Scatterplots & Association — I can create and interpret scatterplots and describe direction, form, and strength of association. - Grade 9 Math

SECTION 1

Historical Context & Motivation

Humans have always looked for patterns. Ancient astronomers tracked the positions of stars, and merchants recorded prices alongside supply, but for centuries there was no visual tool to show whether two measurements moved together. The scatterplot emerged as one of the most powerful ways to see relationships between two numerical variables at a glance. Its development grew alongside the broader history of statistics and data visualization.

1637 • Descartes Introduces the Coordinate Plane

René Descartes published La Géométrie, introducing the Cartesian coordinate system. Plotting points as (x, y) pairs on perpendicular axes gave mathematicians a common language for graphs.

1833 • First Known Scatterplot

English scientist John Herschel plotted orbital data for double stars as individual dots on a coordinate grid, creating what many historians consider the first true scatterplot.

1886 • Galton and Regression

Francis Galton used scatterplots of parents' heights versus children's heights to discover the concept of regression toward the mean. His work showed that a line could summarize the trend in a cloud of dots.

1896 • Pearson's Correlation Coefficient

Karl Pearson formalized the idea of measuring the strength of a linear relationship with the correlation coefficient r, giving scatterplots a precise numerical companion.

2000s • Big Data & Interactive Plots

Modern tools like spreadsheets and coding libraries make it possible to create scatterplots with millions of points in seconds, enabling scientists and businesses to spot trends instantly.

The central question scatterplots help us answer is simple yet powerful: When one variable changes, does the other variable tend to change in a predictable way? Understanding the direction, form, and strength of that relationship is the focus of this lesson.

SECTION 2

Core Principles & Definitions

Before we dive into creating scatterplots, you need a handful of foundational ideas. Each concept below builds on the last, giving you a complete vocabulary for describing any scatterplot you encounter.

Bivariate Data

Bivariate data consists of pairs of numerical values collected from the same individual or observation. For example, (study hours, test score) for each student in a class.

Explanatory & Response Variables

The explanatory variable (x-axis) is the one you think might influence the response variable (y-axis). Think of x as the input and y as the outcome.

Direction of Association

A positive association means both variables increase together. A negative association means one increases while the other decreases.

Form of Association

The form describes the overall shape — linear (follows a straight line), nonlinear (curves), or no clear pattern.

Strength of Association

The strength tells how tightly the points cluster around the trend. A strong association means points hug the pattern closely; a weak association means points are widely scattered.

✦

KEY TAKEAWAY

Think of a scatterplot as a photograph of a relationship. The direction tells you whether the trend goes uphill or downhill, the form tells you whether the path is straight or curved, and the strength tells you how blurry or sharp the photo is. A perfectly sharp, straight uphill photo is a strong, positive, linear association.

SECTION 3

Visual Explanation — Anatomy of a Scatterplot

The diagram below shows a complete scatterplot with every key component labeled. Study it carefully — knowing these parts will help you build and read scatterplots with confidence.

Each cyan dot represents one student's data pair (study hours, test score). The dashed pink trend line summarizes the overall upward pattern. Notice how all points cluster near the line — this indicates a strong, positive, linear association.

In the diagram above, the horizontal axis represents the explanatory variable (study hours) and the vertical axis represents the response variable (test score). Each dot is one student. The overall cloud of dots drifts upward from left to right, which tells us that students who study more tend to earn higher scores. Because the dots follow roughly a straight line, we call this a linear association. Because the points stay close to that line, the association is strong.

SECTION 4

Mathematical Framework

While describing scatterplots visually is important, statistics gives us tools to attach numbers to what we see. Two key formulas help us talk about the trend line and the strength of the relationship.

Line of Best Fit (Least-Squares Regression Line)

LINE OF BEST FIT

ŷ = a + bx

where ŷ = predicted value of the response variable, a = y-intercept (predicted y when x = 0), b = slope (how much y changes for each one-unit increase in x), and x = value of the explanatory variable.

The slope b carries the direction of association: a positive slope means a positive association, and a negative slope means a negative association. The y-intercept a anchors the line on the graph. Together they let you predict y-values for any given x.

Correlation Coefficient

CORRELATION COEFFICIENT

r = (1/(n−1)) × Σ [( xᵢ − x̄ ) / sₓ] × [( yᵢ − ȳ ) / sᵧ]

where n = number of data points, x̄ and ȳ = means of x and y, sₓ and sᵧ = standard deviations of x and y. The value of r always falls between −1 and +1.

You will not need to compute r by hand very often (calculators handle it), but understanding its meaning is essential. An r close to +1 indicates a strong positive linear association, an r close to −1 indicates a strong negative linear association, and an r near 0 means there is little or no linear relationship.

⚠

Important Reminder

The correlation coefficient r measures only linear strength. A dataset can have a strong curved pattern but an r near zero because the relationship is not a straight line. Always look at the scatterplot before trusting r alone.

SECTION 5

Classifying Association — Direction, Form & Strength

Every scatterplot can be described using three characteristics: direction, form, and strength. The diagram below shows six classic patterns you will encounter. Learning to recognize them instantly is one of the most useful skills in statistics.

Six scatterplot patterns: strong positive linear, weak positive linear, strong negative linear, nonlinear (curved), no association, and a cluster pattern. Notice how r can be misleading for nonlinear data.

Characteristic	Options	What to Look For
Direction	Positive, Negative, or None	Does the cloud of points go uphill (positive), downhill (negative), or have no clear slope?
Form	Linear, Nonlinear, Clusters, or No pattern	Do the points follow a straight line, a curve, separate into groups, or scatter randomly?
Strength	Strong, Moderate, or Weak	How tightly do the points cluster around the trend? Tighter = stronger.
Outliers	Present or Absent	Are there any points that fall far away from the overall pattern?

The four things to mention when describing a scatterplot.

Correlation Coefficient Strength Scale

−1 +1

−1 (0%)0 (50%)+1 (100%)

SECTION 6

Worked Example — Describing a Scatterplot

A health teacher collected data on 10 students: the number of hours of sleep each student got the night before a reaction-time test, and their reaction time in milliseconds. The data pairs are shown in the table below.

Student	Sleep (hrs)	Reaction Time (ms)
A	4	410
B	5	380
C	5.5	365
D	6	340
E	6.5	330
F	7	290
G	7.5	275
H	8	260
I	8.5	245
J	9	230

Sleep vs. Reaction Time data for 10 students

Creating and Interpreting the Scatterplot

1
Step 1 — Identify the Variables

The explanatory variable is sleep (hours) because we suspect more sleep influences reaction time. The response variable is reaction time (ms). Place sleep on the x-axis and reaction time on the y-axis.
2
Step 2 — Choose Axis Scales

Sleep ranges from 4 to 9 hours, so a scale from 3 to 10 with increments of 1 works well. Reaction time ranges from 230 to 410 ms, so a scale from 200 to 450 with increments of 50 is appropriate.
3
Step 3 — Plot Each Data Point

For Student A, go to x = 4 and y = 410 and place a dot. Repeat for all 10 students. Each dot represents one (sleep, reaction time) pair.
4
Step 4 — Describe Direction

As sleep increases (left to right), reaction time decreases (dots fall). This is a negative association.

Direction: Negative
5
Step 5 — Describe Form

The dots follow a roughly straight path from upper-left to lower-right. There is no obvious curve, so the form is linear.

Form: Linear
6
Step 6 — Describe Strength

The points stay very close to an imaginary straight line through the data. Very little scatter exists around the trend. A calculator gives r ≈ −0.99. This is a strong association.

Strength: Strong (r ≈ −0.99)
7
Step 7 — Check for Outliers & Write a Summary

No data points fall far from the trend, so there are no obvious outliers. Full description: There is a strong, negative, linear association between hours of sleep and reaction time. Students who slept more tended to have faster (lower) reaction times. No outliers are present.

Strong, negative, linear association with no outliers.

SECTION 7

Strengths & Limitations of Scatterplots

Scatterplots are incredibly versatile, but like every statistical tool, they have boundaries. Understanding what scatterplots can and cannot do will keep you from drawing incorrect conclusions.

Strengths	Limitations
Reveal the direction, form, and strength of a relationship at a glance.	Only display two variables at a time (without special techniques like color coding).
Make outliers easy to spot — they appear as dots far from the main cluster.	Can become cluttered with very large datasets, making patterns hard to see.
Work for any two quantitative variables, regardless of scale or units.	Do not show causation. A strong association does not prove one variable causes changes in the other.
Help detect nonlinear patterns that a correlation coefficient alone would miss.	Not useful for categorical variables — bar charts or mosaic plots work better for those.

Strengths vs. Limitations of Scatterplots

⚡

CORRELATION ≠ CAUSATION

Imagine you notice that ice cream sales and drowning incidents both increase during the same months. A scatterplot would show a strong positive association, but eating ice cream does not cause drowning! A hidden lurking variable — hot weather — drives both. Always ask yourself whether an unseen third factor could explain the pattern before concluding that x causes y.

SECTION 8

Connection to Advanced Topics

The skills you are building now lay the groundwork for more advanced statistical techniques. Here is a quick preview of where scatterplot analysis leads.

What You Know Now	Where It Leads
Eyeballing a trend line on a scatterplot	Least-squares regression — a precise method to calculate the best-fit line using formulas
Describing direction and strength with words	Correlation coefficient (r) and r² — numerical measures of linear strength and the percent of variation explained
Identifying nonlinear patterns visually	Residual plots — plotting leftover errors to confirm whether a linear model is appropriate
Noting that two variables are associated	Inference for regression — using hypothesis tests to decide if the slope is statistically significant (AP Statistics)

How scatterplot skills connect to future coursework

If you continue into AP Statistics or a college statistics course, you will revisit scatterplots with more powerful tools. For now, focus on the visual interpretation — being able to accurately describe direction, form, and strength is the essential first step that everything else builds upon.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A scatterplot shows the age of a used car (x) versus its resale price (y). As the age increases, the prices generally decrease. Describe the direction of the association and explain your reasoning.

Show answer

The association is negative because as one variable (age) increases, the other (resale price) tends to decrease. On the graph, the cloud of points would slope downward from left to right. This makes real-world sense: older cars typically lose value over time due to wear and depreciation.

Question 2: BASIC CALCULATION

The line of best fit for a dataset is ŷ = 12 + 3.5x. (a) What is the slope, and what does it mean? (b) Predict the y-value when x = 6.

Show answer

(a) The slope is 3.5, which means for every one-unit increase in x, the predicted y-value increases by 3.5 units. The positive slope confirms a positive association. (b) Substituting x = 6: ŷ = 12 + 3.5(6) = 12 + 21 = 33. The model predicts a y-value of 33 when x is 6.

Question 3: INTERMEDIATE

A researcher collects data on daily temperature (°F) and the number of hot chocolate cups sold at a café. The correlation coefficient is r = −0.85. Describe the association in a complete sentence using the words direction, form, and strength.

Show answer

There is a strong, negative, linear association between daily temperature and hot chocolate sales. The negative sign of r = −0.85 tells us the direction is negative (as temperature rises, sales fall). The magnitude |0.85| is close to 1, indicating the relationship is strong. Since the correlation coefficient r only measures linear strength, we describe the form as linear — but we should always verify by looking at the actual scatterplot for possible curvature.

Question 4: APPLIED

A coach records the number of practice hours per week and 100-meter dash times (in seconds) for 8 sprinters. The data shows a strong negative linear trend with the equation ŷ = 14.2 − 0.3x. A new athlete practices 12 hours per week. (a) Predict her dash time. (b) Another athlete practices 40 hours per week. Why would it be inappropriate to use the model for this prediction?

Show answer

(a) Substituting x = 12: ŷ = 14.2 − 0.3(12) = 14.2 − 3.6 = 10.6 seconds. (b) Using x = 40 would give ŷ = 14.2 − 0.3(40) = 14.2 − 12 = 2.2 seconds, which is physically impossible for a human 100-meter dash. This illustrates extrapolation — using a model outside the range of the original data. The linear trend likely does not continue at extreme values, so the prediction would be unreliable and unrealistic.

Question 5: CRITICAL THINKING

A student plots shoe size (x) versus GPA (y) for 50 classmates and finds r = 0.62. She concludes that having bigger feet causes higher grades. Identify at least two flaws in her reasoning and suggest a better interpretation.

Show answer

Flaw 1: Correlation does not imply causation. An r = 0.62 shows an association but cannot prove that shoe size causes GPA changes. Flaw 2: A likely lurking variable is age or grade level. Older students (juniors and seniors) tend to have larger feet simply because they are more physically mature, and they may also have higher GPAs because they have taken more courses or matured academically. A better interpretation: the positive association between shoe size and GPA is likely driven by the confounding variable of age, not a direct causal link between foot size and academic performance.

SUMMARY

Lesson Summary

Summary

A scatterplot displays bivariate data as individual points on a coordinate plane, with the explanatory variable on the x-axis and the response variable on the y-axis. To describe any scatterplot, report three things: the direction (positive, negative, or none), the form (linear, nonlinear, or clusters), and the strength (strong, moderate, or weak), along with any outliers.

The correlation coefficient r quantifies the strength and direction of a linear association, ranging from −1 (perfect negative) to +1 (perfect positive). The line of best fit (ŷ = a + bx) lets you make predictions, but beware of extrapolation beyond your data range. Most importantly, remember that correlation does not imply causation — always consider lurking variables before drawing causal conclusions.