Short, practical notes on statistical methods and research practice, written for researchers rather than statisticians.
The JH Stat blog is currently under development. Planned articles will cover applied statistics, health research methods, responsible use of AI in health research and reproducible analysis.
No articles have been published yet. The three cards below are planned topics, shown to indicate the kind of writing this section will contain — they are not published articles and cannot yet be read.
If you would like to be told when writing begins, or if there is a topic you would find useful, get in touch.
Writing will be organised under the following headings.
Planned
Each topic below is in preparation. None has been written or published yet, so there is nothing to read at this stage.
Sample size calculations are often treated as a box to tick for a funder. This piece will look at what the assumptions behind the number really commit a study to, and why a power curve is usually more informative than a single figure.
Dropping incomplete records is the most common way missing data are handled, and usually the least defensible. This piece will set out the mechanisms behind missingness and what changes when multiple imputation is used instead.
Prediction models in health research are increasingly built with machine learning methods. This piece will look at where that genuinely improves performance, where a well-specified regression does just as well, and why calibration matters more than discrimination for clinical use.
If there is a statistical problem your team keeps running into, it is probably worth writing about. Suggestions are welcome — and if you need an answer sooner than the blog can provide one, ask directly.