Setting up a $t$ test for the slope of a regression model means writing clear null and alternative hypotheses about the population slope $\beta$ , checking the four conditions, and identifying that you need a $t$ test for a slope. Most often the null hypothesis is that the slope equals 0, which is the same as saying there is no linear relationship between $x$ and $y$ .

Why This Matters for the AP Statistics Exam

This topic is the setup half of slope inference, and the AP Statistics exam expects you to do this setup correctly before you ever calculate a test statistic. On free-response questions you may be asked to state hypotheses, name the procedure, and verify conditions using a residual plot or computer output. Doing this cleanly sets you up to carry out the test in the next topic and to interpret results without losing easy points. Slope inference falls in a unit that carries a small but real share of the exam, so a solid setup is worth practicing.

more resources to help you study

practice multiple choice FRQ practice & scoring cheatsheets score calculator key terms

Key Takeaways

The correct procedure for testing a regression slope is a t-test for a slope.
The null hypothesis is H0: beta = beta0, and the alternative is Ha: beta < beta0, beta > beta0, or beta != beta0.
Most slope tests use beta0 = 0, which tests whether a linear relationship exists at all.
You must check four conditions: linearity, equal standard deviation of y across x, independence, and approximately normal responses at each x.
Use residual plots and graphs of residuals to support your linearity, equal-variance, and normality checks.
State hypotheses, name the test, and verify all conditions explicitly before calculating anything.

Choosing the Right Test

For inference on a regression slope, the procedure is a t-test for a slope. You use a t-distribution because you are working with quantitative data and you almost never know the true standard deviation of y in the population, so you estimate it from the sample.

A t-test for the slope checks whether the slope is meaningfully different from a hypothesized value. The slope describes the relationship between the explanatory variable (x) and the response variable (y). If the test gives strong evidence that the slope is different from 0, that supports a linear relationship between the two variables. If the evidence is weak, you do not have enough support to say the slope differs from 0.

On many graphing calculators this procedure is listed as LinRegTTest under the STAT > TESTS menu.

Hypotheses

Write your null and alternative hypotheses before doing anything else. For a slope test they look like this:

H0: beta = beta0
Ha: beta < beta0, beta > beta0, or beta != beta0

Here beta is the true slope of the population regression line, and beta0 is the hypothesized value stated in the null hypothesis.

Often the value you care about is 0, because a slope of 0 means x and y have no linear relationship. So a very common pair of hypotheses is:

H0: beta = 0
Ha: beta != 0

Example

Suppose a researcher claims that the slope relating jelly beans eaten per day (x) to pieces of Easter grass on the floor (y) is 40, meaning each additional jelly bean per day is associated with a predicted increase of 40 pieces of grass. To test that specific claim, you would write:

H0: beta = 40
Ha: beta != 40

Pick the alternative (<, >, or !=) based on what the question asks. Use a one-sided alternative only when the question points clearly in one direction.

Conditions

Like other inference procedures, the slope t-test has conditions you must check and state explicitly. There are four:

Linearity: The true relationship between x and y is linear. A residual plot with no clear pattern supports this.
Equal standard deviation: The standard deviation of y does not change as x changes. Look for no "fanning" (no widening or narrowing) in the residual plot.
Independence:
- Data come from a random sample or a randomized experiment.
- When sampling without replacement, check the 10% condition: n is no more than 10% of the population (n <= 0.10N).
Normality: For any particular value of x, the y-responses are approximately normally distributed. Check graphs of the residuals. If the data are skewed, you want a sample size greater than 30. If n is less than 30, the sample data should be free of strong skewness and outliers.

State each condition in context before you move on to calculations.

Scatterplots and residual analysis example

Source: University of Florida

How to Use This on the AP Statistics Exam

Free Response

State hypotheses with correct symbols. Define beta as the slope of the population regression line, and say what beta0 represents in context.
Name the procedure clearly as a t-test for the slope of a regression line.
Verify all four conditions, and connect each one to evidence such as a residual plot, a graph of residuals, or how the data were collected.
Avoid deterministic wording. Say "a predicted increase" rather than implying x forces a change in y.

Common Trap

Do not skip the condition checks. Stating hypotheses and naming the test is not enough; you must verify linearity, equal variance, independence, and normality.
Do not assume beta0 = 0 every time. Read the claim. If a specific slope value is given, beta0 is that value.
Match the alternative hypothesis to the question. Choose <, >, or != based on what is actually being asked, not by habit.

Common Misconceptions

"Failing to reject means there is no relationship." Not finding enough evidence that the slope differs from 0 does not prove the slope is exactly 0 or that the variables are unrelated. You just lack strong evidence for a linear relationship.
"The slope test always uses beta0 = 0." Zero is common, but the null value is whatever the claim states. A claim of slope 40 makes beta0 = 40.
"Residual plots are optional." The linearity and equal-variance conditions rely on residual analysis. Skipping the plot means you cannot fully justify the conditions.
"A significant slope proves causation." Significance supports a linear association. Causation depends on study design, such as a randomized experiment, not on the test alone.
"You can use a z procedure here." Because you estimate the standard deviation of y from the sample, slope inference uses a t-distribution, not z.

Vocabulary

The following words are mentioned explicitly in the College Board Course and Exam Description for this topic.

Term	Definition
alternative hypothesis	The claim that contradicts the null hypothesis, representing what the researcher is trying to find evidence for.
independence	The condition that observations in a sample are not influenced by each other, typically ensured through random sampling or randomized experiments.
linear relationship	A relationship between two variables that can be described by a straight line.
normal distribution	A probability distribution that is mound-shaped and symmetric, characterized by a population mean (μ) and population standard deviation (σ).
null hypothesis	The initial claim or assumption being tested in a hypothesis test, typically stating that there is no effect or no difference.
outlier	Data points that are unusually small or large relative to the rest of the data.
random sample	A sample selected from a population in such a way that every member has an equal chance of being chosen, reducing bias and allowing for valid statistical inference.
randomized experiment	A study design where subjects are randomly assigned to treatment groups to establish cause-and-effect relationships.
regression model	A statistical model that describes the relationship between a response variable (y) and one or more explanatory variables (x).
residual	The difference between the actual observed value and the predicted value in a regression model, calculated as residual = y - ŷ.
sampling without replacement	A sampling method in which an item selected from a population cannot be selected again in subsequent draws.
significance test	A statistical procedure used to determine whether there is sufficient evidence to reject the null hypothesis based on sample data.
skewness	A measure of the asymmetry of a distribution, indicating whether data is concentrated more on one side of the center.
slope	The value b in the regression equation ŷ = a + bx, representing the rate of change in the predicted response for each unit increase in the explanatory variable.
slope of a regression model	The coefficient that represents the rate of change in the predicted response variable for each unit increase in the explanatory variable in a linear regression equation.
standard deviation	A measure of how spread out data values are from the mean, represented by σ in the context of a population.
t-test for a slope	A hypothesis test used to determine whether the slope of a regression model is significantly different from zero, assessing whether there is a statistically significant linear relationship between variables.

Frequently Asked Questions

What test do you use for the slope of a regression model?

Use a t-test for the slope of a regression model. On many calculators, this procedure appears as LinRegTTest.

What are the hypotheses for a t-test for slope?

The null hypothesis is H0: beta = beta0. The alternative is Ha: beta < beta0, beta > beta0, or beta != beta0, depending on the question. Most AP Stats slope tests use beta0 = 0.

What does beta mean in a regression slope test?

Beta is the true slope of the population regression line. A sample slope estimates beta, but the hypothesis test is about the population slope.

What are the conditions for a t-test for slope?

Check linearity, equal standard deviation of y across x, independence, and approximate normality of responses at each x. Residual plots and sample context are usually the evidence you use.

How do residual plots help check conditions for a slope test?

Residual plots help you check whether a linear model is reasonable and whether the spread of residuals is roughly constant. A random scatter with no clear curve or fanning supports the conditions.

What is a common AP Stats mistake when setting up a slope test?

A common mistake is writing hypotheses about the sample slope instead of beta, the population slope. Another is listing conditions without connecting them to the graph, residual plot, or sampling context given in the problem.