Chapter 37 — Quiz

Eighteen questions: thirteen multiple choice, five short answer. Answer key with explanations follows. The quiz emphasizes recognition of methodological strength, not memorization of nutrient claims.

Multiple choice

1. Which of the following is the strongest type of evidence in nutrition research?

a) A single observational study with 30,000 subjects, published in a major journal. b) A randomized controlled trial with 7,500 subjects running over five years. c) A meta-analysis pooling fifty observational studies of diverse design. d) A mechanistic study in cell culture showing how a nutrient affects cancer cells.

2. A food-frequency questionnaire (FFQ) typically yields data with what kind of accuracy?

a) Highly accurate; reliable for individual prediction of nutrient intake. b) Accurate within about 5%, so useful for both group and individual claims. c) Roughly correlated with actual intake (correlations in the 0.3–0.6 range), useful for population-level patterns and weak for individual prediction. d) Almost entirely fabricated; should not be used in research.

3. The Bradford Hill criteria for inferring causation from correlation include all of the following EXCEPT:

a) Strength of the effect. b) Consistency across populations. c) Statistical significance below p = 0.05. d) Biological plausibility.

4. The 2019 Kevin Hall NIH study on ultra-processed foods found that:

a) Subjects ate about the same on both diets but lost more weight on ultra-processed. b) Subjects ate about 500 calories per day more on the ultra-processed diet, despite calorie-matched food on offer, and gained weight. c) Subjects became hyperglycemic on ultra-processed diets within 48 hours. d) Ultra-processed foods produced no measurable effect compared to whole foods.

5. Which of the following is the BEST-evidenced dietary recommendation in this chapter?

a) Eat low-fat foods. b) Eat low-carb foods. c) Get 25–35 g of fiber per day. d) Take a daily multivitamin.

6. The CARET trial of beta-carotene supplementation in smokers and asbestos workers was halted early because:

a) The supplements were contaminated with mercury. b) The supplemented group showed increased lung cancer mortality. c) The placebo group showed faster cancer progression. d) The funding ran out.

7. The 2007 Bjelakovic et al. JAMA meta-analysis of antioxidant supplements found that:

a) All antioxidants reduced all-cause mortality in supplemented populations. b) Beta-carotene, vitamin A, and vitamin E supplementation increased all-cause mortality. c) Vitamin C had a strong protective effect on cardiovascular disease. d) Selenium reduced prostate cancer rates by 40%.

8. The "complete vs incomplete protein" worry of mid-20th-century nutrition was:

a) Confirmed: vegetarians must combine grains and legumes at every meal. b) Largely overturned: amino acid pools allow protein to be balanced across the day, not within a single meal. c) Confirmed for adults but not children. d) Never seriously held by nutrition scientists.

9. The Mediterranean dietary pattern, as evidenced by the PREDIMED trial, showed:

a) About 30% reduction in major cardiovascular events. b) About 70% reduction in cancer. c) No measurable effect compared to the low-fat control. d) A small reduction in obesity but no change in cardiovascular events.

10. Industrial trans fats from partial hydrogenation:

a) Are still considered safe by most health agencies. b) Have been classified as no longer "generally recognized as safe" by the FDA, with required removal from the U.S. food supply by 2018. c) Are unrelated to cardiovascular risk. d) Are now considered beneficial in moderation.

11. Which of these is an example of a surrogate endpoint in a nutrition study?

a) Death from any cause. b) Diagnosis of type 2 diabetes. c) Fasting blood glucose level. d) Heart attack within five years.

12. A reasonable position on saturated fat, given current evidence, is:

a) All saturated fat is poisonous and should be eliminated. b) Saturated fat has no effect on cardiovascular risk. c) Replacing saturated fat with unsaturated fat (olive oil, nuts) reduces cardiovascular risk; replacing it with refined carbohydrates does not. d) Saturated fat is uniformly beneficial and should be the primary calorie source.

13. The Hall et al. 2019 ultra-processed-food study is methodologically strong because it was:

a) A large observational cohort using FFQs. b) A randomized cross-over metabolic-ward study with subjects acting as their own controls under matched-nutrient conditions. c) A meta-analysis of all prior FFQ-based studies. d) A mechanistic cell-culture experiment.

Short answer

14. A friend reads the headline "Eggs Linked to Higher Heart Disease Risk: New Study" and panics. The study was a single observational cohort with 16,000 subjects in one country, with eggs measured by FFQ. The reported relative risk was 1.08 for one egg per day. Walk your friend through how worried (or not) they should be, in 4–6 sentences.

15. Why is it that "calories in, calories out" is true at the level of thermodynamics but incomplete as a guide to weight management? Mention at least two specific mechanisms.

16. Explain in 4–6 sentences why population-level nutrition trials longer than a few months are notoriously hard to interpret, even when they are well-designed at the start.

17. The chapter argues that the framework of "is this a Mediterranean diet?" is the wrong question for someone evaluating a traditional cuisine that has not been studied as much. What is the right question, and why does the question matter?

18. Maya's neighbor sees an ad for a "probiotic" supplement promising "10 billion CFU of immune-boosting bacteria" and asks Maya whether she should buy it. Drawing on the chapter, write Maya's reply in 4–6 sentences.

Answer key

1. b) A randomized controlled trial is the strongest single study type because randomization eliminates most forms of confounding. A meta-analysis of RCTs would be even stronger; a meta-analysis of observational studies (option c) is weaker because it pools designs that share unmeasured confounders. The mechanistic study (d) is suggestive but cannot establish human-population effects.

2. c) FFQ data correlate roughly with actual dietary intake at coefficients of 0.3 to 0.6 — useful for detecting population-level patterns but not for individual prediction. This is the central methodological caution of much of the nutrition-epidemiology literature.

3. c) The Bradford Hill criteria are strength, consistency, specificity, temporality, biological gradient, plausibility, coherence, experiment, and analogy. Statistical significance is not on the list. Bradford Hill was explicit that the criteria were considerations of judgment, not a checklist of statistical rules.

4. b) The Hall study showed that subjects spontaneously ate about 500 cal/day more on the ultra-processed diet than on the unprocessed diet, despite calorie-matched food being offered, and gained weight. The strength of the study was that the diets were nutritionally matched on the variables most diet-comparison studies vary; only the processing differed. This is the strongest single piece of evidence for the ultra-processed-food hypothesis.

5. c) Fiber intake of 25–35 g/day has perhaps the strongest cardiovascular and metabolic evidence base of any single dietary recommendation in this chapter. Most adults in industrialized countries eat about half this amount. The other options range from contested (a, b) to mostly unsupported by RCT evidence (d).

6. b) CARET, published in 1996, was halted early because the beta-carotene plus retinol group showed increased lung cancer mortality compared to placebo — the opposite of what the antioxidant hypothesis predicted. Together with ATBC and SELECT, this trial collapsed the prophylactic-mega-dose-supplement hypothesis.

7. b) The Bjelakovic meta-analysis of 47 RCTs found that beta-carotene, vitamin A, and vitamin E supplementation significantly increased all-cause mortality. The increase was small but real. This is one of the most consequential, and least-publicized, findings in modern nutrition.

8. b) The mid-20th-century worry that vegetarians needed to combine grains and legumes at every meal was based on a misunderstanding of amino acid pool dynamics. Subsequent work showed that the body maintains amino acid pools that allow balance across the day. The 1970s-era "must combine at one meal" rule has been revised.

9. a) PREDIMED showed about a 30% reduction in major cardiovascular events in the Mediterranean groups (with extra-virgin olive oil or with mixed nuts) compared to a low-fat control. After methodological re-analysis in 2018, the cardiovascular benefit was confirmed though somewhat attenuated.

10. b) Industrial trans fats were classified by the FDA as no longer "generally recognized as safe" in 2015, with required removal from the U.S. food supply by 2018. Many other countries have similar bans. The trans-fat story is one of the cases where nutrition science arrived at a confident, replicated answer that translated into successful policy.

11. c) A surrogate endpoint is a measurement that is associated with disease but is not the disease itself. Fasting blood glucose is a surrogate; type 2 diabetes is the actual disease. Death and heart attack are hard endpoints. Surrogate endpoints are often what nutrition trials measure, because they can be detected over shorter time horizons; the inferential leap from "surrogate improved" to "disease reduced" is where many claims fall apart.

12. c) The current reasonable position is that replacing saturated fat with unsaturated fat (olive oil, canola, nuts, seeds) reduces cardiovascular risk; replacing saturated fat with refined carbohydrates does not. The replacement question matters more than the absolute saturated-fat content. The Women's Health Initiative trial helped establish this.

13. b) The Hall study was a randomized cross-over metabolic-ward design — subjects acted as their own controls, the food was carefully matched on nutrients, the conditions were tightly controlled. This is methodologically much stronger than the observational FFQ literature on ultra-processed foods, which is why the result lands with more weight.

14. Sample answer. Your friend should be only mildly worried, and should probably keep eating eggs if they enjoy them. The study is observational (not randomized), used FFQ data (which is noisy at the individual level), and reported a relative risk of 1.08 — a small effect that is well within the range that confounding can produce. The body of evidence on eggs has shifted multiple times in the last forty years, and a single observational study with a small effect should not change anyone's diet. Wait for replication and meta-analyses before re-thinking.

15. Sample answer. "Calories out" is not a constant — total daily energy expenditure varies with body composition, activity, ambient temperature, hormones, and adaptive thermogenesis. After weight loss, resting metabolic rate often drops more than predicted, defending the previous weight. "Calories in" is also regulated — hunger and satiety are mediated by hormones (leptin, ghrelin, GLP-1), gut signals, and central nervous system processing, which respond differently to different foods. A bowl of beans and a candy bar at equal calories produce very different satiety responses, which means real-world calorie intake is a function of food choices, not just willpower.

16. Sample answer. Long nutrition trials face the adherence problem: by year three of a five-year low-fat trial, the "low-fat group" and the "control group" are usually eating diets that are nearly indistinguishable. People go on vacation, get divorced, change jobs, change preferences. The intervention you designed at week one is not the intervention being tested at year five. This means even well-randomized trials often produce muted or null results not because the intervention doesn't work but because the intervention isn't actually being tested. The Women's Health Initiative trial is the canonical example.

17. Sample answer. The right question is is this a varied, mostly-whole-food diet that has sustained healthy people over generations? The Mediterranean pattern is well-studied because Mediterranean populations were studied by predominantly American and European nutrition researchers; the absence of equivalent evidence for Oaxacan, West African, Korean, or Vietnamese diets is a function of who has been studied, not of which diets are nutritionally sound. The framework "is this Mediterranean?" treats one well-studied tradition as the template, when in fact many traditional cuisines have plausibly equivalent properties that haven't been measured. Asking the right question is part of making nutrition science genuinely global.

18. Sample answer. The CFU number on the label is largely unregulated, and there's no guarantee the bacteria are alive when you take the capsule or that they survive your stomach acid. "Boosting your immune system" is not a meaningful clinical claim. The evidence for over-the-counter probiotics doing anything in healthy people, outside narrow indications like C. difficile recurrence, is weak. If you want the modest microbiome-diversity benefit shown in the Stanford 2021 trial, eat actual fermented foods — yogurt, kefir, sauerkraut, kimchi, miso, kombucha — which are cheaper, taste better, and have at least equally good evidence.