Sample Statistics, Margin of Error & Evaluating Claims
Master Digital SAT margin of error, confidence intervals, and evaluating claims. Learn when random sampling proves generalization and random assignment proves causation.
What you'll do in this lesson
A voice-first session with the Crimsora tutor on Sample Statistics, Margin of Error & Evaluating Claims, then targeted practice and FRQs — with the tutor adapting to where you get stuck.
What this lesson covers
On the Digital SAT, one of the most misunderstood question types asks you to interpret a survey result: a sample statistic paired with a margin of error, or a claim about causation. These questions rarely require calculation — they reward careful reading and knowing exactly what a study design permits you to conclude.
In this lesson you will learn to read a margin of error as a plausible range for the true population value, to judge whether a claim is supported by the data, and to separate two ideas the test loves to blur: random sampling (which lets you generalize to a population) and random assignment (which lets you claim cause and effect). Get these distinctions right and you can answer these items in seconds.
In this lesson you will learn to read a margin of error as a plausible range for the true population value, to judge whether a claim is supported by the data, and to separate two ideas the test loves to blur: random sampling (which lets you generalize to a population) and random assignment (which lets you claim cause and effect). Get these distinctions right and you can answer these items in seconds.
Sample Statistics and the Plausible Range
A population is the entire group you want to know about; a sample is the smaller group actually measured. A sample statistic — such as a sample mean or sample proportion — is your best single estimate of the unknown population parameter.
But a single sample almost never lands exactly on the true value. That is why the SAT reports a margin of error. The margin of error creates an interval around the statistic that is a plausible range for the population value. If a survey finds a mean of with a margin of error of , the plausible values run from to .
The correct SAT interpretation sounds like this: "It is plausible that the true population mean is between 31 and 37." Wrong interpretations describe the sample itself ("the sample mean is between 31 and 37" — no, the sample mean is exactly 34) or make an absolute guarantee ("the population mean must be 34").
A key fact the SAT tests: a larger sample size and less variability both produce a smaller margin of error, giving a narrower, more precise estimate. You will not compute the margin of error on the SAT — you interpret it and reason about what changes it.
But a single sample almost never lands exactly on the true value. That is why the SAT reports a margin of error. The margin of error creates an interval around the statistic that is a plausible range for the population value. If a survey finds a mean of with a margin of error of , the plausible values run from to .
The correct SAT interpretation sounds like this: "It is plausible that the true population mean is between 31 and 37." Wrong interpretations describe the sample itself ("the sample mean is between 31 and 37" — no, the sample mean is exactly 34) or make an absolute guarantee ("the population mean must be 34").
A key fact the SAT tests: a larger sample size and less variability both produce a smaller margin of error, giving a narrower, more precise estimate. You will not compute the margin of error on the SAT — you interpret it and reason about what changes it.
Generalization: What Random Sampling Justifies
Whether you can extend results from a sample to a larger population depends entirely on how the sample was selected. If subjects were chosen through random sampling from a population, the results can be generalized to that population — and only that population.
The SAT builds trap answers by generalizing too far. Suppose researchers randomly selected 200 students from Lincoln High School. The results generalize to students at Lincoln High School, not to all students in the state, not to all teenagers, and not to adults. The population you can talk about is the population you sampled from.
When a sample is not random — for example, volunteers, or only people who respond to an online poll — results should not be generalized at all, because the sample may be biased.
On test day, first ask: "Was the sample randomly selected, and from what group?" That single question decides most generalization items.
The SAT builds trap answers by generalizing too far. Suppose researchers randomly selected 200 students from Lincoln High School. The results generalize to students at Lincoln High School, not to all students in the state, not to all teenagers, and not to adults. The population you can talk about is the population you sampled from.
When a sample is not random — for example, volunteers, or only people who respond to an online poll — results should not be generalized at all, because the sample may be biased.
| Sampling method | Can you generalize? | To whom? |
|---|---|---|
| Random sample from a group | Yes | That specific group only |
| Volunteers / self-selected | No | Not reliably to anyone |
| Convenience (whoever is nearby) | No | Not reliably to anyone |
Causation: What Random Assignment Justifies
Generalizing is about who was studied; establishing cause and effect is about how treatments were assigned. To conclude that one variable causes a change in another, subjects must be randomly assigned to treatment groups in a controlled experiment.
Random assignment balances out other factors between groups, so any difference in outcomes can be attributed to the treatment. Without it — for instance in an observational study where researchers just record what people already do — you can identify an association or correlation, but you cannot claim causation, because a lurking variable might explain the link.
The SAT phrases correct causal answers carefully: "the treatment caused the change" appears only when random assignment is present. If the study is observational, the defensible conclusion is limited to "there is an association between" the variables.
Both can appear in one study. A study with random sampling AND random assignment supports both generalizing and a causal claim; a study with neither supports neither.
Random assignment balances out other factors between groups, so any difference in outcomes can be attributed to the treatment. Without it — for instance in an observational study where researchers just record what people already do — you can identify an association or correlation, but you cannot claim causation, because a lurking variable might explain the link.
The SAT phrases correct causal answers carefully: "the treatment caused the change" appears only when random assignment is present. If the study is observational, the defensible conclusion is limited to "there is an association between" the variables.
| Feature | Random sampling | Random assignment |
|---|---|---|
| Purpose | Choosing who is studied | Placing subjects into groups |
| Justifies | Generalization to population | Cause-and-effect claim |
| Absence means | Cannot generalize | Cannot claim causation |
How the SAT Frames These Questions
These items are almost always word problems with an answer set of four interpretive sentences. Your job is to pick the statement that matches exactly what the study design supports — no more, no less.
Common wrong-answer patterns to reject: overgeneralizing beyond the sampled population; asserting causation from an observational study; describing the sample when the question asks about the population; and treating the margin of error as an exact or guaranteed value rather than a plausible range.
A reliable checklist: identify the sample statistic and its margin of error, write the interval as statistic margin. Then ask whether the sample was random (generalization) and whether treatments were randomly assigned (causation). Match those two answers against each choice.
Watch the wording of the plausible range too. "The estimate is likely within the interval" or "it is plausible the population value lies in the interval" are safe. "The population value is definitely" or "exactly" is a trap. The SAT rewards the humble, precise statement every time — statistics gives you plausible ranges and supported inferences, not certainties.
Common wrong-answer patterns to reject: overgeneralizing beyond the sampled population; asserting causation from an observational study; describing the sample when the question asks about the population; and treating the margin of error as an exact or guaranteed value rather than a plausible range.
A reliable checklist: identify the sample statistic and its margin of error, write the interval as statistic margin. Then ask whether the sample was random (generalization) and whether treatments were randomly assigned (causation). Match those two answers against each choice.
Watch the wording of the plausible range too. "The estimate is likely within the interval" or "it is plausible the population value lies in the interval" are safe. "The population value is definitely" or "exactly" is a trap. The SAT rewards the humble, precise statement every time — statistics gives you plausible ranges and supported inferences, not certainties.
Key terms
- Population.
- The entire group of individuals or items a study aims to describe.
- Sample.
- The subset of the population that is actually measured or surveyed.
- Sample statistic.
- A value computed from the sample (such as a mean or proportion) used to estimate the population parameter.
- Margin of error.
- A value added to and subtracted from a sample statistic to form a plausible range for the population parameter.
- Random sampling.
- Selecting sample members by chance from a population; this is what justifies generalizing results to that population.
- Random assignment.
- Randomly placing subjects into treatment groups in an experiment; this is what justifies concluding cause and effect.
- Observational study.
- A study that records existing behavior without assigning treatments, so it can show association but not causation.
- Generalization.
- Extending sample results to the larger population from which the sample was randomly drawn.
Worked example
A researcher randomly selected 150 residents of Maple County and measured their weekly commute times. The sample mean was 42 minutes with a margin of error of 4 minutes at a 95% confidence level. Which conclusion is best supported, and could the researcher claim that living in Maple County causes longer commutes?
First build the plausible range from the statistic and margin of error: and . So it is plausible that the true mean weekly commute time for all Maple County residents is between and minutes.
Next, check generalization. The 150 residents were selected by random sampling from Maple County, so the results generalize — but only to residents of Maple County, not to any wider group such as the whole state.
Now check causation. This is a survey that measured commute times; there is no treatment and no random assignment of residents to conditions. Therefore no cause-and-effect conclusion is possible. The researcher cannot claim that living in Maple County causes longer commutes.
Best supported conclusion: "It is plausible that the mean weekly commute time for all Maple County residents is between 38 and 46 minutes." Any statement claiming an exact value, extending to other counties, or asserting causation would go beyond what the data support.
Next, check generalization. The 150 residents were selected by random sampling from Maple County, so the results generalize — but only to residents of Maple County, not to any wider group such as the whole state.
Now check causation. This is a survey that measured commute times; there is no treatment and no random assignment of residents to conditions. Therefore no cause-and-effect conclusion is possible. The researcher cannot claim that living in Maple County causes longer commutes.
Best supported conclusion: "It is plausible that the mean weekly commute time for all Maple County residents is between 38 and 46 minutes." Any statement claiming an exact value, extending to other counties, or asserting causation would go beyond what the data support.
Practice questions
A nutrition team randomly assigned 300 volunteers to either a high-fiber diet or a standard diet and, after 8 weeks, found the high-fiber group had significantly lower cholesterol. The volunteers were not randomly selected. Which conclusion is most appropriate?
- The high-fiber diet causes lower cholesterol, and this result applies to all adults
- The high-fiber diet caused lower cholesterol among the volunteers in the study
- There is only an association between fiber and cholesterol, with no causal link possible
- The results can be generalized to all adults but no cause can be determined
Answer: The high-fiber diet caused lower cholesterol among the volunteers in the study
Because subjects were randomly assigned to treatment groups, a causal conclusion is justified — the diet caused the change. However, the volunteers were not randomly selected, so results cannot be generalized to all adults. The correct statement claims causation but limits it to the study participants.
A poll of 500 randomly selected voters in a city found that 58% support a new transit plan, with a margin of error of 3 percentage points. Write a statement that correctly interprets this margin of error, and explain why it would be wrong to say exactly 58% of all city voters support the plan.
Answer: It is plausible that the true percentage of all city voters who support the plan is between 55% and 61%.
The plausible range is to . The 58% is only the sample statistic — the best single estimate — not a guaranteed population value. Because a sample rarely matches the population exactly, the margin of error acknowledges uncertainty and gives an interval rather than a single exact figure. Since the sample was random, this range applies to all city voters.
Researchers observed 400 randomly selected teenagers and recorded that those who slept more hours also reported higher grades. Which conclusion is supported?
- Getting more sleep causes higher grades in teenagers
- More sleep causes higher grades, but only for the teenagers studied
- There is an association between more sleep and higher grades among teenagers
- No conclusion about teenagers can be drawn from this study
Answer: There is an association between more sleep and higher grades among teenagers
This is an observational study — researchers only recorded existing behavior with no random assignment to a sleep treatment. So causation cannot be claimed. Because the sample was random, results generalize to teenagers, but the only defensible relationship is an association, not cause and effect.
FAQ
- What is the difference between random sampling and random assignment on the SAT?
- Random sampling is how subjects are chosen from a population, and it is what lets you generalize results to that population. Random assignment is how subjects are placed into treatment groups in an experiment, and it is what lets you claim cause and effect. They answer two different questions: who was studied versus how treatments were given.
- How do I interpret a margin of error?
- Add and subtract the margin of error from the sample statistic to form an interval. That interval is a plausible range for the true population value. For example, a mean of 50 with a margin of error of 2 means it is plausible the population mean is between 48 and 52. It does not guarantee an exact value.
- When can I say a study proves causation?
- Only when subjects were randomly assigned to treatment groups in a controlled experiment. In an observational study, where researchers simply record what already happens, you can report an association but not causation, because an outside variable might explain the relationship.
- Do I need to calculate the margin of error on the Digital SAT?
- No. The SAT gives you the margin of error and asks you to interpret it, form the plausible range, or reason about what would make it smaller. A larger sample size and less variability both shrink the margin of error, producing a more precise estimate.
Learn this with a teacher, not a page
The Crimsora tutor teaches Sample Statistics, Margin of Error & Evaluating Claims live — explaining on a whiteboard, asking you questions, and adapting to where you get stuck.