Choosing and defining variables
- Concept
Saying exactly what is measured, with which instrument and cut-off, so that anyone would get the same result.
What it is
The variables in a study play different roles: the dependent variable (outcome), the independent variable (exposure or intervention) and control variables (possible confounders and other covariates). An operational definition turns a concept (‘hypertension’, ‘physical inactivity’) into a measurable procedure: instrument, technique, timing, cut-off and scale. Without it, two studies using the same word measure different things.
Key facts
- Conceptual definition: what the variable means. Operational definition: how it is measured in this study.
- Example: physical inactivity = less than 150 minutes a week of moderate activity, by a validated questionnaire on the last 7 days.
Worked example
- Data: a cross-sectional study of 250 adults; variable ‘hypertension’. An operational definition is needed before measuring.
- Definition A: mean of 2 readings ≥ 140/90 mmHg or use of antihypertensives. 62 meet it: 62 / 250 = 24.8 %.
- Definition B: a 130/80 mmHg threshold or treatment. 104 meet it: 104 / 250 = 41.6 %.
- With the same data, prevalence changes by 41.6 − 24.8 = 16.8 points because of the definition alone.
- Scales: systolic pressure is continuous quantitative; ‘hypertension yes/no’ is nominal binary; hypertension grades are ordinal.
- Reading: the protocol fixes the definition, a validated device, the technique and the number of readings; otherwise two studies' prevalences cannot be compared.
The role of variables in a study
| Role | What it is | Example |
|---|---|---|
| Dependent (outcome) | What is to be explained or predicted | Blood pressure at 6 months |
| Independent (exposure) | The possible cause or the intervention | Salt intake; an education programme |
| Control (confounder) | Linked to exposure and outcome, without being intermediate | Age, body mass index |
| Intermediate | Lies on the causal path between exposure and outcome | Weight gained because of the diet |
Markers
- DV
- IV
- Confounder
In practice
- When reading a paper, look for the operational definition of the outcome: different definitions explain many figures that do not match.
- Surveillance case definitions are operational definitions shared across the whole country.
Common pitfalls
- Using a concept without defining how it is measured.
- Adjusting for an intermediate variable: it removes part of the real effect.
- Categorising a continuous variable with cut-offs chosen after seeing the data.
Test yourself
Think of the answer, then tap the question to see it.
What does an operational definition include?
The instrument, technique, timing, number of measurements, cut-off and scale of the variable.
In a study of salt and blood pressure, what role does age play?
Control variable: a possible confounder, linked both to salt intake and to blood pressure.
Of 400 people, 80 meet one definition and 140 a broader one. How much does prevalence change?
From 20 % (80/400) to 35 % (140/400): 15 points.
Why is an intermediate variable not adjusted for?
Because it lies on the causal path: adjusting for it removes part of the effect being measured.
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Principles of Epidemiology | Lesson 2 - Section 2 (Centers for Disease Control and Prevention, 2012); Collecting Data (Centers for Disease Control and Prevention (The CDC Field Epidemiology Manual), 2024).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.