Positive and Negative Slopes in Context

The Sign of a Slope Describes a Direction

In “Slope Units and Rates of Change,” you learned that the slope $b$ in $\hat{y}=a+bx$ describes the change in the line’s predicted response for a one-unit increase in the predictor. The sign of $b$ tells you whether that predicted response goes up or down. It also gives a direction to check against the scatterplot.

As in “Positive Versus Negative Association Examples,” think about what happens to the response as the explanatory variable increases. A positive slope describes a line that rises from left to right: larger $x$-values go with larger predicted $y$-values. A negative slope describes a line that falls from left to right: larger $x$-values go with smaller predicted $y$-values.

Definition: A positive regression slope means the line’s predicted response increases as the predictor increases. A negative regression slope means the line’s predicted response decreases as the predictor increases.

The sign describes the direction of the fitted line, not the strength of the association. A slope of $-5$ is not automatically a stronger relationship than a slope of $-1$; their numerical sizes depend on the variables’ units. As discussed in “Slope Units and Rates of Change,” interpret the rate with its response units per predictor unit.

The sign also does not describe every individual observation. A regression line summarizes a pattern, and individual points can lie above or below it. Its slope does not guarantee that every case follows the predicted change exactly, nor does an association by itself show that changing the predictor causes the response to change.

Interpreting a Negative Slope: Car Age and Price

Consider a fictional sample of used cars. Let $x$ be a car’s age in years and $y$ be its sale price in thousands of dollars. A fitted line is:

\hat{y}=31.8-2.45x

The slope is $-2.45$ thousand dollars per year. Because the slope is negative, the line predicts lower prices for older cars. To interpret it in context, name the predictor, its unit, the response, and the response units. Since $2.45$ thousand dollars is $2,450, a suitable interpretation is: For each additional year of age, the line predicts that a used car’s sale price will be $2,450 lower.

The word “predicts” matters. This describes the fitted line, not a promise about an individual car. Two cars of the same age can have different prices, and a particular older car might sell for more than a particular younger one. Condition, mileage, make, and other factors can matter. As covered in “Why Correlation Does Not Imply Causation,” the negative association alone does not establish that age causes a particular price change.

A downward pattern in the scatterplot is consistent with this negative slope: moving from left to right means looking at older cars, and the overall pattern tends toward lower prices. The sign check is about the direction of the overall linear pattern, not whether every point descends as you move right.

Worked Example: Car Age and Predicted Price

For the fictional used-car model above, compare the predicted prices for cars that are 2 and 6 years old. Use the predictions to verify the meaning of the slope.

For a 2-year-old car, substitute $x=2$ into the line:

\hat{y}=31.8-2.45(2)=31.8-4.9=26.9

The predicted price is $26.9$ thousand dollars, or $26,900. For a 6-year-old car:

\hat{y}=31.8-2.45(6)=31.8-14.7=17.1

The predicted price is $17.1$ thousand dollars, or $17,100. The car age increased by $6-2=4$ years, while the predicted price changed by $17.1-26.9=-9.8$ thousand dollars. The rate across those four years is:

\frac{-9.8\text{ thousand dollars}}{4\text{ years}} =-2.45\text{ thousand dollars per year}

This agrees with the slope. In context, the line predicts a $9,800 lower price for the 6-year-old car than for the 2-year-old car, a difference of $2,450 per additional year of age according to the model. These are predicted prices; actual sale prices may differ.

Interpreting a Positive Slope

A positive slope uses the same interpretation structure, but the predicted response increases as the predictor increases. For example, a fictional community garden records plant height and the number of weeks since seedlings were planted. If a fitted line has a positive slope, it predicts that plants tend to be taller at later weeks. The slope’s number and units tell how much taller the line predicts them to be for each additional week.

The scatterplot check works in the same way. With the predictor on the horizontal axis and the response on the vertical axis, a positive overall linear pattern rises from left to right. Check that the variables are assigned to the intended axes before deciding whether the direction agrees. As you practiced in “Writing Descriptions in Context,” describe the pattern using the actual variables and units rather than only saying “it goes up.”

Worked Example: Practice Time and Successful Shots

A fictional sports program models the number of successful shots out of 50, $y$, from weekly practice time $x$, in hours. The fitted line is:

\hat{y}=18+3.2x

Interpret the slope and check its direction using predictions for 2 and 7 hours of practice.

The slope is $3.2$ successful shots per practice hour, and it is positive. Thus, for each additional hour of weekly practice, the line predicts 3.2 more successful shots out of 50. Calculate the two predictions:

\hat{y}(2)=18+3.2(2)=24.4\text{ successful shots}

\hat{y}(7)=18+3.2(7)=40.4\text{ successful shots}

Practice time increased by $7-2=5$ hours, and the predicted number of successful shots increased by $40.4-24.4=16$. The rate is:

\frac{16\text{ successful shots}}{5\text{ hours}} =3.2\text{ successful shots per hour}

The positive slope and the increasing predictions agree: the predicted response rises as practice time rises. A context-appropriate interpretation is: For each additional hour of weekly practice, the line predicts 3.2 more successful shots out of 50 for participants represented by this model. This does not mean every participant gains exactly 3.2 shots, or that practice time alone explains differences in performance.

The slope does not have to be a whole number even though the observed shot counts are whole numbers. A regression line can give fractional predicted values; these are model predictions, not necessarily actual counts for an individual.

Check the Slope Sign Against the Scatterplot

For a clear overall linear pattern, the slope’s sign should agree with the scatterplot’s direction when you read the horizontal axis from smaller $x$ to larger $x$. If the pattern descends from left to right, a negative slope is consistent. If it rises, a positive slope is consistent. This is a direction check, not a substitute for examining form, strength, and unusual features as in the earlier tutorial on describing a scatterplot with DUFS.

A disagreement is a reason to pause and investigate rather than to change a sign by instinct. Confirm that the graph and equation refer to the same variables and observations. Check which variable is on each axis, whether the equation has been copied correctly, and whether the horizontal scale has been read in the intended direction. A reversed or mislabeled axis can make a correct equation appear inconsistent.

Also consider whether the plot has a curved pattern, an unusual influential observation, or subgroups. As discussed in “How Outliers Change the Correlation” and “Lurking Patterns: Time and Subgroups,” unusual points and group patterns can complicate what the full set of points appears to show. A visual impression can be misleading; check that you are describing the same overall linear pattern the fitted line summarizes. If the mismatch remains, do not claim that the line matches the plot without resolving it.

Worked Example: Investigating a Sign Mismatch

A fictional building-energy analysis uses outdoor temperature $x$, in degrees Fahrenheit, to predict daily heating energy use $\hat{y}$, in kilowatt-hours. The scatterplot shows a clear downward linear pattern: days with higher temperatures tend to have lower heating energy use. A report gives this equation:

\hat{y}=88+0.8x

Does the equation agree with the scatterplot? Use two temperatures to check.

The reported slope is $+0.8$ kilowatt-hours per degree Fahrenheit. Its positive sign says that the line predicts higher heating energy use as outdoor temperature increases. For temperatures of $40^\circ\text{F}$ and $60^\circ\text{F}$, the reported equation gives:

\hat{y}(40)=88+0.8(40)=120\text{ kilowatt-hours}

\hat{y}(60)=88+0.8(60)=136\text{ kilowatt-hours}

The reported line predicts an increase of $136-120=16$ kilowatt-hours as temperature rises by $60-40=20^\circ\text{F}$. Its rate is $16/20=0.8$ kilowatt-hours per degree Fahrenheit. That is an upward direction, so it conflicts with the described downward scatterplot.

One possible transcription error is the sign of the slope. If the intended equation were $\hat{y}=88-0.8x$, then:

\hat{y}(40)=88-0.8(40)=56\text{ kilowatt-hours}

\hat{y}(60)=88-0.8(60)=40\text{ kilowatt-hours}

That corrected line predicts a decrease of $40-56=-16$ kilowatt-hours over the $20^\circ\text{F}$ increase, agreeing in direction with the plot. This does not prove that the corrected equation is the actual fitted line; the data and model output would need to be checked. The correct conclusion from the information given is that the reported positive slope and the described downward pattern are inconsistent.

Common Mistakes and AP Exam Tips

Reversing the direction. A negative slope means predicted $y$ decreases as $x$ increases. It does not mean both variables decrease together. State which variable is the predictor and which is the response.
Leaving out context and units. “The slope is $-2.45$” is not a complete interpretation. Say that for each additional year of car age, the line predicts a $2,450 lower sale price.
Interpreting the sign as strength. The sign gives direction. The slope’s numerical size depends on measurement units, so do not use it alone to compare how strong two associations are.
Claiming every individual follows the line. A fitted slope describes the change in predicted response. It does not guarantee that every older car sells for less than every younger car.
Ignoring an apparent mismatch. If a downward plot is paired with a positive slope, check the axes, variables, data, and equation. Do not simply repeat both descriptions as if they agree.
Turning association into causation. A negative slope for car age and price describes an observed pattern in the data. It does not show, by itself, that age causes the price difference.

For full-credit communication, name both variables, use the slope’s sign to state the predicted direction, include the amount and units of the rate, and refer to a one-unit increase in the predictor. When asked to check a plot, explain whether its overall direction agrees with the slope; if not, identify the mismatch and what should be checked.

Key takeaway: A positive slope means the line’s predicted response rises as the predictor increases; a negative slope means it falls. Check that this direction agrees with the scatterplot’s overall linear pattern, using the correct variables and axes.

Check Your Understanding

For each question, focus on the predictor, response, slope sign, and direction in context.

A line predicts resale price in dollars from the age of a bicycle in years and has slope $-35$. What does the sign tell you about the line’s predicted response as age increases?
A model predicts weekly water use in liters from the number of people in a household. Its slope is positive. Describe the direction of the predicted relationship in context.
A scatterplot of car age on the horizontal axis and sale price on the vertical axis has a downward linear pattern. Is a negative or positive slope consistent with that direction?
A report gives a positive slope for a plot that appears to rise from left to right. Name two things you could check before deciding whether the equation and plot agree.
Why is “each additional year makes every car’s price decrease” too strong an interpretation of a negative slope in a car-age model?