If the significance value is greater than the alpha value (we’ll use.
SPSS runs two statistical tests of normality – Kolmogorov-Smirnov and Shapiro-Wilk. Here’s what you need to assess whether your data distribution is normal. The Explore option in SPSS produces quite a lot of output. You’re now ready to test whether your data is normally distributed. This should now look something like this.
Now click Continue, which will take you back to the Explore dialog box. In this box, you want to make sure that the Normality plots with tests option is ticked, and it’s also sensible to select both descriptive statistics options (Stem-and-leaf and Histogram). Once you’ve got the variable you want to test for normality into the Dependent List box, you should click the Plots button. However, since we can perfectly well test for normality without adding in this extra complexity, we’ll just leave the box empty. In our example, Dog Owner, our independent variable, has two levels – owner and non-owner – so we could add Dog Owner to the Factor List box, and look at our dependent variable split on that basis. The Factor List box allows you to split your dependent variable on the basis of the different levels of your independent variable(s). You can either drag and drop, or use the blue arrow in the middle. To begin, click Analyze -> Descriptive Statistics -> Explore… This will bring up the Explore dialog box, as below.įirst, you’ve got to get the Frisbee Throwing Distance variable over from the left box into the Dependent List box. Our example data, displayed above in SPSS’s Data View, comes from a pretend study looking at the effect of dog ownership on the ability to throw a frisbee.įrisbee Throwing Distance in Metres (highlighted) is the dependent variable, and we need to know whether it is normally distributed before deciding which statistical test to use to determine if dog ownership is related to the ability to throw a frisbee.