Evaluating a Claim About Vitamin C and Colds
A magazine article claims: 'Taking 1000 mg of vitamin C daily cuts your chance of catching a cold by half.' The article cites a study in which 200 volunteers were given either 1000 mg vitamin C or a placebo for 12 weeks. Participants recorded whether they developed cold symptoms. Results: 28 of 100 in the vitamin C group reported a cold; 40 of 100 in the placebo group reported a cold. The article does not mention that participants knew which treatment they received, that the study was funded by a supplement company, or that the difference was not statistically significant (p = 0.08). A second, larger study found no significant difference.
1. Calculate the reported reduction and evaluate whether the headline 'cuts your chance by half' is justified.
Read the worked answer
Cold incidence: vitamin C group 28/100 = 28%; placebo group 40/100 = 40%. Absolute reduction = 40% − 28% = 12 percentage points. Relative reduction = 12/40 = 30%, not 50%. The headline overstates the effect even on the study's own raw figures.
Further, p = 0.08 means the difference is not statistically significant at the conventional 0.05 threshold; a difference this large could plausibly arise by chance in samples of 100. So the claim of any real effect is not established by this study. The 'half' figure appears to be an unsupported exaggeration, and the article should have reported the absolute and relative reductions and the uncertainty.
A defensible statement would be: 'In this study, 28% of the vitamin C group and 40% of the placebo group reported colds; the difference was not statistically significant and may be due to chance.'
2. Identify features of the study design that limit the conclusions that can be drawn, and explain how each could bias the result.
Read the worked answer
Lack of blinding: participants knew their treatment. People who believe they are taking a protective supplement may be more likely to notice and report cold symptoms, or conversely may under-report; expectation effects can inflate or deflate the difference. A double-blind design (neither participants nor assessors know allocation) removes this.
Self-reported outcome: 'reported a cold' is subjective; there was no objective confirmation (e.g. viral testing or clinical assessment). This increases measurement error and susceptibility to bias.
Funding source: supplement-company funding does not automatically invalidate the study, but it creates a potential conflict of interest and should be disclosed so readers can judge.
Sample size and power: 200 participants gives limited power to detect a modest effect, which may explain the non-significant p value.
Single study: one study is weak evidence; the larger second study finding no difference suggests the first result may be a chance finding.
3. The article's author argues that because the second study was larger, it must be more reliable. Assess this reasoning.
Read the worked answer
Larger sample size generally increases precision and power, reducing the chance of a false negative and narrowing confidence intervals, so the second study is likely to provide a more precise estimate. However, size alone does not guarantee validity. A large study can still be biased if it is unblinded, poorly randomised, or uses subjective outcomes; a small well-designed study may be more trustworthy than a large flawed one.
The right approach is to assess both studies on design quality (randomisation, blinding, objective outcomes, pre-registration), consistency with other evidence, and effect size with confidence intervals. If the larger study was well designed and found no significant difference, that is stronger evidence against a large effect. The most defensible conclusion is that vitamin C at 1000 mg daily probably does not halve cold risk, and any small effect remains uncertain.
This illustrates that scientific claims should be judged on the weight of evidence, not on a single headline or a single study's size.