Common Mistakes Describing Scatterplots

Good Descriptions Do More Than Name a Direction

In “Effect of Axis Scale on Appearance,” you learned that display choices can change how a scatterplot looks without changing the data. Even when the display is fair, a description can still be misleading if it leaves out context, treats steepness as strength, or ignores an unusual point. These errors can turn a partly correct observation into an incomplete or inaccurate account of the data.

As in “Describing a Scatterplot With DUFS,” a careful description considers direction, unusual features, form, and strength. “Writing Descriptions in Context” also established that the variables, their units, and the individuals matter. This tutorial focuses on mistakes that can occur while using those ideas: not saying what the variables measure, using “strong” to mean “steep,” and describing only the majority pattern while overlooking a point that departs from it.

A description should report what the plot shows, not what you assume caused it. An association between two variables does not, by itself, show that changes in one variable cause changes in the other. For example, students who spend more time practicing might also differ in prior experience, access to help, or other factors that are not displayed.

Mistake 1: Leaving Out the Context

A statement such as “There is a moderate positive linear association” gives useful information about direction, form, and strength, but it could describe many different data sets. It does not tell the reader who or what was measured, what the variables mean, or how their values are expressed. On an exam, a general statistical description should be tied to the situation.

Check that your wording identifies the observational units, the explanatory variable, and the response variable. Then include units when the variables have them. If the explanatory variable increases, say what tends to happen to the response variable. Use cautious wording such as “tends to increase” rather than suggesting that every observation follows an exact rule.

Notice that the final description does not simply attach “tomato” to a generic statement. It names the plants, both variables, their units, and the trend. It also avoids claiming that each plant with more sunlight must have a higher yield.

Mistake 2: Confusing Strength With Steepness

Strength describes how closely the points follow the overall pattern. Steepness describes how much the response tends to change as the explanatory variable changes, relative to the units on the axes. A pattern can rise sharply and still have substantial scatter; another can rise more gradually while its points hug the pattern closely. These are different features of a scatterplot.

The apparent steepness also depends on the measurement units and the axes, as you learned in “Effect of Axis Scale on Appearance.” For example, measuring study time in minutes instead of hours changes the numerical amount of response per unit of explanatory variable. That unit change does not, by itself, make the association stronger or weaker. Check the scales before comparing steepness, and judge strength by the scatter around the pattern.

The endpoint calculations help illustrate the distinction, but they should not replace a visual description of the full pattern. A full-credit comparison tells the reader what is being compared, describes the amount of scatter, and does not use “steeper” as a synonym for “stronger.” If axis ranges or plot proportions differ, mention that the visual comparison of steepness is limited.

Mistake 3: Ignoring an Unusual Point

A description can be misleading if it summarizes the main cluster but says nothing about a point that lies far from the overall pattern. In “Spotting Outliers in Bivariate Data,” you learned that a bivariate outlier is a point far from the overall pattern, not merely a point with an unusually large or small value on one axis. When a plot has such a point, report it as part of the description.

Do not pretend the point is absent, and do not let it replace the description of the other observations. First describe the overall pattern, then locate the unusual observation using both variables and units. Unless there is a justified reason to correct or exclude a recorded value, a description should not silently remove it.

An unusual point can change how a reader sees the whole plot, so it deserves explicit attention. But “there is an outlier” is not a complete description either: say which observation appears unusual and how it departs from the pattern. Keep the overall trend and the exception distinct.

Other Wording Traps

Some errors are not about a particular feature of the plot. They arise when a description overstates what the display shows or uses vague language instead of statistical terms. Before finalizing a sentence, check whether each claim is supported by the points and whether a reader can tell what the variables mean.

• Calling an association causal: “More tutoring causes higher scores” goes beyond a scatterplot. Say that students with more tutoring hours tend to have higher scores, and do not claim cause and effect based only on the association.
• Using “strong” without describing scatter: Strength is about how closely points follow the overall pattern. A steep angle or a large change in response values does not establish strong association.
• Describing only direction: “The association is positive” leaves out form, strength, context, and unusual features. Add those details when they are visible and relevant.
• Calling every extreme coordinate an outlier: A point far to the right may simply have a large \(x\)-value while still following the pattern. Judge whether it is far from the overall relationship, as well as whether its individual coordinates are unusual.
• Overgeneralizing from a few points: Use “tends to” or “overall” when the pattern has variation. Do not say that every increase in \(x\) is matched by an increase in \(y\) unless the plotted values actually support that claim.
• Omitting units: “More time is associated with a higher result” is less informative than naming the type of time, the result, and their units.

Revise a Description Before You Submit It

A useful final check is to read your sentence as if the reader cannot see the plot. Would they know what each point represents? Could they tell which variable is explanatory and which is the response? Have you described the direction and form, judged strength by the scatter rather than the angle, and mentioned any clear unusual feature?

The aim is not to make every description long. It is to include the details needed for an accurate summary. A concise response can earn credit when it is specific, contextual, and cautious about what the data support.

Common Mistakes and AP Exam Tips

A response can use the right statistical vocabulary and still be incomplete. These checks help turn a quick visual impression into a description that communicates the plot accurately.

• Do not write only “positive and strong.” Full-credit communication identifies the individuals and variables in context, gives the units, and describes the overall form as well as direction and strength.
• Do not equate steepness with strength. Say how the response tends to change as the explanatory variable increases if you are describing steepness. Separately explain whether the points are close to or scattered around the pattern.
• Do not omit a clear exception. Describe the main pattern and then identify an unusual point by its approximate values and its position relative to that pattern.
• Do not claim that association proves cause. “The variables are associated” or “higher values of the explanatory variable tend to occur with higher response values” is appropriate. A causal explanation is not established by the scatterplot alone.
• Do not make a plot comparison without checking the axes. As in “Effect of Axis Scale on Appearance,” check units, ranges, and plot proportions before comparing apparent steepness or strength.

A strong AP response is not just a list of adjectives. For instance: “For these six students, tutoring hours per week and practice-test score in points show a positive, roughly linear pattern among five students, but the student with 6 tutoring hours and a score of 46 points lies well below it.” This names the context, describes the main pattern, and accounts for the unusual observation without asserting a cause.