Blog categories
Subscribe to new blog posts.Analysis
- Draft: Replication success as predictive utility
- Interactions between continuous variables
- Confidence intervals for standardised mean differences
- Which predictor is most important? Predictive utility vs. construct importance
- R tip: Ordering factor levels more easily
- Classifying second-language learners as native- or non-nativelike: Don't neglect classification error rates
- Why reported R² values are often too high
- On correcting for multiple comparisons: Five scenarios
- The problem with cutting up continuous variables and what to do when things aren't linear
- Covariate adjustment in logistic regression — and some counterintuitive findings
- Some tips on preparing your data for analysis
- Silly significance tests: Tests unrelated to the genuine research questions
- Assessing differences of significance
- Calibrating p-values in 'flexible' piecewise regression models
- Analysing pretest/posttest data
Teaching
- Automatise repetitive tasks
- Some illustrations of bootstrapping
- What data patterns can lie behind a correlation coefficient?
- Surviving the ANOVA onslaught
- Explaining key concepts using permutation tests
- A purely graphical explanation of p-values
Reporting
- Tutorial: Adding confidence bands to effect displays
- Tutorial: Plotting regression models
- Common-language effect sizes
- Tutorial: Drawing a dot plot
- Tutorial: Drawing a boxplot
- Tutorial: Drawing a line chart
- Tutorial: Drawing a scatterplot
- Silly significance tests: The main effects no one is interested in
- Some advantages of sharing your data and code
- Drawing a scatterplot with a non-linear trend line
- A more selective approach to reporting statistics
- Thinking about graphs
- Why I don't like standardised effect sizes
- Overaccuracy and false precision
- Some alternatives to bar plots
- Silly significance tests: Tautological tests
- Silly significance tests: Balance tests
Design
- Confidence interval-based optional stopping
- Creating comparable sets of stimuli
- Abandoning standardised effect sizes and opening up other roads to power
- The Centre for Open Science's Preregistration Challenge: Why it's relevant and some recommended background reading
- Experiments with intact groups: spurious significance with improperly weighted t-tests
- Causes and consequences of unequal sample sizes
- Analysing experiments with intact groups: the problem and an easy solution
- Covariate adjustment in logistic mixed models: Is it worth the effort?
- Controlling for confounding variables in correlational research: Four caveats
- Power simulations for comparing independent correlations
- More on why I don't like standardised effect sizes
- The curious neglect of covariates in discussions of statistical power