 Hey, everyone, welcome back. So remember last time we were talking a little bit about prediction. Prediction is one of the levels of scientific understanding. Prediction is very good for us. Well, it's not the most rigorous level of scientific understanding. You can actually do quite a lot with predictive value. So when we're talking about prediction, what we're typically referring to is correlation. And if you hear about scientific studies, especially psychological studies, you may hear things like one variable is correlated with another variable, or one phenomenon is associated with another phenomenon. Association is very often another way of saying correlation. When we're talking about finding relationships between two variables, very often we're talking about these correlations and we're talking about the correlation being a kind of mathematical interpretation or depiction of the strength of the relationship between two variables. So a correlation is going to tell you, it's this mathematical equivalent of both the direction of the relationship and the strength of the relationship. In this case, I'm talking about the correlational coefficient. So the values that I'm showing you here on the screen, I have a graph here on the left or up and down axis, also known as the y-axis, we see that there's one variable. And on the bottom axis, the one that goes left and right, horizontal, also known as the x-axis, we see that there are little data points that are depicted. And if you can see this or if I'm describing this for you, we see that the data starts pretty high early on and then as you go further into the study, as time goes on or as the x-value increases, the y-value decreases. So again, we're talking about the level to which one value predicts the other and we're also talking about a level of directionality. Knowing that correlational coefficient actually allows researchers to be able to communicate to one another and to be able to predict the value of an unknown variable. So while I'm showing you a number of data points in here, I could potentially use something called inferential statistics to infer the value and location of an unknown variable. If I know the x-value, I could potentially predict the value of something on the y-axis even if I didn't collect those data. Now talking about those relationships and the directionality, we sometimes have these terms that we use, positive correlation and negative correlation. When we say positive correlation, that's something where when you increase the value on the x-axis, you're gonna see a corresponding increase in the value on the y-axis. So as you go to the right, you're also gonna go up in the graph. So you tend to see this kind of 45 degree angle originating right there at zero, zero and kind of going up at a 45 degree angle. A perfect positive correlation, it's gonna be depicted as the whole value one as a correlational coefficient means that you can perfectly predict as you go up one unit on the x-axis and you're gonna go up one unit on the y-axis. A negative correlation is kind of the opposite where when you go up one unit on the x-axis, you're actually gonna go down a unit on the y-axis. So as one variable increases, the other decreases. A perfect negative correlation is negative one. And so again, that's the correlational coefficient. A negative correlation of negative one means that as you go up one unit on the x-axis, you go down one unit on the y. The closer you get to those whole values one and negative one, the stronger and more predictive your data are. When we say no correlation, we mean that you just simply can't predict knowing one variable what the other variable is gonna be. And so no correlation data is actually not super strong for us. And so remember just super quickly, those correlational coefficients range from negative one to one. The closer you get to negative one or one, the stronger your value is gonna be. If you have a negative one, correlational coefficient, that means you had a perfect negative correlation. If you have a positive one, correlational coefficient, you had a perfect positive correlation as one thing increased, the other increases the same amount. And then if you have no correlation whatsoever, it's gonna be close to the value zero. Finally, what I want to caution you with is though predictions are really valuable, though they can be really valuable, and sometimes that's the highest level of scientific understanding that we can get to. I want you guys to remember that correlation does not equal causation. And many people have heard this before, but what that means is it is entirely possible that you can have one variable that's increasing at the same time as another variable, and that relationship can be very, very strong. But that doesn't necessarily mean that changing one variable produces the change in the other variable. There is a video assigned in the study guide for this. But another example of this correlation does not equal causation is an idea of the fact that ice cream sales are positively correlated with drowning. And it doesn't mean that just because you're buying ice cream you're more likely to drown. It means that there's a third variable that occurs at the same time. So as the temperature increases, ice cream sales go up. As the temperature increases, people go swimming more often and there are more accidents and drownings. For that matter is the temperature increases because it tends to be summer and summer also produces more light. We also see that ice cream sales are associated with drownings and are also associated with home invasions. But that doesn't necessarily mean that having more ice cream gets y'all hopped up and makes you go into your neighbor's house and steal their TV. It just means that all of those things tend to co-vary together and one can help predict the other even though it doesn't necessarily cause it. If you really wanna know if you have a causal variable the only way you can possibly know that is by doing an experiment. And as time goes on we're gonna talk a lot about experiments. This time around we're gonna stick with this idea of a classic psychological study which tends to be a between groups design. So come on back, we'll talk about between groups designs and we'll talk about the importance of random assignment.