Causal reasoning and alternative explanations
Evaluate whether an observed relationship supports a causal conclusion.
Covered in this guide
- Distinguish correlation, sequence and causation.
- Generate plausible alternative explanations.
- Choose a suitable comparison or test.
Causal reasoning asks whether changing one factor produced a change in another. A relationship alone is insufficient: the proposed cause must precede the outcome, competing causes must be considered and a credible comparison is usually needed.
A correlation is not yet a cause
Causal reasoning asks whether changing one factor produced a change in another. A relationship alone is insufficient: the proposed cause must precede the outcome, competing causes must be considered and a credible comparison is usually needed.
A campus study that ignored self-selection
Employees who use an optional learning portal receive higher performance ratings. Portal use may improve performance, but high-performing employees may also be more likely to use optional training.
Try this
Design a practical check that would improve confidence in either explanation without claiming that one new observation proves the issue completely.
Review answer guidance
Compare similar employees before and after portal use, account for prior performance and role, and examine whether the amount and timing of use predict later improvement. A randomised invitation could strengthen the design if it is ethical and operationally feasible.
Alternative explanations in the task
Name at least one other cause that could produce the same pattern. If the design cannot distinguish them, do not claim a causal effect.