Scatter Plots - PSAT Math | Varsity Tutors

SECTION 1

Historical Context & Motivation

Long before computers could crunch millions of data points in seconds, scientists and mathematicians needed a way to see relationships between two variables. If you measured the temperature outside every hour and also counted how many ice-cream cones a shop sold, how would you figure out whether those two quantities are connected? The answer that emerged over centuries is deceptively simple: plot one variable on a horizontal axis and the other on a vertical axis, then place a dot where each pair of values meets. The resulting picture — a scatter plot — can instantly reveal patterns that columns of raw numbers hide.

1686 • Early Coordinate Geometry

Building on René Descartes' coordinate system, Edmund Halley plotted barometric pressure against altitude, creating one of the earliest known scatter-style graphs to study atmospheric trends.

1833 • Statistical Graphics Emerge

English scientist John Herschel published a scatter plot comparing star brightness to orbital distance. This helped establish graphical methods as legitimate scientific tools.

1886 • Galton & Regression

Sir Francis Galton plotted the heights of parents versus children, discovering the concept of regression toward the mean and pioneering the idea of fitting a line to scattered data.

1900s • Pearson Formalizes Correlation

Karl Pearson developed the correlation coefficient (r), giving scatter plots a numerical companion that quantifies the strength and direction of a linear relationship.

Today • PSAT & Modern Data Literacy

Scatter plots appear frequently on the PSAT and SAT. The College Board expects students to read, interpret, and draw conclusions from scatter plots as part of Problem Solving & Data Analysis.

The central question scatter plots answer is both practical and powerful: Do two variables move together, and if so, how? Understanding this concept equips you not only for the PSAT but also for interpreting data in science classes, news articles, and everyday decision-making.

SECTION 2

Core Principles & Definitions

Before you can master scatter-plot questions on the PSAT, you need a clear vocabulary. A scatter plot is a graph of ordered pairs (x, y) drawn on a coordinate plane. Each dot represents one observation — one person, one trial, one measurement. The horizontal axis (x-axis) usually carries the independent variable (the factor you think might influence the other), and the vertical axis (y-axis) carries the dependent variable (the factor that responds). Understanding these roles is the first step in reading any scatter plot correctly.

Association (Correlation)

The overall pattern between x and y. An association can be positive (both increase), negative (one increases while the other decreases), or no association.

Line of Best Fit

A straight line drawn through the scatter plot that minimizes the overall distance between itself and every data point. On the PSAT, this line is often provided, and you interpret its slope and y-intercept.

Outlier

A data point that falls far away from the general pattern. Outliers can strongly influence the line of best fit and may signal unusual circumstances or measurement errors.

Strength of Association

How tightly the dots cluster around a line. A strong association shows dots close to the line; a weak association shows dots widely scattered.

Form of Association

The shape of the pattern — linear (dots follow a straight line) or nonlinear (dots follow a curve). The PSAT mostly tests linear associations.

SECTION 3

Visual Explanation — Anatomy of a Scatter Plot

The diagram below shows a scatter plot with all its key components labeled. Study the position of each element — the PSAT expects you to read scatter plots quickly and accurately.

This scatter plot shows hours of study on the x-axis and test score on the y-axis. Each cyan dot is one student. The dashed violet line is the line of best fit. The red dot at roughly (3, 13) is an outlier — a student who studied three hours but scored much lower than predicted.

Notice how the dots climb from the lower-left corner toward the upper-right corner. That upward slope tells you the association is positive: as study hours increase, test scores also tend to increase. The dots hug the dashed line fairly tightly, so the association is strong. If the dots were spread out in a wide cloud, we would call it weak. And because the overall path is straight rather than curved, the association is linear. On the PSAT, you will be asked to describe scatter plots using these three descriptors: direction (positive or negative), strength (strong or weak), and form (linear or nonlinear).

SECTION 4

Mathematical Framework — Line of Best Fit

When a scatter plot shows a linear association, we can describe the trend with an equation. The PSAT frequently provides the equation of the line of best fit (also called the least-squares regression line) and asks you to interpret its slope and y-intercept in context. You are not expected to calculate the line yourself; instead, you must understand what each part of the equation means.

LINE OF BEST FIT

ŷ = mx + b

ŷ = predicted value of the dependent variable; m = slope (change in y for each one-unit increase in x); x = value of the independent variable; b = y-intercept (predicted y when x = 0).

The slope is the single most-tested piece of the equation. It tells you the rate of change. For example, if the equation is ŷ = 6.5x + 30 where x is hours studied and ŷ is the predicted test score, the slope of 6.5 means: "For each additional hour of study, the predicted test score increases by 6.5 points." The PSAT loves asking you to put the slope into a real-world sentence just like that one.

SLOPE INTERPRETATION

m = (y₂ − y₁) / (x₂ − x₁)

Pick any two points on the line (not necessarily data points). The slope equals the vertical change divided by the horizontal change. A positive slope means y increases as x increases; a negative slope means y decreases as x increases.

PREDICTION WITH THE LINE

ŷ = 6.5(7) + 30 = 45.5 + 30 = 75.5

To predict a test score for a student who studies 7 hours, substitute x = 7 into the equation. The model predicts a score of approximately 75.5. Remember: this is a prediction, not a guarantee.

ℹ

PSAT Tip

SECTION 5

Types of Association — A Visual Guide

Not every scatter plot looks the same. The PSAT may show you a scatter plot and ask you to classify its association. The diagram below illustrates the four most common patterns you will encounter: positive linear, negative linear, nonlinear, and no association.

Four panels compare positive linear, negative linear, nonlinear, and no association patterns. The checklist below the panels summarizes the four questions to ask when classifying any scatter plot.

When you encounter a scatter plot on the PSAT, run through the four-step checklist shown in the diagram: direction, form, strength, and outliers. Practicing this mental routine will help you answer classification questions in under thirty seconds, which is valuable time saved for harder problems later in the section.

SECTION 6

Worked Example — Interpreting a Line of Best Fit

Let's work through a PSAT-style problem from start to finish. Suppose a scatter plot shows the relationship between the number of months a salesperson has been employed (x) and the number of sales they make per month (y). The line of best fit for the data is ŷ = 3.2x + 12.

Worked example

1
Step 1 — Identify the Variables

The independent variable (x) is the number of months employed. The dependent variable (y) is the number of sales per month. The equation ŷ = 3.2x + 12 models the predicted sales based on employment duration.
2
Step 2 — Interpret the Slope

The slope is 3.2. In context: for each additional month of employment, the model predicts that the salesperson makes approximately 3.2 more sales per month. Because the slope is positive, there is a positive association between months employed and sales.

Slope = 3.2 → predicted sales increase by 3.2 per additional month employed
3
Step 3 — Interpret the Y-Intercept

The y-intercept is 12. When x = 0 (a brand-new employee with zero months of experience), the model predicts they would make 12 sales in that first month. Always check whether the y-intercept makes sense in context — here it does, because a new salesperson can still make some sales.

Y-intercept = 12 → a new hire is predicted to make 12 sales in month 0
4
Step 4 — Make a Prediction

How many sales does the model predict for someone employed for 10 months? Substitute x = 10: ŷ = 3.2(10) + 12 = 32 + 12 = 44.

Predicted sales at 10 months = 44
5
Step 5 — Find a Residual

Suppose the actual data point for a 10-month employee is 48 sales. The residual is the difference between the actual value and the predicted value: residual = actual − predicted = 48 − 44 = 4. A positive residual means the actual value is above the line of best fit.

Residual = 48 − 44 = +4 (above the line)

SECTION 7

Strengths, Limitations & Common Pitfalls

Scatter plots are one of the most versatile tools in data analysis, but like any tool, they have limitations. Understanding where scatter plots shine and where they can mislead you will help you avoid common traps on the PSAT.

Strengths	Limitations	PSAT Pitfalls to Avoid
Instantly reveal the direction and strength of a relationship between two variables.	Only show two variables at a time; cannot display three or more dimensions easily.	Confusing association with causation. A scatter plot showing that ice-cream sales and sunburns both increase does not mean ice cream causes sunburns.
Make outliers easy to spot visually.	With many overlapping data points, individual values become hard to distinguish.	Extrapolating far beyond the data. If x ranges from 1 to 10, predicting y at x = 50 is unreliable.
A line of best fit provides a simple predictive model.	A linear model is inappropriate if the actual relationship is curved.	Misidentifying a nonlinear pattern as linear. Always check whether the scatter curves before choosing a linear answer.
Can display real-world data directly — each dot is a real observation.	Sensitive to scale choices; stretching an axis can exaggerate or flatten a trend.	Reading the wrong axis. Double-check which variable is on which axis before answering.

Strengths, limitations, and common PSAT pitfalls related to scatter plots

SECTION 8

Connection to Advanced Concepts

The scatter plot skills you build for the PSAT form the foundation for more advanced statistical topics you will encounter in AP Statistics, college courses, and data-driven careers. The table below maps each PSAT-level concept to its advanced counterpart so you can see the bigger picture.

PSAT Concept	Advanced Version	What Changes
Describing association as positive/negative	Pearson correlation coefficient (r)	The direction and strength are captured by a single number between −1 and +1.
Line of best fit (ŷ = mx + b)	Least-squares regression analysis	You learn how to calculate m and b by minimizing the sum of squared residuals.
Residual = actual − predicted	Residual plots and diagnostic checks	You plot residuals to check whether a linear model is appropriate — random scatter in residuals confirms linearity.
Nonlinear patterns (curve)	Exponential and polynomial regression	When the data curves, you use logarithmic or power transformations to straighten the scatter before fitting a line.
Association ≠ causation	Confounding variables and experimental design	In AP Statistics, you learn to identify lurking variables and design experiments that can establish causation.

PSAT scatter plot skills mapped to their advanced statistical counterparts

You do not need to master any of these advanced topics for the PSAT, but knowing they exist can motivate you: every scatter-plot question you practice now is building intuition for powerful tools used in medicine, economics, engineering, and virtually every field that relies on data.

SECTION 9

Practice Problems

Work through the following five problems in order. They increase in difficulty, mirroring the range you might see on the actual PSAT. Write out your reasoning before checking the answers.

Question 1: CONCEPTUAL

A scatter plot displays the relationship between daily high temperature (°F) and the number of hot chocolates sold at a café. The dots trend from the upper left to the lower right of the scatter plot. What type of association does the scatter plot show between daily high temperature and hot chocolate sales?

A. Positive association
B. Negative association
C. No association
D. Cannot be determined

Show answer

The correct answer is B) Negative. Step-by-step reasoning: The dots trend from the upper left to the lower right, meaning that as the x-variable (daily high temperature) increases, the y-variable (hot chocolate sales) decreases. A line of best fit drawn through these points would have a negative slope, indicating a negative association. This also makes intuitive sense — on hotter days, fewer people want a warm drink. Why the other choices are wrong: A) Positive association would mean both variables increase together, which would show dots trending from lower left to upper right — the opposite of what is described. C) No association would mean the dots are scattered randomly with no clear pattern, but a clear upper-left to lower-right trend is described. D) Cannot be determined is incorrect because the described trend provides sufficient information to identify the direction of the association.

Question 2: BASIC CALCULATION

The line of best fit for a scatter plot is ŷ = −0.8x + 50, where x is the number of absences during a semester and ŷ is the predicted final exam score. What is the predicted final exam score for a student who was absent 10 times?

A. 34
B. 38
C. 42
D. 46

Show answer

Substitute x = 10 into the equation: ŷ = −0.8(10) + 50 = −8 + 50 = 42. The correct answer is C. The model predicts a final exam score of 42. The slope of −0.8 tells us that each additional absence is associated with a predicted decrease of 0.8 points on the exam. The y-intercept of 50 represents the predicted score for a student with zero absences. Choice A (34) results from incorrectly multiplying −0.8 × 10 = −16 instead of −8, then subtracting: 50 − 16 = 34. Choice B (38) results from a partial arithmetic error, subtracting 12 instead of 8: 50 − 12 = 38. Choice D (46) results from using the wrong sign, adding instead of subtracting: 50 − 4 = 46, or misreading the slope as −0.4 and computing −0.4(10) + 50 = 46.

Question 3: INTERMEDIATE

A line of best fit for a data set is given by ŷ = −0.8x + 50, where x is the number of absences and ŷ is the predicted final exam score. A student with 15 absences actually scored 30 on the final exam. What is the residual for this student's data point?

A. −8
B. 8
C. −12
D. 12

Show answer

First, find the predicted score: ŷ = −0.8(15) + 50 = −12 + 50 = 38. The residual equals actual − predicted = 30 − 38 = −8. A negative residual means the student scored 8 points below the line of best fit's prediction. On a scatter plot, this data point would appear below the trend line.

Question 4: APPLIED

A researcher collects data on the age of used cars (in years) and their resale prices (in thousands of dollars) for cars between 1 and 12 years old. The scatter plot shows a strong negative linear association, and the line of best fit is ŷ = −1.5x + 25. A friend says, "According to this model, a 20-year-old car would be worth −5 thousand dollars, so the dealer would pay you to take it!" Which of the following best explains the flaw in the friend's reasoning?

A. The slope of the line of best fit should be positive, not negative, for used car data.
B. The model is being used to extrapolate far beyond the range of the data, making the prediction unreliable.
C. The y-intercept of 25 was calculated incorrectly and causes the model to produce negative values.
D. The association between age and resale price is not strong enough to make any predictions.

Show answer

The correct answer is B. Step-by-step: The line of best fit is ŷ = −1.5x + 25, built from data on cars aged roughly 1 to 12 years. Plugging in x = 20 gives ŷ = −1.5(20) + 25 = −30 + 25 = −5. While the arithmetic is correct, x = 20 is far outside the range of the original data. Using a model to predict values far beyond the observed data range is called extrapolation, and it is unreliable because the linear trend may not continue. In reality, a car's resale value approaches zero but cannot be negative. Choice A is wrong because a negative slope correctly reflects that older cars are worth less. Choice C is wrong because the y-intercept of 25 is part of the given model and is not the source of the error. Choice D is wrong because the problem states the association is strong, so strength of association is not the issue.

Question 5: CRITICAL THINKING

A student models the same data set two ways. In the first model, a line of best fit is used, and its residual plot shows a clear U-shaped pattern — residuals are negative, then positive, then negative as x increases. In the second model, a curve of best fit is used, and its residual plot shows points scattered randomly above and below zero with no clear pattern. Which conclusion is best supported by these residual plots?

A. The linear model is more appropriate because it produces a recognizable U-shaped pattern in the residuals, confirming a consistent fit.
B. The curved model is more appropriate because randomly scattered residuals indicate the model captures the data's pattern well, leaving only random error.
C. The linear model is more appropriate because simpler models are always preferred regardless of the residual pattern.
D. Neither model is appropriate because an ideal model would produce residuals that are all equal to zero.

Show answer