Stating Hypotheses for a Population Mean

Start With the Population Mean

In “Complete Conditions Check for a Mean Inference Problem,” you practiced identifying the population mean and the observations relevant to a mean inference procedure. The next step is to state hypotheses about that mean. Hypotheses are claims about a population parameter—not about the sample mean you happened to observe.

This tutorial focuses on writing the hypotheses clearly. You will first define the parameter in context, then write a null hypothesis using equality and an alternative hypothesis that represents the question of interest. A well-stated parameter makes clear whose mean is being discussed and what is being measured.

Definition: The population mean $\mu$ is the average value of a quantitative variable for the entire population named in the problem. In a hypothesis test, the hypotheses state claims about this population mean.

For example, “the average fill is 500 mL” is not yet a complete parameter definition. A clearer definition names the relevant bottles and production process: let $\mu$ be the mean fill volume, in milliliters, of all bottles filled by a specified production line during a specified period. That definition identifies the population and the units.

How to Write the Two Hypotheses

The null hypothesis, written $H_0$, gives a reference value for the population mean. In a one-sample mean test, it is written with equality, such as $H_0:\mu=500$. The alternative hypothesis, written $H_a$, expresses the departure from that reference value that the question asks about. Common forms are $H_a:\mu<500$, $H_a:\mu>500$, and $H_a:\mu\ne500$.

Formula: A typical pair of hypotheses for a population mean is $$ H_0:\mu=\mu_0 \qquad\text{and}\qquad H_a:\mu<\mu_0,\ \mu>\mu_0,\ \text{or}\ \mu\ne\mu_0, $$ where $\mu_0$ is the reference value stated in the problem. Define $\mu$ in context before using it in the hypotheses.

The equality in $H_0$ matters. It specifies the reference value against which the data will be assessed. The alternative states what kind of difference or departure would be relevant to the question. For example, a concern that a filling line is delivering less than its target points toward an alternative below the target. Whether the question calls for a lower, higher, or any-difference alternative—and how to distinguish those choices—is the focus of “One-Sided Versus Two-Sided Alternatives for a Mean.”

The number in the hypotheses comes from the claim or benchmark in the problem, not from the observed sample mean. If a label says 500 mL, that claim supplies the reference value. A sample mean such as $\bar{x}=496.8$ mL does not replace the 500 in $H_0$, and it does not by itself determine the alternative. The sample results are used later to assess evidence against the null.

Writing routine: Define $\mu$ in words, identify the reference value $\mu_0$, write $H_0:\mu=\mu_0$, and then use the question’s wording to state the alternative. Check that both hypotheses concern the same population mean and use the same units.

When stating hypotheses, include the parameter definition alongside the symbols. Writing only $H_0:\mu=500$ leaves the reader to guess what $\mu$ measures and which population it describes. A concise definition supplies that context without repeating every detail of the study.

Worked Examples

Worked Example: Investigating a 500 mL Bottle-Filling Target

A fictional beverage company uses a filling line intended to place an average of 500 mL in each bottle. A quality-control team randomly samples 40 bottles from one production shift. The sample mean is $\bar{x}=496.8$ mL and the sample standard deviation is $s=6.4$ mL. The team wants to investigate whether the line is underfilling on average.

State. Let $\mu$ be the mean fill volume, in milliliters, of all bottles filled by this line during the production shift.

Plan. Translate the target and concern into claims about $\mu$. The stated target supplies the reference value, 500 mL. The concern is underfilling, so the question is about a mean below that target.

Do. The null hypothesis represents the target mean, and the alternative represents underfilling: $$ H_0:\mu=500\text{ mL} \qquad\text{and}\qquad H_a:\mu<500\text{ mL}. $$ The sample summaries, $\bar{x}=496.8$ mL and $s=6.4$ mL, are not inserted into these hypotheses. They describe the sample; the hypotheses describe the population mean.

Conclude. These hypotheses ask whether the mean fill volume for all bottles from this shift is below 500 mL. They do not, by themselves, establish that the line is underfilling. That conclusion would require carrying out and interpreting the appropriate test.

Worked Example: Checking a 60 g Package Claim

A fictional snack maker advertises that its packages contain an average of 60 g of product. A quality analyst randomly selects 25 packages from a production run and measures their contents. The sample mean is 59.4 g, and the sample standard deviation is 1.1 g. The analyst wants to know whether the production-run mean differs from the advertised average.

Define the parameter. Let $\mu$ be the mean product mass, in grams, in all packages from this production run.

Identify the reference value and question. The advertised average provides the reference value, 60 g. The question asks whether the mean differs from 60 g, without specifying only a lower or higher departure.

Write the hypotheses. $$ H_0:\mu=60\text{ g} \qquad\text{and}\qquad H_a:\mu\ne60\text{ g}. $$

Interpret the pair. The null describes a production-run mean of 60 g. The alternative describes a production-run mean that is not 60 g. The observed sample mean of 59.4 g is below 60 g, but that fact does not change the hypotheses: they are set by the stated claim and question, before assessing the sample evidence.

Worked Example: A Clinic’s Average Waiting-Time Goal

A fictional community clinic aims to keep its mean weekday walk-in waiting time at 20 minutes. The clinic selects 36 weekday visits at random from a defined month and records each wait. The sample mean is 18.7 minutes, and the sample standard deviation is 5.2 minutes. Administrators want to investigate whether the mean wait is now below the goal.

Define the parameter. Let $\mu$ be the mean waiting time, in minutes, for all weekday walk-in visits to this clinic during the defined month.

Translate the question. The reference value is 20 minutes. “Below the goal” asks whether the population mean is less than 20 minutes.

Write the hypotheses. $$ H_0:\mu=20\text{ minutes} \qquad\text{and}\qquad H_a:\mu<20\text{ minutes}. $$

Check the scope. The parameter refers to weekday walk-in visits during the defined month, not all patients at the clinic, visits in other months, or the mean wait among only the 36 sampled visits. The sample mean, 18.7 minutes, is a sample statistic; it is not the population parameter named in the hypotheses.

What the Hypotheses Do—and Do Not—Say

A hypothesis is a statement about a parameter, so it does not report a sample result. In the bottle example, $\mu$ is the unknown average for the defined population of bottles. The observed $\bar{x}$ is calculated from the 40 sampled bottles. These quantities are related, but they are not interchangeable: hypotheses are written about $\mu$, while $\bar{x}$ is evidence that will be considered in the test.

The hypotheses also need a clear population boundary. If a question concerns one production shift, define $\mu$ for that shift. Do not silently expand the claim to every bottle the company produces. Similarly, if a clinic’s data come from a particular month, define the population in a way that matches the intended claim.

Units help reveal whether the parameter has been defined correctly. A fill-volume mean should be measured in milliliters, a package-mass mean in grams, and a waiting-time mean in minutes. The reference value and parameter should refer to the same variable, population, and units. Comparing a mean measured in minutes with a reference value stated in hours without converting units would make the hypotheses inconsistent.

Key distinction: $H_0$ and $H_a$ are claims about the population mean $\mu$. The sample mean $\bar{x}$, sample standard deviation $s$, and sample size $n$ describe the sample and do not belong in the hypotheses.

Writing the hypotheses does not yet tell you whether the data provide convincing evidence for the alternative. It also does not tell you whether the conditions for the intended test are satisfied. As practiced in “Complete Conditions Check for a Mean Inference Problem,” condition checks concern how the data were collected, independence, and the shape or sample-size evidence. Those checks are separate from stating $H_0$ and $H_a$.

Common Mistakes and AP Exam Tips

Defining $\mu$ as the sample mean. “Let $\mu$ be the mean of the sample” confuses a parameter with a statistic. A full-credit definition identifies the mean for the population named in context.
Putting $\bar{x}$ in a hypothesis. A statement such as $H_0:\bar{x}=500$ is not a hypothesis about the population mean. Write $H_0:\mu=500$; use $\bar{x}$ only to describe the sample.
Leaving out equality in the null. For a one-sample mean test, write the null at the stated reference value, such as $H_0:\mu=500$, rather than using a directional inequality for the null.
Choosing the alternative from the sample result. A sample mean below the target does not automatically mean the alternative should be “less than.” Use the question’s stated concern or research question to identify the departure being investigated.
Giving symbols without context. The pair $H_0:\mu=20$, $H_a:\mu<20$ is incomplete if the reader cannot tell what $\mu$ measures or which population it refers to. Define $\mu$ in words and include units.
Changing the population while writing. If the study concerns one shift, month, school, or production run, do not define the parameter for a broader population unless the question supports that scope.

A strong AP response is brief but complete: define the population mean in context, state both hypotheses using that parameter, include the reference value and units, and make the alternative match the question. You do not need to calculate a test statistic or make a conclusion when the task asks only for hypotheses.

Key takeaway: Define $\mu$ as the population mean in context, write the null with equality at the stated reference value, and express the question of interest in the alternative. Keep the hypotheses about $\mu$, not the sample statistic $\bar{x}$.

Check Your Understanding

For each situation, define the population mean in context and write an appropriate pair of hypotheses.

A fictional orchard’s manager wants to know whether the mean mass of apples in this week’s harvest differs from 150 g. State the parameter and hypotheses.
A school is investigating whether the average time students spend traveling to school is more than 25 minutes. Define $\mu$ for the students of interest and write the hypotheses.
A fictional water-treatment team wants to investigate whether the mean processing time for a defined month is below its 12-minute target. State the parameter and hypotheses.
A sample of 32 devices has a mean battery life of 9.6 hours. The manufacturer’s stated average is 10 hours, and the question asks whether the population mean battery life is different from 10 hours. Write the hypotheses and explain why 9.6 does not replace the null value.
In your own words, explain why defining $\mu$ with a population and units makes a hypothesis pair clearer.