SAT Evaluating Statistical Claims
Last updated: May 2, 2026
Evaluating Statistical Claims questions are one of the highest-leverage areas to study for the SAT. This guide breaks down the rule, the elements you need to recognize, the named traps that catch most students, and a memory aid that scales to test day. Read it once, then practice the same sub-topic adaptively in the app.
The rule
On the Digital SAT, evaluating statistical claims means asking three questions about a study: (1) Was the sample randomly selected from the population the claim describes? (2) Were treatments randomly assigned to subjects? (3) Does the margin of error or confidence interval actually support the strength of the conclusion? Random selection lets you generalize to a population; random assignment lets you claim cause and effect; and the conclusion can never be stronger than the interval the data produce.
Elements breakdown
Random Selection (Generalizability)
Whether subjects were chosen at random from a defined population determines which group the results apply to.
- Identify the population sampled
- Check if selection method was random
- Limit conclusions to that population
- Reject claims about broader groups
Common examples:
- Random sample of 400 Glenmark High seniors $\to$ conclusions apply to Glenmark High seniors only, not all teens
Random Assignment (Causation)
Whether subjects were randomly assigned to treatment vs. control determines whether you can claim cause and effect.
- Look for the phrase 'randomly assigned'
- If assignment is random, causal language is allowed
- If only observation, use 'associated with' or 'linked to'
- Never infer cause from a correlational study
Margin of Error and Confidence Intervals
A margin of error builds an interval around a sample statistic; the conclusion must respect that interval.
- Compute interval as estimate $\pm$ margin
- Larger sample $\to$ smaller margin of error
- Higher confidence $\to$ wider interval
- Conclusions must stay inside the interval
Common examples:
- Estimate $42\%$ with margin $\pm 3\%$ gives the interval $[39\%, 45\%]$
Matching Conclusion to Study Design
The strongest defensible conclusion is fixed by how the study was run; stronger language is wrong even if the data look impressive.
- Random sample + random assignment $\to$ cause-and-effect for the population
- Random sample only $\to$ association generalizes to population
- Random assignment only $\to$ cause-and-effect for participants
- Neither $\to$ description of the participants only
Common patterns and traps
Population Overreach
The wrong answer takes a result from a narrow, specific sample and projects it onto a much broader group. The data may be sound for the sampled population, but the conclusion swaps that group for 'all teenagers,' 'all Americans,' or 'students in general.' This is the most common trap on R&W-flavored statistics items and shows up constantly in Math evaluating-claims items too.
A choice that begins 'all students' or 'most adults' when the study only sampled one school, one city, or one workplace.
Causation From Correlation
The wrong answer uses cause-and-effect verbs ('causes,' 'leads to,' 'improves,' 'reduces') even though the study only observed an association. If the problem never says 'randomly assigned,' you cannot claim causation. The correct answer will use neutral language like 'associated with,' 'linked to,' or 'correlated with.'
A choice saying 'the program raised test scores' when subjects chose to enroll themselves rather than being randomly assigned.
Ignoring the Margin of Error
The wrong answer states the sample statistic as if it were the exact population value, dropping the $\pm$ margin entirely. The SAT rewards conclusions that live inside the interval $[\text{estimate} - \text{margin}, \text{estimate} + \text{margin}]$. Any choice that pins down a single number as the truth has overreached.
A choice asserting 'exactly $47\%$ of residents agree' when the survey reported $47\%$ with margin $\pm 3\%$.
Misreading the Confidence Interval
The wrong answer flips what the interval describes — for instance, claiming it gives the probability that an individual person holds a view, or that $95\%$ of people fall inside it. The interval is a range of plausible values for the population parameter (like a mean or proportion), not for any single person.
A choice saying '$95\%$ of customers spend between \$40 and \$48' when the interval is for the mean spending of all customers.
Smaller-Margin-Means-Bigger-Sample Logic
Many evaluating-claims items ask which of two studies gives a more reliable estimate. A larger random sample (with the same confidence level) produces a smaller margin of error. Watch for choices that reverse this or that compare two samples on irrelevant traits like 'age range' instead of size.
Two studies with the same confidence level — pick the one with the larger random sample because its margin of error will be smaller.
How it works
Treat every statistical-claims question as a matching exercise between the study's design and the answer's verb. Suppose researcher Marta Reyes randomly selects $250$ adults from the city of Bridgeport and finds that $62\%$ support a new transit fee, with a margin of error of $\pm 4$ percentage points at the $95\%$ confidence level. Because the sample is random, you can generalize to Bridgeport adults, but only inside the interval $[58\%, 66\%]$. A choice that says 'most adults nationwide support the fee' fails on selection (wrong population). A choice that says 'exactly $62\%$ of Bridgeport adults support it' ignores the margin of error. A choice that says 'the fee causes adults to support transit' invents causation from an observational poll. The right answer will say something modest, like 'between $58\%$ and $66\%$ of Bridgeport adults likely support the fee.'
Worked examples
A researcher randomly selected $400$ adults from the residents of Halverston County and asked whether they had visited the public library in the past year. Of those surveyed, $58\%$ said yes. The margin of error for this estimate was $\pm 3$ percentage points at the $95\%$ confidence level.
Which of the following is the most appropriate conclusion based on the study?
- A Exactly $58\%$ of all adults in Halverston County visited the public library in the past year.
- B It is plausible that between $55\%$ and $61\%$ of all adults in Halverston County visited the public library in the past year. ✓ Correct
- C It is plausible that between $55\%$ and $61\%$ of all adults in the United States visited a public library in the past year.
- D Visiting the public library caused $58\%$ of Halverston County adults to support library funding.
Why B is correct: The sample was random and drawn from Halverston County adults, so the result generalizes to that population only. The margin of error of $\pm 3$ produces the interval $[55\%, 61\%]$, and a defensible conclusion must respect that interval rather than pinning down a single number. Choice B states the population correctly and stays inside the interval.
Why each wrong choice fails:
- A: This treats the sample statistic $58\%$ as the exact population value and ignores the margin of error of $\pm 3$ percentage points. (Ignoring the Margin of Error)
- C: The sample was drawn only from Halverston County, so the conclusion cannot be extended to all adults in the United States. (Population Overreach)
- D: The study was an observational survey, not a randomized experiment, so causal language about library funding is not supported, and the claim invents a question the study didn't ask. (Causation From Correlation)
At Fairlow Middle School, $120$ sixth-graders volunteered for an after-school reading program; the remaining $180$ sixth-graders did not volunteer. At the end of the year, the volunteers' average score on a standardized reading test was $7$ points higher than the non-volunteers'.
Which of the following is the most appropriate conclusion based on the study?
- A The after-school reading program caused sixth-graders at Fairlow Middle School to score $7$ points higher on the reading test.
- B The after-school reading program would cause any sixth-grader in the country to score about $7$ points higher.
- C At Fairlow Middle School, sixth-graders who participated in the program had higher average reading scores than those who did not. ✓ Correct
- D At Fairlow Middle School, the program had no effect on reading scores.
Why C is correct: Students self-selected into the program rather than being randomly assigned, so any difference in scores might come from underlying differences between volunteers and non-volunteers (motivation, prior skill, parental involvement). Without random assignment, only an associational claim is defensible, and choice C makes exactly that claim, limited to the school where the data were collected.
Why each wrong choice fails:
- A: Because participation was voluntary, the program was not randomly assigned, so a cause-and-effect conclusion is not justified — the score gap could be due to who chose to enroll. (Causation From Correlation)
- B: The study involved only one school's sixth-graders, so the conclusion cannot be projected onto sixth-graders nationwide, and it also wrongly asserts causation. (Population Overreach)
- D: The volunteers scored $7$ points higher on average, which is a real difference in the data; saying the program had 'no effect' contradicts what was actually observed in the sample.
Two researchers each estimated the average number of hours per week that adults in Carrowville spend exercising. Researcher Fei Liu used a random sample of $200$ adults and reported a $95\%$ confidence interval of $4.2 \pm 0.6$ hours. Researcher Ben Okafor used a random sample of $800$ adults from the same population and reported a $95\%$ confidence interval at the same confidence level.
Which of the following best compares the margins of error of the two studies?
- A Okafor's margin of error is larger than Liu's because his sample is larger.
- B Okafor's margin of error is smaller than Liu's because his sample is larger. ✓ Correct
- C The two margins of error are equal because both studies used the same confidence level.
- D The two margins of error cannot be compared without knowing the standard deviation of exercise hours.
Why B is correct: For random samples drawn from the same population at the same confidence level, the margin of error decreases as the sample size increases. Okafor's sample of $800$ is four times the size of Liu's sample of $200$, so his margin of error will be smaller — roughly half, since margin of error scales with $\frac{1}{\sqrt{n}}$.
Why each wrong choice fails:
- A: This reverses the relationship between sample size and margin of error; a larger random sample produces a smaller, not larger, margin of error. (Smaller-Margin-Means-Bigger-Sample Logic)
- C: Confidence level alone does not fix the margin of error; sample size also matters, and the two samples here differ in size. (Smaller-Margin-Means-Bigger-Sample Logic)
- D: Even without an exact standard deviation, you can still compare margins of error qualitatively because both samples come from the same population at the same confidence level, so the larger sample must yield the smaller margin.
Memory aid
Use the 3-S check: Sample (who was picked?), Setup (random assignment?), Span (does the interval support the claim?). If any S fails, the conclusion fails.
Key distinction
Random selection lets you generalize; random assignment lets you say 'caused.' These are independent, and the SAT will test both separately.
Summary
Match the verb of the conclusion to the design of the study and the width of the interval — never overreach on population, causation, or precision.
Practice evaluating statistical claims adaptively
Reading the rule is the start. Working SAT-format questions on this sub-topic with adaptive selection, watching your mastery score climb in real time, and seeing the items you missed return on a spaced-repetition schedule — that's where score lift actually happens. Free for seven days. No credit card required.
Start your free 7-day trialFrequently asked questions
What is evaluating statistical claims on the SAT?
On the Digital SAT, evaluating statistical claims means asking three questions about a study: (1) Was the sample randomly selected from the population the claim describes? (2) Were treatments randomly assigned to subjects? (3) Does the margin of error or confidence interval actually support the strength of the conclusion? Random selection lets you generalize to a population; random assignment lets you claim cause and effect; and the conclusion can never be stronger than the interval the data produce.
How do I practice evaluating statistical claims questions?
The fastest way to improve on evaluating statistical claims is targeted, adaptive practice — working questions that focus on your specific weak spots within this sub-topic, getting immediate feedback, and revisiting items you missed on a spaced-repetition schedule. Neureto's adaptive engine does this automatically across the SAT; start a free 7-day trial to see your sub-topic mastery climb in real time.
What's the most important distinction to remember for evaluating statistical claims?
Random selection lets you generalize; random assignment lets you say 'caused.' These are independent, and the SAT will test both separately.
Is there a memory aid for evaluating statistical claims questions?
Use the 3-S check: Sample (who was picked?), Setup (random assignment?), Span (does the interval support the claim?). If any S fails, the conclusion fails.
What's a common trap on evaluating statistical claims questions?
Generalizing past the sampled population
What's a common trap on evaluating statistical claims questions?
Claiming causation without random assignment
Ready to drill these patterns?
Take a free SAT assessment — about 15 minutes and Neureto will route more evaluating statistical claims questions your way until your sub-topic mastery score reflects real improvement, not luck. Free for seven days. No credit card required.
Start your free 7-day trial