Scatter Plots - SAT Math | Varsity Tutors

SECTION 1

Historical Context & Motivation

Humans have always wanted to see patterns in data, but for most of history, raw tables of numbers were the only option. The scatter plot changed that by giving researchers — and now SAT test-takers — a powerful visual tool for spotting relationships between two variables at a glance. Understanding where this idea came from helps reveal why the graph works the way it does.

1686 • Early Coordinate Graphing

Building on René Descartes' coordinate system, English astronomer Edmond Halley plotted barometric pressure against altitude — one of the first known uses of plotting paired data on axes.

1833 • Statistical Graphics Emerge

John Herschel published scatter-style plots relating orbital data of double stars, demonstrating that visual patterns in plotted points could reveal scientific relationships.

1886 • Galton & the Line of Best Fit

Sir Francis Galton introduced the concept of regression while studying the heights of parents and children. His scatter plots and fitted lines became the foundation of modern correlation analysis.

1900s • Pearson Formalizes Correlation

Karl Pearson developed the correlation coefficient (r), giving mathematicians a precise numerical way to quantify the strength and direction of trends visible in scatter plots.

2005–Present • Scatter Plots on the SAT

The redesigned SAT emphasizes real-world data literacy. Scatter plots now appear regularly in the Problem Solving & Data Analysis domain, testing students' ability to interpret trends, identify outliers, and evaluate lines of best fit.

The central question scatter plots address is straightforward: when two quantities change together, how can we visualize and describe their relationship? Whether you're comparing study hours to test scores or temperature to ice cream sales, the scatter plot turns abstract numbers into a picture you can analyze in seconds.

SECTION 2

Core Principles & Definitions

Before you can tackle SAT scatter plot questions confidently, you need to lock down a handful of foundational ideas. Every scatter plot question on the test relies on one or more of these core principles, so understanding them thoroughly will pay off across many problems.

Bivariate Data

A scatter plot displays bivariate data — each data point represents two measurements collected from the same individual or observation. The horizontal axis shows the independent variable (the input), and the vertical axis shows the dependent variable (the outcome).

Correlation & Direction

The overall trend of the points reveals the correlation. A positive correlation means both variables increase together. A negative correlation means one increases while the other decreases.

Strength of Association

If data points cluster tightly around an imaginary trend line, the association is strong. If they spread widely, the association is weak. If the points form a random cloud with no pattern, there is no correlation.

Line of Best Fit

A line of best fit (also called a trend line or regression line) is the straight line that best represents the overall pattern. It minimizes the total distance between itself and all the data points.

Outliers

An outlier is a data point that falls far from the overall trend. Outliers can pull a line of best fit toward them, potentially distorting the model. The SAT may ask you to identify outliers or describe their effect.

SECTION 3

Visual Explanation — Reading a Scatter Plot

The diagram below shows a typical scatter plot you might encounter on the SAT. It displays the relationship between hours spent studying per week and exam scores for a group of 15 students. Notice how the data points generally rise from left to right, suggesting a positive correlation. The dashed line of best fit captures the overall upward trend, and one point sits noticeably far from the line — that's the outlier.

Each cyan dot represents one student's data pair (hours, score). The dashed purple line is the line of best fit. The red dot near (4, 53) is an outlier — a student who studied 4 hours but scored unusually low compared to the trend.

When reading the diagram, start by identifying the axes and their scales. The horizontal axis represents the independent variable (hours studied), while the vertical axis represents the dependent variable (exam score). The upward slope of the data cloud and the line of best fit confirm a strong positive correlation. On the SAT, you might be asked to describe this trend, predict a score for a given number of study hours, or explain why the outlier doesn't fit the pattern.

SECTION 4

Mathematical Framework

The SAT doesn't require you to calculate a regression line from scratch, but you do need to understand the equation of the line of best fit and use it to make predictions. Here are the key equations and what each part means.

LINE OF BEST FIT (SLOPE-INTERCEPT FORM)

ŷ = mx + b

ŷ = predicted value of the dependent variable, m = slope (rate of change per unit increase in x), x = value of the independent variable, b = y-intercept (the predicted y-value when x = 0).

SLOPE

m = (y₂ − y₁) / (x₂ − x₁)

The slope describes the direction and steepness of the trend. A positive slope means y increases as x increases. A negative slope means y decreases as x increases. On the SAT, slope is often interpreted in context: for example, 'for each additional hour studied, the predicted exam score increases by m points.'

RESIDUAL

Residual = Actual y − Predicted ŷ

A residual measures how far a single data point falls above or below the line of best fit. A positive residual means the actual value is above the line; a negative residual means it is below. Large residuals often indicate outliers.

ℹ

SAT Tip: Interpreting Slope & Intercept in Context

SECTION 5

Types of Association & Correlation Patterns

Not every scatter plot shows the same kind of relationship. The SAT may present you with data that trends upward, downward, curves, or shows no pattern at all. Being able to classify the association quickly is essential. The diagram below illustrates the four most common patterns you'll encounter.

Four typical scatter plot patterns. The strong positive and strong negative patterns are linear. The curved pattern suggests a nonlinear model, and the random scatter indicates no association.

Pattern	Direction	Shape	Best Model
Strong Positive	Upward (↗)	Linear	ŷ = mx + b (m > 0)
Strong Negative	Downward (↘)	Linear	ŷ = mx + b (m < 0)
Nonlinear	Varies	Curved	Exponential or quadratic
No Correlation	None	Random cloud	No model fits

Summary of common scatter plot association types

ℹ

Correlation ≠ Causation

SECTION 6

Worked Example

Let's walk through an SAT-style scatter plot problem step by step. Suppose a scatter plot shows the relationship between the number of hours a group of employees commuted per week (x) and the number of sick days they took per year (y). The line of best fit is given as ŷ = 1.5x + 2.

Worked example

1
Step 1 — Understand the Context

The independent variable (x) is weekly commute hours, and the dependent variable (y) is annual sick days. The equation ŷ = 1.5x + 2 models the trend. We're asked: (a) What does the slope 1.5 mean in context? (b) Predict the number of sick days for an employee who commutes 6 hours per week. (c) If one employee commutes 6 hours and actually took 14 sick days, what is the residual?
2
Step 2 — Interpret the Slope

The slope m = 1.5 tells us the rate of change. For each additional hour of commuting per week, the model predicts an increase of 1.5 sick days per year. Notice we say 'is associated with' or 'the model predicts,' not 'causes.' This language matters on the SAT.

Slope interpretation: each additional commute hour per week is associated with 1.5 more sick days per year.
3
Step 3 — Make a Prediction

Substitute x = 6 into the equation: ŷ = 1.5(6) + 2 = 9 + 2 = 11. The model predicts that an employee commuting 6 hours per week will take approximately 11 sick days per year.

ŷ = 1.5(6) + 2 = 11 sick days
4
Step 4 — Calculate the Residual

The residual equals the actual value minus the predicted value. The employee actually took 14 sick days, but the model predicted 11. Residual = 14 − 11 = 3. A positive residual means the actual value is above the line of best fit — this employee took more sick days than the model predicted.

Residual = 14 − 11 = +3 sick days

★

Key takeaway

STRATEGY RECAP

SECTION 7

Common Pitfalls & SAT Traps

The College Board designs scatter plot questions to test whether you truly understand the data or are just picking an answer that sounds right. Here are the most common mistakes students make and how to avoid them.

Common Mistake	Why It's Wrong	Correct Approach
Claiming causation from correlation	A scatter plot shows association only — there may be lurking variables	Use 'associated with' or 'predicts' language; only experiments can prove causation
Confusing slope with y-intercept	Slope describes rate of change; y-intercept describes the starting prediction	Slope answers 'how much does y change per unit of x?' and y-intercept answers 'what is ŷ when x = 0?'
Extrapolating far beyond the data	The model is only reliable within the range of the given data	Only predict within or very close to the observed x-values; flag any extreme extrapolation
Ignoring outliers when describing the trend	Outliers can distort your impression of the overall pattern	Describe the trend based on the majority of points; note outliers separately
Forcing a linear model on curved data	A straight line poorly represents data that curves	If the points curve, recognize the association is nonlinear and a different model (exponential, quadratic) is more appropriate

Five common scatter plot errors on the SAT

★

Key takeaway

REMEMBER

SECTION 8

Connection to Advanced Topics

The scatter plot concepts you learn for the SAT form the foundation of much deeper statistical analysis. As you progress into AP Statistics or college-level coursework, you'll encounter more sophisticated tools that build directly on what you already know.

SAT Level Concept	Advanced Extension	What Changes
Line of best fit (visual)	Least-squares regression	You'll compute the exact line that minimizes the sum of squared residuals
Describing correlation as 'strong' or 'weak'	Correlation coefficient (r)	You'll calculate r, a number from −1 to 1 that precisely quantifies strength and direction
Residuals	Residual plots	You'll plot all residuals to check whether a linear model is truly appropriate
Two-variable scatter plot	Multiple regression	You'll model y using two or more independent variables simultaneously
Nonlinear pattern recognition	Transformations & curve fitting	You'll apply logarithmic or power transformations to linearize curved data

How SAT scatter plot skills extend into advanced statistics

The good news is that mastering scatter plots at the SAT level gives you a strong conceptual foundation. When you eventually encounter the correlation coefficient r or coefficient of determination r², you'll already understand what they're measuring because you've been visually estimating correlation strength on scatter plots. Similarly, the residual calculation you practice now is the exact same building block used to derive the least-squares regression line in AP Statistics.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A scatter plot shows the relationship between the number of hours of sleep students get per night and their GPA. The data points generally trend upward from left to right. Which of the following best describes the relationship shown?

A. There is a negative linear association between sleep and GPA.
B. There is a positive linear association between sleep and GPA.
C. More sleep causes higher GPAs.
D. There is no association between sleep and GPA.

Show answer

The correct answer is (B). Because the points trend upward from left to right, the scatter plot displays a positive association — as sleep hours increase, GPA tends to increase as well. Choice (A) is wrong because the direction is upward, not downward. Choice (C) is a classic SAT trap: the scatter plot shows association, not causation. We cannot conclude that sleep causes higher GPAs from observational data alone. Choice (D) contradicts the upward trend described in the problem.

Question 2: BASIC CALCULATION

The line of best fit for a scatter plot comparing advertising spending (in thousands of dollars) to monthly revenue (in thousands of dollars) is ŷ = 4.2x + 12. What monthly revenue does the model predict for a company that spends $5,000 on advertising?A) $21,000 B) $33,000 C) $17,200 D) $60,000

Show answer

B) $33,000. Substitute x = 5 into the equation: ŷ = 4.2(5) + 12 = 21 + 12 = 33. The model predicts monthly revenue of $33,000.

Question 3: INTERMEDIATE

A scatter plot shows data for 20 cities, comparing average daily temperature (°F) on the x-axis to natural gas consumption (therms per household) on the y-axis. The line of best fit is ŷ = −0.8x + 95. One city has an average temperature of 50°F and actual gas consumption of 60 therms. Which of the following best describes the residual for this city?

A. The residual is −5, meaning the city consumes 5 fewer therms than the model predicts.
B. The residual is 5, meaning the city consumes 5 more therms than the model predicts.
C. The residual is 55, meaning the model predicts a consumption of 55 therms above the city's actual usage.
D. The residual is −55, meaning the city consumes 55 fewer therms than the model predicts.

Show answer

Step 1: Find the predicted value using the line of best fit: ŷ = −0.8(50) + 95 = −40 + 95 = 55 therms. Step 2: Calculate the residual: Residual = Actual − Predicted = 60 − 55 = 5. The correct answer is (B).

Question 4: APPLIED

A researcher collects data on 30 used cars, plotting the car's age (in years) on the x-axis and its resale value (in thousands of dollars) on the y-axis. The scatter plot shows a strong negative linear association. The line of best fit is ŷ = −2.1x + 28.5, where ŷ is the predicted resale value and x is the car's age. Which of the following is the best interpretation of the slope of the line of best fit in this context?

A. For each additional $2,100 in resale value, the model predicts the car's age increases by 1 year.
B. For each additional year of age, the model predicts the car's resale value decreases by $2,100.
C. For each additional year of age, the model predicts the car's resale value increases by $2,100.
D. A brand-new car has a predicted resale value of $2,100.

Show answer

The slope is −2.1, meaning the x-variable (age) increases by 1 and the y-variable (resale value) changes by −2.1 thousand dollars. In context, for each additional year of age, the model predicts the car's resale value decreases by $2,100. Answer: B

Question 5: CRITICAL THINKING

Two researchers each study the relationship between weekly exercise hours and resting heart rate using 50 adults. Both researchers obtain the same line of best fit, ŷ = −3x + 85, but Researcher A's scatter plot shows a strong negative linear association while Researcher B's scatter plot shows only a moderate negative linear association. Which of the following best explains how both researchers can have the same line of best fit but different strengths of association?

A. Researcher A used more data points than Researcher B, so the line of best fit is more accurate for Researcher A's data.
B. Researcher A's data points are clustered more tightly around the line, producing smaller residuals, while Researcher B's data points are more spread out from the line, producing larger residuals.
C. The line of best fit has a steeper slope for Researcher A's data, which causes the association to appear stronger.
D. Researcher B's data contains only positive residuals, which weakens the association compared to Researcher A's data.

Show answer

B — The line of best fit represents the average trend in the data and is determined by minimizing the total squared residuals. Two datasets can produce the same line yet differ in association strength based on how closely the individual points cluster around that line. Researcher A's points are tightly packed near the line (small residuals → strong association), while Researcher B's points are more spread out from the line (larger residuals → moderate association). The direction and slope of the trend are the same, but the scatter around the line differs. Choice A is wrong because both researchers used the same number of data points (50 adults each), so sample size does not explain the difference in association strength. Choice C is wrong because both researchers obtained the identical line of best fit, ŷ = −3x + 85, meaning the slope is the same for both. Choice D is wrong because residual sign (positive or negative) does not determine association strength; it is the magnitude of the residuals and how closely points cluster around the line that matters.