Skip to Main Content

MASH : Maths and Stats Help

Statistical Tests

We have provided some guides to help you with carrying out a number of commonly used statistical tests.
This list of statistical tests is not exhaustive.

We show you the process to carry out the test in SPSS, examples of the output that it will produce, and an example of how you could report it using the correct APA style.

If there is a test you would like to see here, please email us your suggestion.

Testing for a Difference

We can explore differences between groups in multiple ways, depending on the structure of our groups. If we divide our data between subjects (putting different people/participants in different groups) we can compare between these groups and we have a between-subjects design. If we have one participant group and we measure them multiple times, we can test if there is a difference in the repeated measurement and therefore, we have a repeated-measures design.

In some subjects, like Psychology, a repeated-measures design is called a within-subjects design.

Parametric Tests

Non Parametric Tests

Testing for a Relationship

We can explore relationships in two ways; a correlation or a regression. Correlations quantify or measure the strength of the relationship between two variables. Regression expresses the relationship between two (or more) variables in the form of an equation.

Button Table

Parametric or Non-Parametric

Data may be described as as parametric, for data to be parametric it must possess four characterises sometimes called the Parametric Assumptions. Data is parametric if it is normally distributed, homogenous, interval level (or higher) and independent.

Parametric tests, such as an Independent Samples t-Test, should only be used on parametric data, if your data is non-parametric you should use a non-parametric alternative, such as the Mann-Whitney U-Test

For more information on determining if your data is parametric or not, please our guide:

You can also directly test if the data is normally distributed by conducting a normality test, such as a Shapiro-Wilk test. Other normality tests can be used, choices of which are often determined by research field and sample size.

Testing for Reliability

Sometimes we use a series of questions and attempt to combine them into a single scale. These scales are great, as they allow us to perform higher levels of analysis such as ANOVAs and Regressions but we need to ensure they are reliable. We can test the overall reliability of a scale by calculating the intercorrelation of the items.