In this video, I'm going to demonstrate how to perform a power analysis. Specifically, I'm going to show you how to perform a power analysis for a two-sample t-test. Where we're going to do an a priori power analysis for a two-sample t-test. There are two general methods for performing power analyses that one method is to use simulations which are very useful when you have complex experimental designs. In our case, since we're imagining the experiments that can be analyzed with a two-sample t-test. That's far too simplest scenario to warrant simulations. And so in this case, we can use existing software. And I'm going to demonstrate how to perform our power analysis using the software g power. G power is freely downloadable from this website, and it's available both for Mac and for Windows operating system, operating systems. And the website also has tutorials available which greatly facilitate using g power. Now if we open G power, you'll see a display very much like this. And to perform our power analysis, we basically just have to fill in the options at the bottom of this screen. And the first choice that we have to make involves choosing the type of power analysis we want to perform. And when we click on that tab, you'll see that there's a number of types of power analysis at g power can perform, including a priori, post-talk and other types of power analysis as well. We're just going to choose an a priori power analysis. In the next decision that we're going to make is we're going to choose the type of test family wants to go to. And so we're going to go up and click on this tab. And when we do, we can see that that g power can perform power analysis for a variety of test families and create an F tests, Chi-squared tests, et cetera. Since we want through a t-test, we're going to choose the t-tests family. Once you've chosen that, our g power is going to make a number of tests that for all fall within the t-test family available within this tab, our statistical tests. So once you've chosen a t-test family, we can then go to the tab beside it under the option statistical test. And you'll see that there's a wide variety of tests that all fall within the t-test family, including non-parametric tests and regression and correlation. We can do a paired t-test, we can do one sample t-tests, et cetera. Although these options I've just pointed to you where for the non-parametric versions, a paired t-test I should have pointed up here, apologies. In our case, we just want to do a simple two-sample t-test. So we're going to choose this option where we're comparing the difference between two independent means. So where we have two groups. And once you've chosen that, we've completed the decisions at the top part of at the top part of our of the lower portion of the screen. And now our job is to fill in these options at the bottom. And I'm going to start by just highlighting the bottom three windows here. So the window in the middle is where you can specify alpha or your error probability, and that refers to your type one error rate. And we're going to choose a type I error rate of 0.05, which is standard in the biological sciences. Although you might want to choose a more stringent error rate, something like 0.005, which is recommended by some authors. The second window from the bottom is where we can specify the amount of power that we want in our experiment to have. And you can just fill in whatever number you want here. Between 01, I've chosen a power of 95%, which is generally considered a high powered experiment. The last option at the bottom here is where we can specify the allocation ratio for the sample sizes if our two groups. So when the allocation ratio of N2 over N1, this roof just refers to the ratio of the sample sizes in our two different groups. And if we want, our sample size has to be the same in our two different groups, then we would just fill in a one here as I've done. Because if the sample size in group 2 is the same as the sample size and group one. Then dividing one number by the same value gives you one. Now we filled in most of the options down here, and we're going to focus on specifying effect size. And here you can see we've got the effect size d. And d here refers to Cohen's d, which is the effect size for this type of analysis. So it's appropriate for t-tests. And when specifying this, you could simply input a value here. Based on, for example, some of the recommendations made by Cohen, where Cohen has made recommendations for effect sizes for deed it specif that corresponds to low, medium and high, or small, medium and large effect sizes. Or we can go to this button here on the left where we can say determine. And if we click on that button, as we're going to imagine we've just done. Then another window will pop out on the right, which allows us to specify the means and standard deviations for that we expect to have in our experiment. So I'm just filling in the options down here at the bottom, we're assuming our sample sizes for our two groups are equal. And I've specified the mean for the first group to be 14.5 and the mean for the second group to be 16. And then I've said that the standard deviation within each of those groups is three. And once I've done that, I can simply click on this bottom option here it says Calculate and transferred to the main window. And that will cause the, the power that corresponds to these values here to up here in this position here, which in this case is equal to 0.5. Our last option is to choose whether or not we want to have a one-tailed test or two-tailed test. And I would advise that in the vast majority of circumstances, we want to choose a two-tailed test, as I've done here. Once you've done that, you can just press Return on your computer and output like this will result. And you can see here that if we want to have 95 percent power with the means and standard deviations that we've specified. We need to have a sample size of a 105 individuals within each of our two different groups, which depending on your study system, maybe a lot or maybe very manageable or even small. So that's how to perform a power analysis for a two-sample t-test. See me want to do an, a priori analysis and perform it in G power. I hope this was useful. Thank you.