5.4 Biased and Unbiased Point Estimates

A point estimator is a sample statistic you use to estimate a population parameter, and it is unbiased when its average value across all possible samples equals the true parameter. The sample mean and sample proportion are unbiased estimators, and bias is about being centered on the target while variability is about how spread out the estimates are.

Unbiased Estimator AP Stats

In AP Statistics, an unbiased estimator is a statistic whose sampling distribution is centered at the population parameter it estimates. That means the estimator is correct on average across many random samples, even though any single sample estimate can be too high or too low.

For Topic 5.4, focus on the difference between bias and variability. Bias is about center, while variability is about spread. Increasing sample size usually reduces variability, but it does not fix a biased sampling method.

more resources to help you study

practice multiple choiceFRQ practice & scoringcheatsheetsscore calculatorkey terms

Why This Matters for the AP Statistics Exam

This topic sets up everything that comes later in sampling distributions and inference. Once you understand that a statistic like x̄ or p̂ is just one value pulled from a whole distribution of possible values, the rest of Unit 5 makes more sense, and confidence intervals and hypothesis tests in later units build directly on it.

On the exam you can expect to:

• Explain whether an estimator is unbiased and justify it in context.
• Calculate point estimates from sample data.
• Compare estimators using both bias and variability.

These show up in multiple-choice questions and in free-response questions where you have to write a clear reason, not just give a number.

Key Takeaways

• An estimator is unbiased if, on average, its value equals the population parameter it is estimating.
• A sample statistic, like x̄ or p̂, is a point estimator of the matching population parameter.
• Every estimator has variability because each sample is only part of the population, so estimates change from sample to sample.
• Bias is about center (is the distribution of estimates centered on the parameter?), and variability is about spread.
• A larger sample size reduces variability, but it cannot remove bias caused by a flawed sampling method.
• A good estimator has low bias and low variability.

What Unbiased Really Means

An unbiased estimator produces estimates that, on average, land on the true population parameter. If you repeatedly took samples and calculated the statistic each time, the mean of all those estimates would equal the parameter.

Say you want the mean height of every student in your school. You take a sample and find its mean height. If the long-run average of your sample means equals the true school mean, your estimator is unbiased. If your method consistently runs too high or too low, it is biased.

In notation, the sample mean is unbiased when the mean of the sampling distribution of x̄ equals the population mean μ. The same idea applies to proportions: the average of all possible p̂ values equals the population proportion p.

Variability in a Sampling Distribution

A sampling distribution has low variability when the statistics from different samples are close to each other. You can never reach zero variability because each sample is only a subset of the population, so there is always some sampling error.

The fix for high variability is straightforward: increase your sample size. Bigger samples produce sample statistics that cluster more tightly around the center.

Bias and Variability Together

These two ideas describe different problems, so keep them separate.

• Bias is about whether the distribution of estimates is centered on the parameter. If the whole distribution sits to one side of the parameter, the estimator is biased.
• Variability is about spread. The more spread out the estimates are, the more variability the estimator has.

Increasing sample size lowers variability. It does not fix bias. If your sampling method is flawed, collecting more data just gives you a tighter cluster around the wrong value.

A bullseye is a useful picture. The center is the true parameter, and each shot is an estimate.

• Low bias, high variability: shots scatter around the bullseye but average out near it.
• Low variability, high bias: shots cluster tightly but land away from the bullseye.
• The goal is low bias and low variability: estimates that are both accurate and consistent.

A Quick Note on Skewness

Skewness measures how symmetric a distribution is. A symmetric distribution looks roughly the same on both sides of its center, like a bell curve. A skewed distribution is lopsided, with more values bunched on one side.

A sampling distribution that is centered on the parameter points toward low bias, but symmetry alone does not guarantee an unbiased estimator. The sampling method and how the data were collected also matter, which connects back to study design in Unit 3.

How to Use This on the AP Statistics Exam

Free Response

When a question asks whether an estimator is unbiased, anchor your answer to the definition: an estimator is unbiased if its average value across all possible samples equals the parameter. Tie that to whether the sampling method is random and representative. Write the reason in context, not just "yes" or "no."

Problem Solving

To find a point estimate, calculate the matching sample statistic. Use x̄ to estimate μ and p̂ to estimate p. Report it with the correct context and units.

Common Trap

If you learn the true parameter and your estimates keep landing on one side of it, that signals bias, not just bad luck. A single sample being off is normal variability; a consistent pattern of being too high or too low is bias.

Practice Problem

Suppose you are asked to estimate the mean income of all households in your town. You use a sample of 100 households selected with a random sampling method. After collecting the data, you calculate the sample mean income to be $50,000.

a) Explain whether or not this sample is biased, and give a reason for your answer.

b) Explain whether or not the sample mean income is an unbiased estimator of the population mean income, and give a reason for your answer.

c) Suppose you later learn that the true population mean income is actually $55,000. Explain how this information affects your conclusions about the bias of the sample and the estimator in parts (a) and (b).

d) Discuss one potential source of bias that could have affected the results of this study, and explain how it could have influenced the estimate of the population mean income. (This connects to Unit 3: Collecting Data.)

Answer

a) Since the sample of 100 households was selected using a random sampling method (such as an SRS, stratified, or cluster sample), the sampling method is not biased because it tends to produce samples that are representative of the population.

b) The sample mean income is an unbiased estimator of the population mean income when the sample is selected randomly and is representative. Because a random sampling method was used, the sample mean income is an unbiased estimator, meaning its average value across all possible samples equals the population mean.

c) A single sample mean of $50,000 differing from the true mean of $55,000 does not by itself prove bias. Estimators have natural variability, so one sample landing below the true value is expected. Bias would be shown by a consistent pattern of underestimates across many samples, not by one sample being off.

d) One potential source of bias is nonresponse bias, which occurs when certain groups are more or less likely to respond. If higher-income households are more likely to respond, the estimate could be pushed too high. If lower-income households are more likely to respond, the estimate could be pushed too low. Either way, the sampling and response process, not the estimator itself, would be the source of the bias.

Common Misconceptions

• "Unbiased means the sample matches the population exactly." Not true. Unbiased is about the long-run average of the statistic equaling the parameter, not any single sample hitting it.
• "A bigger sample removes bias." A bigger sample reduces variability, but it cannot fix bias from a flawed sampling method. More data just tightens the cluster around the wrong value.
• "One estimate that is off proves the estimator is biased." A single estimate being above or below the parameter is normal variability. Bias is a consistent, systematic pattern across many samples.
• "Low variability means unbiased." Estimates can be tightly clustered yet centered in the wrong place. Bias is about center, variability is about spread.
• "Bias and skewness are the same thing." Skewness describes the shape of a distribution. Bias describes whether the center of the estimates lands on the true parameter.

Related AP Statistics Guides

• Unit 5 Overview: Sampling Distributions
• 5.1 Introducing Statistics: Why Is My Sample Not Like Yours?
• 5.5 Sampling Distributions for Sample Proportions
• 5.2 The Normal Distribution, Revisited
• 5.3 The Central Limit Theorem
• 5.7 Sampling Distributions for Sample Means