 In this video, we're going to do some simple parametric tests. We're going to do a t-test, a one-way analysis of variance, and simple linear regression. Let's get started. I'm going to go to Analyze, Compare Means, and I have an independent sample t-test. It's the one that I want to choose. Let's reset there, and let's see what we want to do. Let's have a look at the age column. Age is our test variable, and we want that to be grouped by whether a patient has diabetes or not. And that is going to be my grouping variable. I've got to tell SPSS what the data point values were, though. So, remember, I can only have two groups for a normal t-test, and the values there were yes and no. So, it's going to go down diabetes and make two groups at a finding yes and no values. I'm going to click Continue. The options there. I want 95% confidence intervals. That's okay. I'm going to hit OK, and there we go. We get a nice little table that gives us two tests, actually. We have the assumption for equal variances and the assumption for unequal variances. We didn't do something like a Bartlett's test here just to test for the assumption of equal variance or not. SPSS is going to give us both of those. We've got to decide which one to use. And if we move all the way across, if we look at the assumption for equal variances, we see students t-test p-value here of 0.425. If our alpha value was 0.05, this is not statistically significant difference between the two groups within and without diabetes, as far as the age variable is concerned. We see the mean difference and we see the 95% confidence intervals around the difference. So a simple students t-test. If we fail this assumption of equal variance, we see slightly different results and we've got to decide which one to use. Now, let's have more than two groups and we're still going to do parametric tests. So we're doing this under the assumption that our sample was taken from a population in which that variable was normally distributed. So we're really going for those assumptions. I'm not testing them here and you'd remember from the way that we simulated these data, not all of them that come from a normal distribution. So we're not looking into that at the moment. I'm purely doing some simple parametric tests without checking for the assumptions. So let's go to analyze. We are going to compare means and all the way down we find one way and over. Now in this instance, let's reset. And I want age to be my dependent variable and my independent factor is going to be grade. Dependent meaning it is going to predict or be going to be analyzed according to what is found inside of the grade. I'm not going to choose any of these. We'll just choose OK and there we go. We get our one way analysis of variance. We get our F statistic and we get our significant values. We see that we had a significant value of 0.736. Once again, if our alpha value was 0.05, we would not see this as statistically significant. Last up, let's just do some linear regression. First of all, let's just draw a graph. I'm going to choose the chart builder. OK, I'm just going to hit reset here. And what I'm interested in for linear regression is just a scatter plot. So let's just do a normal scatter plot. And we've got to have a numerical variable on each of these two axes. So let's take age and let's take temperature. What's the correlation between age and temperature? That's all I want for now. I'm going to click OK. And if we look at this chart age versus temperature, we see that there really isn't a correlation for this on our normal scatter plot. How would we analyze this? We would go to analyze and we would go down all the way to regression and we just want a simple linear regression. There we go, linear regression. And already you're populated. I do have age and heart rate. It doesn't matter which way around these are. My method is just not a normal inter method. If I go to statistics, I just want the estimates and the confidence intervals. That's quite OK. I'm not interested in a r squared value here. And I'm taking my assumptions for parametric tests. So we're talking about Pearson correlation here. And if we move all the way down, we see our p-value here under sick for significance. The heart rate there against the age with a p-value of 0.468 is not statistically significant. So that's just a p-value for a normal correlation, a linear correlation. So those would be the three standard parametric tests. Again, I have not tested for the assumptions that I could use parametric tests. And if I should rather have used non-parametric tests, we just skip those assumptions. But this is how easy it is in SPSS to do a simple t-test, a simple one-way ANOVA and simple linear regression.