Statistics support resources
We have a wide range of statistics resources using SPSS, Jamovi, R and SAS for most commonly used statistical tests but also recommend videos created by the stats support team at Brunel University on Excel and SPSS in the BrunelASK youtube channel. If you are not sure which test to use, try our interactive Choosing the right statistical test
We have a range of Jamovi video and paper resources on Getting started with Jamovi and carrying out
Summary Statistics and Graphs
We have a few resources on summarising data in SPSS but also recommend BrunelASK videos for a wider range in
Testing Averages of Different Groups
These tests are used to compare the averages of completely different groups of people or subjects. Independent t-tests are for comparing two groups and ANOVA for 3+. ANCOVA allows a continuous covariate to be added.
Techniques include: Summarising continuous data, Independent t-tests, one and two way ANOVA and ANCOVA
Testing Averages of Paired or Repeated Measurements
If you have collected multiple measurements of the same variable e.g. weight before or after a diet,
Techniques include: Paired t-test and repeated measures ANOVA and their respective non-parametric techniques the Wilcoxon signed rank and Friedman test which can be used for ordinal or skewed data
Correlation and linear regression
Use these techniques when looking at relationships between continuous variables or the impact of several continuous or binary independent (predictor) variables on one continuous dependent (outcome) variable.
Techniques include: Correlation, scatterplots, simple and multiple linear regression.
Testing Categorical (nominal) Data
If you have two categorical (nominal) variables, look for an association using Chi-squared test of association. If you dependent has only two categories (binary), you can test several binary or continuous predictors using logistic regression.
Techniques include: Chi-squared tests and logistic regression
Non-parametric tests (ordinal or skewed data)
If your dependent variable is ordinal (e.g. strongly disagree - strongly agree) or your data are skewed, use a non-parametric test for comparing groups or repeated measurements.
Techniques include: Independent groups - Mann-Whitney (2 groups) or Kruskall-Wallis (3+). Multiple measurements of the same variable - Wilcoxon signed rank (2) and Friedman (3+)
If you are using a mean or total score of a set of ordinal questions, use Cronbach's alpha to test for internal consistency. If you are testing for agreement between observers or methods, use Intraclass correlation for continuous and Kappa for categorical.
Techniques include:e.g. Cronbachs Alpha, Intraclass correlation, Kappa
Choosing the right statistical test
We have summary self-help sheets on statsitical terminology for choosing the right test, summary tables and our interactive test chooser which takes you to the appropriate resource at the end.
Data manipulation in SPSS
We have a few resources on manipulating data in SPSS here but also recommend videos created by the stats support team at Brunel University on topics such as reverse coding, creating or reducing categories and creating a scale mean.
These data sets are used in the self-help resources in case you want to follow through the process.