What this SAT skill is testing
Questions about observational studies and experiments test whether you can judge what a study design actually proves. The arithmetic is usually light. The real work is matching the conclusion to the way participants were selected and treated.
Core rule: random selection supports generalizing to a population; random assignment supports a cause-and-effect conclusion.
Study types and valid conclusions
- Observational study: researchers observe existing behavior or conditions. It can reveal an association, but it cannot establish causation.
- Controlled experiment: researchers assign treatments and compare outcomes. Random assignment helps balance other variables and can support causation.
- Random sample: members of a population have a known chance of selection. A representative random sample supports generalizing results to that population.
- Convenience or voluntary-response sample: participants are easy to reach or choose themselves. These samples can be biased.
Worked SAT-style example
A school randomly selects 120 students from all enrolled students. Half are randomly assigned to use a new study app, and half use the existing method. The app group earns a higher average quiz score.
What is justified? Random selection makes the sample suitable for generalizing to the school's students. Random assignment makes a causal conclusion about the app reasonable. The conclusion should still stay within the population represented by the sample: students at that school.
A four-question decision method
- Who was studied? Identify the population and the sample.
- How were they selected? Look for random sampling, convenience sampling, or voluntary response.
- Was a treatment assigned? If not, the study is observational.
- What does the answer claim? Check separately for generalization and causation.
Common SAT traps
- Claiming that one variable caused another because they are associated.
- Generalizing from a narrow sample to a much broader population.
- Assuming a large sample is automatically unbiased.
- Confusing random assignment with random selection.
- Ignoring a confounding variable that could explain the result.
Fast review checklist
Circle the words randomly selected and randomly assigned whenever they appear. Then write two quick labels: generalize? and cause? This turns a paragraph-long study into two small decisions.
In advanced questions, the best answer is often the most carefully limited statement. Be skeptical of choices containing words such as proves, always, or all people unless the design truly supports them.
