7. Kruskal-Wallis test and Friedman’s ANOVA

Amy Atkinson

Lab preparation

Before the lab, please watch the following short video. This walks you through how to perform a Kruskal-Wallis test and Friedman’s ANOVA in R.

If you want to have a play around with the script yourself, the R markdown script and datasets can be downloaded here.

Lab

Overview

I’ll provide you with one or two research questions each week which will require you to complete the statistical tests covered in the lectures. You can work in groups or individually.

You can write your script as a .R or Rmd file. Use the lab preparation video and script, lecture slides, and previous content covered in the statistics modules to help you.

The presentation given at the start of the lab can be downloaded here.

Datasets

The datasets for this lab can be downloaded here.

Research Question 1

You are a psychology lecturer. You hear that the library is offering three statistics courses. You are interested in whether students who attend the courses perform significantly differently from each other.

You recruit 18 people and assign each one to a course. After the courses are finished, you ask them to write an R script. You time how long it takes students to complete the task. You are interested in whether there is a significant effect of course on the time taken to complete the task.

Research Question 2

You are a developmental psychologist. You are interested in whether working memory develops between 15 and 17 years of age.

You recruit a sample of adolescents and test them on a working memory task when they are 15 years of age, 16 years of age, and 17 years of age.

You then examine whether there is a significant effect of age on working memory score.

Hints and tips

Your script should aim to answer and interpret both of these research questions. Start a new session on the server, then load in the required libraries (tidyverse, cowplot, ggpubr, rstatix) and the datasets.

For each research questions, you will need to:

  1. Perform normality checks

  2. Explore your data (e.g. descriptive statistics, a plot)

  3. Conduct the statistical test

  4. Calculate an effect size

  5. Conduct post-hoc tests (where appropriate)

  6. Interpret the output

Model script

A model script showing one way of answering the research questions above using R will be available here from 9am on Monday of Week 18.

Feedback on student scripts

Here is some feedback on a few scripts submitted by students.

Back to top