Scientific research processes
Scientific research processes
This domain tests evidence-based practice: formulating answerable questions, appraising studies, and applying findings without overreach.
Core rules
- Use a focused question (population, intervention, comparator, outcome) to search efficiently.
- Rank evidence by design and risk of bias; randomized controlled trials and systematic reviews generally outrank uncontrolled case reports for causal claims.
- Read effect sizes and confidence intervals alongside p-values; statistical significance is not practical importance.
- Check sample representativeness, blinding, randomization, and outcome validity before generalizing to your athletes.
- Document search decisions and monitoring protocols so another practitioner can audit them.
Worked example
A supplement study reports a significant 2% sprint improvement (p = 0.04) in 12 recreationally trained men. Reasoning: the effect is small, the sample is not elite, and the confidence interval may include trivial effects. Before adopting it, check whether the outcome matches your sport, whether blinding held, and whether the effect exceeds typical timing error. The defensible conclusion is "promising but not ready for team-wide use."
Traps
- Treating correlation as causation.
- Overvaluing p-values while ignoring effect size and precision.
- Generalizing from untrained or single-team samples.
- Confusing a surrogate measure with competitive performance.
- Failing to check whether the control condition was credible.