Skip to main content

3.1 Estimators

add to ChatGPT
Verified for the 2027 examUpdated September 2026Fiveable Content Team
Fiveable

📊AP Statistics Unit 3 Review

QR code for AP Statistics practice questions

3.1 Estimators

An estimator is a rule for using sample data to estimate a population parameter. In the revised AP Statistics course, this topic focuses on point estimators, especially sample statistics like p^\hat{p}, and on the idea that an estimator is unbiased when it does not systematically overestimate or underestimate the true parameter.

Confidence intervals and hypothesis tests later in the unit both depend on estimating a population parameter from a sample statistic, which is why this topic opens the inference unit. The exam may ask you to identify a point estimate, explain why an estimator is or is not unbiased, or compare estimators using bias and variability.

Unbiased Estimators Center on the Parameter

  • A sample statistic is a point estimator for the matching population parameter.
  • A point estimate is the actual numerical value of the statistic from one sample.
  • An estimator is unbiased if, on average across repeated samples, it neither overestimates nor underestimates the parameter.
  • Bias is about center; variability is about spread.
  • A larger sample size usually reduces variability, but it does not fix a biased method.

Parameter, Statistic, Estimator, and Estimate

These words are related but not interchangeable.

  • A parameter is the population value you want to know.
  • A statistic is a value computed from a sample.
  • An estimator is the statistic viewed as a rule for estimating the parameter.
  • An estimate is the actual numerical answer from one sample.

For example, if pp is a population proportion, then p^\hat{p} is a point estimator for pp. Once you calculate p^\hat{p} from one sample, that number is your point estimate.

What It Means for an Estimator to Be Unbiased

An estimator is unbiased if, over many repeated samples, its average value equals the true population parameter.

That does not mean every single sample estimate lands exactly on the parameter. Individual sample estimates still vary. Unbiasedness is a long-run property.

For revised Unit 3, that matters because later inference procedures assume you understand that a sample proportion can be off in one sample without the estimator itself being biased.

Bias Versus Variability

These are different ideas:

  • Bias asks whether the estimator is centered on the true parameter.
  • Variability asks how spread out the estimates are from sample to sample.

An estimator can be:

  • unbiased but highly variable
  • biased but tightly clustered
  • unbiased and low-variability

The ideal estimator has low bias and low variability.

Why Sample Size Matters

Increasing sample size usually makes the sampling distribution tighter, so estimates vary less from sample to sample.

But a larger sample does not repair a biased sampling or response process. If the method is flawed, more data just gives a more stable wrong answer.

Judging Bias From Repeated Samples

MCQ:

  • Identify which sample statistic estimates which population parameter.
  • Distinguish between a biased method and an unbiased estimator with natural variability.
  • Watch for questions where a single estimate misses the parameter; that alone does not prove bias.

Free Response:

  • Define unbiasedness in long-run language: "on average across repeated samples."
  • Use the correct parameter and statistic in context.
  • Separate comments about sampling method from comments about the estimator itself.

Common Trap: A single estimate being too high or too low is not enough to prove the estimator is biased. Bias is about a repeated pattern across many samples.

Common Misconceptions

  • Misconception: Unbiased means every sample estimate is correct.
    Actually: It does not.
  • Misconception: A larger sample removes bias.
    Actually: It only reduces variability.
  • Misconception: Low variability means unbiased.
    Actually: A tight cluster can still be centered on the wrong value.
  • Misconception: Bias and skewness are the same thing.
    Actually: They are not.

Practice Questions for 3.1

1 of 10

A random sample of 50 households in a town was surveyed about the number of pets they own. The results are shown in the bar chart above.

Based on the sample data, what is the point estimate for the population mean number of pets per household?

A.

1.00

B.

1.12

C.

11.2

D.

0.40

Vocabulary

The following words are mentioned explicitly in the AP® course framework for this topic.

Term

Definition

biased

A property of an estimator where the average value of the estimator does not equal the population parameter being estimated.

estimator

A statistic used to estimate or approximate the value of a population parameter based on sample data.

population

The entire group of individuals or items from which a sample is drawn and about which conclusions are to be made.

population parameter

A numerical characteristic of an entire population, such as the mean, proportion, or standard deviation.

sample

A subset of individuals or items selected from a population for the purpose of data collection and analysis.

sample statistic

A numerical value calculated from sample data that is used to estimate the corresponding population parameter.

statistic

Numerical summaries or measures calculated from sample data, such as mean, median, or standard deviation.

unbiased

A property of an estimator where the average value of the estimator equals the population parameter being estimated.

variability

The spread or dispersion of data values in a distribution.

See Fiveable more often in Google

Mark Fiveable as a preferred source and Google is more likely to show you our guides in Top Stories, with a "preferred" badge in AI Mode and AI Overviews.

Prefer us on Google