The dependent variable (outcome) is what you want to explain or predict. Independent variables (predictors, exposures) are what you think influences it.
Why it matters
This choice decides which test or model you use, and which way you write your conclusion.
Key ideas
- Outcome = Y, the result you measure
- Predictor = X, the possible influence
- One study can have many predictors
- The roles depend on your question, not on the data itself
- Don't confuse with confounders, which affect both
Real-life research example
Question: "Does daily steps count affect fasting glucose?" Steps = independent variable, glucose = dependent variable.
When to use
- Decide the outcome before you open the software
When NOT to use
- Don't swap roles just because it gives a nicer p-value
Quick takeaway
Outcome is what you explain; predictors are what you use to explain it.
Try it yourself
For the question "Do students who sleep more score higher in exams?", name the outcome and the predictor. Add one other variable that could confuse the answer.
Model answer
Outcome: exam score. Predictor: hours of sleep. A possible confounder: study time, because it may relate to both sleep and score.