Teaching resources

Analysing quantitative data: An introduction for language researchers

These lecture notes as well as the datasets used are available from GitHub.

Contents:

1. Software
2. Working with datasets
3. Fundamentals of probability theory
4. Descriptive statistics of a univariate sample
5. Random samples
6. Estimating estimation uncertainty
7. Another look at the mean
8. Adding a predictor
9. Group differences
10. Differences between differences
11. Several predictors in one model
12. The logic of significance testing
13. The t-test
14. Analysis of variance
15. Calculating statistical power
16. Silly tests
17. Questionable research practices
18. Within-subjects experiments
19. Logistic regression
20. Recommendations for self-study

Quantitative methodology: An introduction

The script for the course on quantitative methodology I teach is available here.

Contents:

1. Association and causality.
2. Constructing a control group.
3. Alternative explanations.
4. Inferential statistics 101. (The course is not a statistics course, but there’s no avoiding talking about p-values given their omnipresence.)
5. Increasing precision.
6. Pedagogical interventions.
7. Within-subjects experiments.
8. Quasi-experiments and correlational studies.
9. Constructs and indicators.
10. Questionable research practices.

I’ve also included two appendices:

• Reading difficult results sections.
• Reporting research transparently.

Introduction to the general linear model

These are the lecture notes for a summer school module I taught. They are available from GitHub.

Contents:

1. Nuts and bolts: General linear model equation; optimisation criteria (least absolute deviations, least squares, maximum likelihood); estimating uncertainty (bootstrapping, i.i.d. normality assumption).
2. Adding a predictor: Interpretation of parameter estimates and regression lines; confidence bands.
3. Group differences: Dummy variables (treatment coding, sum coding); bootstrapping without homoskedasticity.
4. Interactions.
5. Multiple predictors: Confounding variables; control variables; posttreatment variables.
6. The basic of logistic regression: Linear probability model; odds, odds ratios, log-odds.

Working with datasets and visualising data in R: Two primers

I wrote a primer on working with datasets in R and another on visualising data in R. You can find the datasets used in the primers on GitHub.

Visualising statistical uncertainty using model-based graphs

I wrote a tutorial about visualising the statistical uncertainty in statistical models for the BICLCE 2019 conference in Bamberg. You can find the tutorial here: Visualising statistical uncertainty using model-based graphs.

Contents:

1. Why plot models, and why visualise uncertainty?
2. The principle: An example with simple linear regression
   • Step 1: Fit the model
   • Step 2: Compute the conditional means and confidence intervals
   • Step 3: Plot!
3. Predictions about individual cases vs. conditional means
4. More examples
   • Several continuous predictors
   • Dealing with categorical predictors
   • t-tests are models, too
   • Dealing with interactions
   • Ordinary logistic regression
   • Mixed-effects models
   • Logistic mixed effects models
5. Caveats
   • Other things may not be equal
   • Your model may be misspecified
   • Other models may yield different pictures

Miscellaneous tutorials

The blog archive contains a number of tutorials.

cannonball (R package)

The cannonball package bundles a couple of functions that I use when teaching introductory courses in quantitative methodology and statistics. These include

• plot_r() for drawing different scatterplots with the same correlation coefficient,
• walkthrough_p() and walkthrough_blocking(), which both aim to help students see the connection between an experiment’s design and its analysis,
• clustered_data() for simulating data from cluster-randomised experiments,
• parade() and associated functions for helping researchers check the assumptions of their statistical models.

More information is available on GitHub.