Two Versions of the Same Finding
You are comparing two texts about remote work and promotion. The news article, 900 words, is headlined 'Remote work causes career stagnation'. It says the evidence 'proves' that staff who work from home are less likely to be promoted. The primary study's abstract reports a cross-sectional survey of 2,100 employees in three firms. It found that employees who worked remotely three or more days per week reported lower promotion rates (odds ratio 0.72, 95% confidence interval 0.58–0.89). The authors state that the design cannot establish causation and that 'reverse causality cannot be excluded'. The article quotes a management consultant but not the study authors, omits the confidence interval, and does not mention that all three firms are in one sector. You must decide which claims are supported and prepare a correction note for a student newsletter that plans to repeat the headline.
1. Identify two specific mismatches between the news article and the study abstract. For each, explain why the mismatch changes the meaning.
Read the worked answer
First mismatch: the headline says 'causes', but the study is cross-sectional and the authors explicitly say causation cannot be established and reverse causality cannot be excluded. 'Causes' turns an association into a causal claim.
Second mismatch: the article says the evidence 'proves' the effect, but the abstract reports an odds ratio with a confidence interval, not proof. It also omits that the survey covered three firms in one sector, which limits generalisation.
A third, subtler mismatch is the sourcing: the article quotes a consultant rather than the authors, so the interpretation is not checked against the researchers' caveats. Each mismatch matters because it changes what a reader should do: a proven cause justifies intervention, while an uncertain association justifies further study.
2. Explain, for a student audience, what the odds ratio 0.72 and the 95% confidence interval 0.58–0.89 do and do not tell us.
Read the worked answer
The odds ratio of 0.72 means that, in this sample, employees who worked remotely three or more days per week had lower odds of promotion than the comparison group - about 28% lower odds, not a 28% lower probability. The 95% confidence interval 0.58–0.89 is the range of values compatible with the data; because it does not cross 1.0, the association is statistically significant at the conventional threshold. But it does not tell us the causal direction, the size of the effect for any individual, or whether the result would hold in other firms or sectors. Nor does it rule out confounding: remote workers may differ in role, seniority, or visibility. Students should read it as 'evidence of an association', not 'proof of harm'.
3. Write a correction note of no more than 55 words that the newsletter could publish.
Read the worked answer
Correction note (52 words): 'Our headline overstated the study. The survey found an association between remote work and lower promotion odds (OR 0.72, 95% CI 0.58–0.89), not proof that remote work causes stagnation. The design cannot rule out reverse causality, and the sample covered three firms in one sector. We apologise for the misleading framing.'