot, maybe even too many playlists. But it's there for your organizational purposes. Use it. Please.