Inferences & Claims
From Statistics
Master how to evaluate survey design, judge statistical claims, and draw valid conclusions from data on the PSAT.
What Are Statistical Inferences?
A statistical inference is a conclusion drawn about a population based on data collected from a sample. On the PSAT, you will decide whether a claim is supported by the study described.
Why Random Sampling Matters
For a conclusion about a population to be valid, the sample must be representative. The gold standard is a random sample, where every member of the population has an equal chance of being selected.
Sources of Bias
Bias occurs when the sample does not accurately represent the population. This makes any inference unreliable. The PSAT tests whether you can spot biased study designs.
Experiments vs. Observational Studies
The PSAT frequently asks: "Can this study establish causation?" The answer depends entirely on the study type.
The Two Core Rules of Inference
Understanding Margin of Error
The margin of error tells you how much the sample result might differ from the true population value. A survey result of 60% with a margin of error of \(\pm 4\%\) means the true value is likely between 56% and 64%.
Association vs. Causation
| Association | Causation | |
|---|---|---|
| Study type | Observational study | Controlled experiment |
| Assignment | No random assignment | Random assignment |
| Language | "linked to," "correlated with" | "caused by," "due to" |
| Confounders | May be present | Controlled for |
| PSAT answer | "is associated with" | "is likely caused by" |
How to Evaluate a Claim
Example 1: School Survey
Example 2: Fertilizer Experiment
Example 3: Margin of Error
Scope of Inference Matrix
Sample Size & Variability
The PSAT may ask how sample size affects the reliability of results. Larger samples reduce random variability and produce more consistent estimates.
Common Mistakes
Summary Checklist
| Question to Ask | If Yes | If No |
|---|---|---|
| Random sample? | Generalize to population | Only about the sample |
| Random assignment? | Can claim causation | Association only |
| Large sample? | Smaller margin of error | Larger margin of error |
| Intervals overlap? | No significant difference | Plausible difference |
| Volunteer sample? | Likely biased | Less bias concern |
Question 1 of 5
A nutritionist randomly selects 300 adults from a city of 50,000 and records their daily fruit intake. She finds that the average is 2.1 servings per day with a margin of error of \(\pm 0.3\) servings. She did not assign any dietary plans.
Which conclusion is best supported by this study?
Question 2 of 5
A teacher randomly assigns her 40 students into two groups. Group A uses a new study app for 4 weeks; Group B uses traditional textbooks. Group A scores an average of 8 points higher on the final exam. The students were not randomly selected; they are all from one class.
Which statement is best supported?
Question 3 of 5
A town puts up a sign-up sheet at the library asking residents to volunteer for a survey about park usage. 150 people sign up. Which problem most threatens the validity of generalizing the results to all town residents?
Question 4 of 5
Survey A polls 400 randomly selected voters and finds 52% support a ballot measure (\(\pm 4\%\)). Survey B polls 1,600 randomly selected voters and finds 52% support (\(\pm 2\%\)). What is the most accurate comparison?
Question 5 of 5
Researchers randomly select 500 employees from a large corporation and randomly assign them to either a flexible-hours group or a standard-hours group for 3 months. The flexible-hours group reports significantly higher job satisfaction.
Which conclusion is most supported?