Sample Size and Representativeness: A study's findings are more reliable if the sample size is large and the demographics (age, sex, ethnicity) match the target population. Small or non-representative samples may lead to biased results.
Statistical Significance: Researchers use statistical tests to determine if the differences observed in data are likely due to the risk factor or just random chance. The presence of error bars or standard deviations helps in comparing mean values.
Control of Variables: Valid studies must account for confounding variables—other factors that might influence the outcome. For example, a study on smoking should also consider the participants' diet and exercise levels.
| Concept | Correlation | Causation |
|---|---|---|
| Definition | A statistical link or association between two variables. | One variable directly causes the change in another. |
| Evidence | Observed patterns in data (e.g., as A increases, B increases). | Requires experimental proof and a biological mechanism. |
| Conclusion | "There is a link between smoking and CVD." | "Smoking causes CVD." (Requires much stronger evidence) |
Describe vs. Conclude: When asked to 'describe' data, state exactly what the numbers show (e.g., 'As X increases, Y increases'). When asked to 'conclude', interpret the meaning (e.g., 'There is a positive correlation').
Check the Baseline: Always identify the control group or the baseline relative risk (usually ). This allows you to quantify the increase in risk for other groups.
Critique the Methodology: If asked to evaluate a study, look for weaknesses such as small sample sizes, lack of statistical testing, or failure to control for other risk factors.
Avoid Absolute Language: Unless a causal link is explicitly proven and stated, use terms like 'association', 'link', or 'correlation' rather than 'causes'.
