Research, Evidence, and Data Literacy in Nutrition
Core concepts
Understand study designs (randomized controlled trials, cohort, case-control, cross-sectional, ecological), their strengths and biases, and the difference between association and causation. Know key terms: confounding, selection bias, measurement error, generalizability, statistical versus clinical significance, relative versus absolute risk, sensitivity/specificity, and p-values and confidence intervals. Understand levels of evidence and systematic reviews/meta-analyses. Recognize that nutrition epidemiology is largely observational and prone to confounding by lifestyle and diet measurement error.
Traps in questions
- Treating a cross-sectional association as causal.
- Confusing relative risk with absolute risk (a large relative change can be a small absolute change).
- Assuming statistical significance equals practical importance.
- Overlooking that self-reported dietary data are subject to recall and social-desirability bias.
Rule of thumb
Before accepting a nutrition claim, ask: what design, what population, what comparison, and how large is the absolute effect?
Worked example
A cohort reports a 50% higher relative risk of a disease with a food, but the baseline risk is 2 per 10,000: the absolute increase is 1 per 10,000-clinically small despite the dramatic relative figure.