Choosing Evidence for a Claim About Saturated Fat
A student writes: "Replacing saturated fat with unsaturated fat lowers heart-disease risk." She gathers three sources. Source A is a randomized controlled trial in 4,000 adults over 5 years comparing a butter-based diet with an olive-oil-based diet, measuring cardiac events. Source B is a blog post by a fitness influencer citing "many studies" but naming none. Source C is an observational study of 12,000 people showing that those who reported eating more olive oil had fewer cardiac events, but the authors note that olive-oil eaters also exercised more and smoked less. The assignment requires at least one source that supports a causal claim.
1. Rank the three sources for supporting a causal claim and justify the ranking.
Read the worked answer
A > C > B.
- A (RCT): Randomization balances confounders between groups, so a difference in cardiac events can be attributed to the diet manipulation. This is the strongest causal evidence.
- C (observational): Shows an association, but the authors themselves flag confounders (exercise, smoking). Correlation here cannot separate olive oil's effect from a healthier overall lifestyle.
- B (blog): No named studies, no methods, no data - not evidence at all, just assertion.
Only A meets the 'supports a causal claim' requirement.
2. The student wants to cite Source C as proof that olive oil causes fewer cardiac events. Rewrite that use of the source so it is defensible.
Read the worked answer
Reframe C as consistent with rather than proof of causation, and name the limitation: "In a 12,000-person observational study, higher olive-oil intake was associated with fewer cardiac events, though the authors noted that olive-oil consumers also exercised more and smoked less, so the association may partly reflect lifestyle rather than olive oil itself." This preserves the finding while acknowledging the confounders the authors identified.
3. If Source A found a 15% relative reduction in cardiac events in the olive-oil group, and the butter group had 200 events out of 2,000, how many events would the olive-oil group have, and what is the absolute risk reduction?
Read the worked answer
Butter-group risk = 200/2,000 = 10%. Relative reduction 15% → olive-oil risk = 10% × 0.85 = 8.5%. Events in olive-oil group = 8.5% × 2,000 = 170. Absolute risk reduction = 10% − 8.5% = 1.5 percentage points (30 fewer events per 2,000 people).
This shows why relative and absolute numbers must be reported together: "15% lower" sounds large, but the absolute drop is 1.5 points.