1/n Research thought for the day: People looking for small effects in small samples are chasing phantasms.

Partly because there is *no* statistical method that can reliably detect small intervention effects in small samples, it is generally unethical to conduct such studies.

2/n Even if the intervention is likely to be harmless, research takes time and costs money. Putting people through research studies that are unlikely to yield enough information to guide decisions about whether to use the intervention is seldom justifiable.

3/n Some reasons such studies *might* be justified include piloting the feasibility of a research design in a particular setting, and/or helping with the development of budding scientists embarking on their first research projects.

As a person who has been involved in studying aspects of injury prevention and performance for the last quarter-century it pains me to say this, but there is a *lot* of dross published in the field of sports medicine and sports science.

If an intervention study is worth doing, it should be done well. This means, among other things, consulting experts in study design and analysis before you begin, and ensuring you have reasonable grounds to decide whether or not to use the intervention at the end!

One of the implications of viewing injury aetiology from a “complex systems” perspective is the realisation that what you are seeing locally may bear the same relationship to the general pattern as an eddy in a river does to the overall flow.

It is a real (local) phenomenon, but provides little basis for inferring how things are more generally.