The Sleep-Study Summary and the Overreaching Conclusion
A student summarizes a study for a science class.
Researchers tracked 240 high-school students for six weeks. Students who reported sleeping fewer than six hours on school nights scored lower on a weekly attention test than students who reported seven or more hours. The gap held after the researchers accounted for caffeine intake and screen time. The study shows that insufficient sleep causes attention problems in teenagers, and that later school start times would fix them.
The teacher marks two phrases: shows that... causes and would fix them. The study was observational: students reported their own sleep hours, and attention was measured once per week. No students were assigned to sleep conditions.
1. Explain why "shows that insufficient sleep causes attention problems" overstates what an observational study can support, and rewrite the claim to match the evidence.
Read the worked answer
The study is observational: students reported their own sleep, and no one was assigned to sleep conditions. Observational data can show an association - short sleepers scored lower - but cannot establish causation, because a third factor could drive both. The researchers controlled for caffeine and screen time, which narrows the alternatives but does not eliminate them: for example, students under heavy academic or family stress might both sleep less and score lower, and stress was not measured.
A claim matched to the evidence:
The study finds an association between short sleep and lower attention scores among teenagers, one that persisted after accounting for caffeine and screen time.
This keeps the finding and the control while dropping causes. Note that the control for two variables strengthens the association; it does not license a causal claim.
2. Evaluate the claim that "later school start times would fix them." Identify two separate problems with it.
Read the worked answer
Two distinct problems:
It is not tested by this study at all. The study measured sleep hours and attention; it never manipulated start times. Even if short sleep caused low attention, the study provides no evidence that changing school schedules would change sleep hours - students could go to bed later instead.
"Would fix them" is a certainty claim about an untested intervention. Fix implies the problem would be eliminated, which no single observational study could support even for a tested intervention.
A defensible version would be a hypothesis, clearly labeled: One possibility worth testing is that later start times could increase sleep and, in turn, support attention - but this study does not examine that. The judgement the exam tests: an association between A and B does not establish that intervening on A will change B.
3. A classmate says, "The sample was 240 students, which is large, so the causal claim is fine." Explain the error in that reasoning.
Read the worked answer
The classmate confuses sample size with study design. A large sample reduces random error - it makes the association more precisely estimated. It does nothing to rule out confounding, which is a structural feature of observational designs. A study of 240,000 self-reported sleepers would still be unable to assign causation, because the same unmeasured third factors (stress, illness, household schedule) would still be present.
What would support a causal claim is a design that manipulates sleep - for example, randomly assigning students to different sleep-duration conditions or to different start times - so that confounders are balanced across groups. The 240-student figure is relevant to how confident we are in the correlation, not to whether the correlation is causal.