The Digital SAT tests whether you understand the difference between observational studies and experiments — and crucially, when you can claim causation vs. mere association. This is a reasoning skill, not a computation skill.
Core Concepts
Observational Study
Researchers observe subjects without intervening. They measure variables as they naturally occur.
Example: Surveying coffee drinkers about their sleep patterns.
Can conclude: association/correlation. Cannot conclude: causation.
Experiment
Researchers impose a treatment on subjects and observe the effects. A well-designed experiment includes:
- Treatment and control groups.
- Random assignment to groups.
- Blinding (subjects don't know their group).
Example: Randomly assigning people to drink coffee or water, then measuring sleep.
Can conclude: causation (if well-designed with random assignment).
Key Distinction
| Observational Study | Experiment | |
|---|---|---|
| Treatment imposed? | No | Yes |
| Random assignment? | No | Yes |
| Can infer causation? | No | Yes |
| Can show association? | Yes | Yes |
Confounding Variables
A confounding variable is a third variable that affects both the independent and dependent variables, creating a false appearance of a direct relationship.
Example: Ice cream sales and drowning rates are correlated — but the confounding variable is hot weather.
Random Assignment vs. Random Selection
- Random selection (sampling) → results generalise to the population.
- Random assignment (to treatment groups) → causation can be inferred.
Best studies have BOTH.
Strategy Tips
Tip 1: "Causes" or "Due to" = Requires an Experiment
If the conclusion uses causal language, it's only valid if the study was a randomised experiment.
Tip 2: Surveys = Observational
Surveys never assign treatments. They observe.
Tip 3: Look for Random Assignment
If subjects were randomly assigned to groups, it's an experiment.
Worked Example: SAT-Style
A study found that people who exercise regularly have lower rates of depression. Can we conclude exercise prevents depression?
No — this is observational. People who exercise may differ in other ways (lifestyle, income, etc.). Confounding variables prevent a causal conclusion.
Worked Example: Example 2
Researchers randomly assigned 200 patients to receive either a new drug or a placebo. The drug group showed significantly better outcomes. Can we conclude the drug caused the improvement?
Yes — this is a randomised experiment with a control group (placebo).
Worked Example: Example 3
A company surveys its employees and finds those who take breaks have higher productivity. The company concludes breaks cause higher productivity. Is this valid?
No — this is observational. Perhaps more productive workers feel they can afford to take breaks.
Practice Problems
Problem 1
A study tracks students' TV watching and grades. Those who watch less have higher GPAs. Can we say TV watching causes lower grades?
Problem 2
Researchers randomly assign gardens to receive fertiliser A or B, then measure plant growth. What type of study is this?
Problem 3
A study randomly selects 1000 people and asks about diet and health. Is this an experiment?
Want to check your answers and get step-by-step solutions?
Common Mistakes
- Claiming causation from observational data. Only experiments with random assignment can show causation.
- Confusing random sampling with random assignment. They serve different purposes.
- Ignoring confounding variables. Always consider what other factors might explain the result.
Key Takeaways
Observational studies observe without treatment → association only.
Experiments impose treatments with random assignment → causation.
Random assignment enables causal conclusions.
Random selection enables generalisation.
Confounding variables explain apparent associations without true causation.
On the SAT: if the study is observational, the answer is NEVER "causes."