Bluebook SAT Test Evaluating Statistical Claims: Observational Studies and Experiments 5 — Questions and Answers
Question 1: A study randomly assigns 200 participants to either a sleep restriction group (6 hours/night) or a normal sleep group (8 hours/night) for 2 weeks. The restricted sleep group performs worse on memory tests. What is the most valid conclusion?
- Sleep restriction causes memory impairment in all adults.
- For people similar to those in this study, sleep restriction likely causes worse memory performance. (Correct answer)
- Memory tests are unreliable for measuring sleep effects.
- There is only a correlation between sleep and memory, not a causal relationship.
Correct answer: For people similar to those in this study, sleep restriction likely causes worse memory performance.
Random assignment supports a causal conclusion; however, generalizability is limited to populations similar to the study participants.
Question 2: A researcher observes that counties with higher rates of library visits also have higher graduation rates. She claims that library use causes higher graduation rates. What is wrong with her reasoning?
- Graduation rates are not a valid outcome to measure.
- The association might be explained by a confounding variable, such as overall community investment in education. (Correct answer)
- She needs a larger county-level sample to make any claim.
- Observational studies cannot show any kind of relationship.
Correct answer: The association might be explained by a confounding variable, such as overall community investment in education.
Counties with more libraries likely invest more in education overall, making community investment a confounding variable that could explain both library visits and graduation rates.
Question 3: Participants in a study are randomly selected from 10 different cities and randomly assigned to either an intervention or control group. The intervention group shows a significant positive outcome. To which population can results be most validly generalized?
- All people in the world
- People in cities similar to the 10 cities studied (Correct answer)
- Only the people who were in the intervention group
- All adults in the country where the cities are located
Correct answer: People in cities similar to the 10 cities studied
Random selection from 10 cities supports generalization to populations in similar cities; random assignment supports causal inference but does not expand the scope of generalizability.
Question 4: A study of 10,000 adults finds that vegetarians have lower rates of type 2 diabetes than non-vegetarians. A nutritionist claims this proves vegetarian diets prevent diabetes. What is the most valid critique?
- The study is too large; smaller studies are more valid.
- Vegetarians may differ from non-vegetarians in other healthy behaviors (exercise, lower smoking) that reduce diabetes risk, making causation unclear. (Correct answer)
- Diabetes rates cannot be observed in non-experimental studies.
- The nutritionist's claim is valid because the sample size is large enough to prove causation.
Correct answer: Vegetarians may differ from non-vegetarians in other healthy behaviors (exercise, lower smoking) that reduce diabetes risk, making causation unclear.
Confounding lifestyle factors (e.g., exercise, smoking, stress management) common among vegetarians could explain the lower diabetes rate, undermining a causal claim.
Question 5: A randomized experiment gives one group a mindfulness app and the other no app, then measures anxiety. The app group reports lower anxiety. Which of the following would most limit generalizing this result to all adults?
- The random assignment process introduced bias.
- Participants volunteered for the study, so they may not represent all adults. (Correct answer)
- The experiment lasted too long.
- The anxiety measure was self-reported.
Correct answer: Participants volunteered for the study, so they may not represent all adults.
Volunteers may be more motivated or open to technology than the general adult population, limiting how well the results generalize beyond similar volunteers.
Question 6: An article reports: 'A study found people who take daily vitamins are healthier than those who don't.' The article concludes vitamins cause better health. Which question would most directly test whether this conclusion is valid?
- How many vitamins were taken per day?
- Was this an observational study or a randomized experiment with random assignment? (Correct answer)
- Were the vitamins taken in the morning or evening?
- What brand of vitamins was used in the study?
Correct answer: Was this an observational study or a randomized experiment with random assignment?
The key question is whether the study used random assignment; without it, healthier people may already be more likely to take vitamins, making causation impossible to conclude.
Question 7: A city randomly selects 500 households and surveys them about neighborhood crime and home security systems. They find a negative correlation: more security systems, fewer reported crimes. A council member says security systems reduce crime. What is the main limitation?
- The sample of 500 households is too small for city-wide conclusions.
- Wealthier neighborhoods may have more security systems and also have fewer crimes for unrelated reasons, making the causal claim unjustified. (Correct answer)
- Correlation can never be found in survey data.
- City councils cannot use survey data for policy decisions.
Correct answer: Wealthier neighborhoods may have more security systems and also have fewer crimes for unrelated reasons, making the causal claim unjustified.
Neighborhood wealth is a confounding variable that could explain both higher security system ownership and lower crime rates, making the causal claim unsupported.
Bluebook SAT Test Evaluating Statistical Claims: Observational Studies and Experiments 5
A study randomly assigns 200 participants to either a sleep restriction group (6 hours/night) or a normal sleep group (8 hours/night) for 2 weeks.
The restricted sleep group performs worse on memory tests.
What is the most valid conclusion?