SAT Margin of Error & Sampling: Complete Guide

Margin of error and inference from samples are part of SAT Problem-Solving and Data Analysis.

College Board specifically includes:

• inference from sample statistics;
• margin of error;
• evaluating statistical claims;
• observational studies and experiments.

If you want to round out your Math prep, our guides on linear equations and quadratic equations cover other major skill areas, and the rest of our SAT Math guides cover everything else tested.

Population vs sample

Population

The entire group you want to understand.

Example: all 4,000 students at a school.

Sample

The smaller group actually measured.

Example: 200 selected students.

We use the sample to estimate something about the population.

Why random sampling matters

Suppose a school wants to estimate how many students use the bus.

It surveys only students waiting at the bus stop.

That sample is biased.

It overrepresents bus riders.

A random sample is more likely to represent the full population.

Margin of error

Suppose a poll estimates:

[ 52%\pm4% ]

A simple interpretation is that the plausible interval around the estimate is:

[ 48%\text{ to }56% ]

The SAT focuses mainly on interpreting the margin—not calculating advanced statistical formulas.

For a refresher on using your calculator efficiently on questions like this, see the Desmos on the Digital SAT guide.

Sample size and margin of error

All else equal:

larger random samples tend to produce smaller margins of error.

This is a frequent test idea.

If two properly designed surveys sample the same population:

• Survey A: 100 people
• Survey B: 1,000 people

Survey B generally has the smaller margin of error.

Bigger sample does not fix bias

Surveying 10,000 people from the wrong group can still produce a bad estimate.

Sample quality matters as well as sample size.

Random sample vs random assignment

These solve different problems.

Random sampling

Helps generalize from a sample to a population.

Random assignment

Helps support cause-and-effect conclusions in an experiment.

Do not confuse them.

Observational study

Researchers observe what already happens.

This can show an association.

It usually cannot establish causation as strongly as a well-designed randomized experiment.

Experiment

Researchers impose treatments.

If subjects are randomly assigned to treatment groups, differences can support a causal conclusion more strongly.

Original practice

1

A poll reports 61% ± 3%. What interval is represented?

2

Two random samples study the same population. One has 150 people and one has 1,500. Which would generally have the smaller margin of error?

3

A school surveys only members of the chess club to estimate how many students play chess. Main problem?

A. sample too random
B. selection bias
C. margin of error is always zero
D. random assignment

4

What does random assignment help support?

A. generalizing to every population
B. causal conclusions between treatments
C. eliminating all error
D. increasing survey response rate

Answers

1. 58% to 64%
2. The sample of 1,500
3. B
4. B

Common traps

• thinking a larger sample automatically removes bias;
• confusing random sampling with random assignment;
• treating observational correlation as proof of causation;
• reversing the meaning of margin of error.

Official source

• College Board Problem-Solving and Data Analysis
  https://satsuite.collegeboard.org/sat/whats-on-the-test/math/types/problem-solving

---

Trademark note: SAT® is a registered trademark of College Board. MastaPrep is not affiliated with, endorsed by, or sponsored by College Board.