Statistics with R
The purpose of this page is to provide an educational resource for researchers to support them in conducting basic quantitative analyses using R. It is also a resource to support students taking my quantitative research methods courses and as such there will be regular updates in line with these courses.
The reason R has been selected is because it is a free and open source tool (https://rstudio.com/products/rstudio/download/), and there is a large community of R users who are regularly developing new packages and creating additional educational resources.
The following subpages are developed by knitting RMarkdown documents as .html files and embedding them in this site.
Importantly, they have all been developed as educational material, so are not all comprehensive tutorials and the full utility of packages of functions.
Computing descriptive statistics
Visualisations of descriptive statistics
Testing the assumption of normality
Testing the assumption of equality of variances
Adjusting for multiple comparisons and Type I errors
Comparing mean differences between two groups
Dependent and independent samples t-tests and non-parametric alternatives with effect sizes
Identifying and treating univariate outliers
Comparing mean differences between three or more groups
One-way ANOVA, Welch's one-way test and the Kruskal-Wallis rank sum test with effect sizes
Identifying and treating multivariate outliers
Multiple regression and the consolidation of the above tests as linear models
Controlling for variables
ANCOVA as a linear model
Developing a visualisation
The basic code for creating a visualisation using ggplot, including adding multiple geoms, adding colour and fill, facets, and combining plots
Making adjustments to visualisations
Changing themes, adding labels (titles, etc.), selecting colour palettes, transparency, choosing different plot types
Building information rich plots
Adding appropriate geoms, including confidence intervals, including formulae, highlighting information and adding labels
Cohen's conventional effect sizes
Power analyses for tests of two means, correlations, balanced one-way ANOVA models, and general linear models.