Evaluating a conclusion asks how strongly the evidence supports it. Consider the sample size, the range of conditions tested, the uncertainty, the number of anomalies, and whether other variables might explain the results. A conclusion is only as strong as its weakest link.
Good evaluations also consider who did the research and whether they had an interest in a particular result, which could cause bias. Results that are reproduced by independent groups using different methods are much more convincing. Suggest specific further work that would test the conclusion more severely.
Worked example
A company's own small study says its sports drink improves running times. How should this be evaluated?
- The sample is small, so chance could explain the result.
- The company may have a bias towards a positive result.
Answer:An independent, larger, controlled study is needed.
Key words
- bias
- a tendency to favour a particular result
- independent
- carried out by people with no interest in the result
- reproduced
- repeated by others with similar results
- further work
- new tests to check a conclusion
New questions every time.