(a) you need to know whether your data are parametric, as this will determine which test you can use

(b) you need to know whether there are outliers that might distort your analyses, and

(c) it gives you some idea of what results to expect from your later analyses, which acts as a checking mechanism

The first thing to do is to have a look at your data on a graph (this is often referred to as eyeballing your data). On SPSS this is done with the

**Graphs**menu, from which you pick the type of graph you want. Make sure you are familiar with producing bar graphs, histograms, and scatterplots.

SPSS also has facilities for exploring your data in other ways, accessed from

**Analyse > Descriptive Statistics**. The

**Descriptives**option is useful for getting summary information about your data. But particularly useful for our purposes is the

**Frequencies**option. This not only displays the same key statistics as the Descriptives command, but also provides graphs showing the distribution of your data - you can even overlay a normal distribution onto it.

Finally, you can also use

**Analyse > Descriptive Statistics > Explore**to objectively test whether your data are normally distributed. The key thing is to click the button marked 'Plots...' and select the option 'normalcy plots with tests'. This does a test called the Kolmogorov-Smirnov test. If this is significant then the distribution is not normal.