
make a scatter plot, describe the association, fit a linear model with technology and interpret its slope and intercept in context.
Instruction
When both variables are numbers, plot each pair as a point. The result, a scatter plot, shows whether and how the variables move together.
Data (illustrative, shaped like typical measurements). Eight students measured height and arm span in centimeters:
Height: 152, 158, 160, 165, 168, 172, 178, 183
Arm span: 150, 157, 162, 163, 170, 171, 180, 181
Describing the association. Describe four things. Direction: positive, because taller students tend to have longer arm spans. Form: linear, because the points cluster around a straight line rather than a curve. Strength: strong, because the points are close to the line. Unusual features: no point stands far apart from the pattern.
Fitting a line. You can fit a line by eye, choosing two points that follow the trend. Using (158, 157) and (178, 180), the slope is (180 − 157) ÷ (178 − 158) = 23 ÷ 20 = 1.15. A better method, used by every spreadsheet and graphing tool, is the least-squares line: the line that makes the sum of the squared vertical distances from the points as small as possible. For these data technology gives
predicted arm span ≈ 1.02 × height − 3.76
In a spreadsheet, enter the data in two columns, make a scatter chart and add a linear trendline with its equation; in Desmos, type the table and then y1 ~ m x1 + b.
Interpreting the slope. The slope 1.02 has units of centimeters of arm span per centimeter of height. It means: for each additional centimeter of height, arm span is predicted to be about 1.02 cm longer. Notice the language: “predicted” or “on average”, not “will be”, because individual students vary around the line.
Interpreting the intercept. The intercept −3.76 is the predicted arm span for a height of 0 cm. That is meaningless: no one is 0 cm tall, and the data only cover heights from 152 to 183 cm. Say so. An intercept has a useful interpretation only when an input of zero makes sense and lies near the data.
Predicting. For a student 175 cm tall: 1.02(175) − 3.76 ≈ 174.9 cm, using the unrounded coefficients. That is an interpolation, inside the range of the data, and reasonably trustworthy. Predicting for a 100 cm child would be extrapolation, and the model may not hold there.
Practice. For each data set in the lab: make the scatter plot, describe direction, form and strength, fit a least-squares line with technology, write a sentence interpreting the slope with units, decide whether the intercept means anything, and make one prediction inside the data’s range.
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
The model predicted arm span ≈ 1.02 × height − 3.76 has slope 1.02. Which interpretation is correct?
Using the model, a student 175 cm tall has a predicted arm span of about cm.
The intercept −3.76 means a newborn’s predicted arm span is −3.76 cm.
Measure the height and forearm length of at least six people (family members or friends count), or use the data set in this lesson’s reading. Make a scatter plot, describe direction, form and strength, fit a line with technology and interpret its slope with units.
How confident are you that you can fit a linear model to a scatter plot and interpret its slope and intercept in context?
Practice
Work these on paper or in your notebook, then open Check your answer. Aim for all of Fluency and Application; try at least one Challenge.
Printable version: this unit’s practice workbook (PDF)
Fluency
Build speed and accuracy with the core skill.
- Use technology to find the least-squares line and the correlation coefficient for (2, 11), (6, 21), (7, 20), (13, 36), (14, 36).
Check your answer
Answer: ŷ = 6.672 + 2.158x; r = 0.992Interpret the slope in context: the predicted change in y for each one-unit increase in x. - Use technology to find the least-squares line and the correlation coefficient for (1, 6), (3, 5), (10, 18), (11, 19), (14, 22).
Check your answer
Answer: ŷ = 3.075 + 1.401x; r = 0.982Interpret the slope in context: the predicted change in y for each one-unit increase in x. - Use technology to find the least-squares line and the correlation coefficient for (7, -12), (10, -14), (11, -17), (13, -16), (14, -19).
Check your answer
Answer: ŷ = −5.700 − 0.900x; r = −0.912Interpret the slope in context: the predicted change in y for each one-unit increase in x. - Use technology to find the least-squares line and the correlation coefficient for (4, 11), (5, 12), (6, 18), (8, 25), (10, 27), (12, 37), (13, 35).
Check your answer
Answer: ŷ = −0.560 + 2.912x; r = 0.981Interpret the slope in context: the predicted change in y for each one-unit increase in x. - Use technology to find the least-squares line and the correlation coefficient for (1, -2), (2, -2), (9, 3), (12, 7), (13, 7), (14, 5).
Check your answer
Answer: ŷ = −2.947 + 0.700x; r = 0.959Interpret the slope in context: the predicted change in y for each one-unit increase in x. - Using predicted arm span ≈ 1.021 × height − 3.757, predict the arm span for heights 150, 170 and 190 cm. Which prediction is extrapolation?
Check your answer
Answer: 149.4, 169.8 and 190.2 cm; 190 cm is outside the data (152-183 cm).
Application
Use the skill in context. Show your reasoning.
- Six students report hours studied and test scores: (1, 62), (2, 68), (3, 71), (4, 75), (5, 82), (6, 85). Find the least-squares line and predict the score for 4.5 hours.
Check your answer
Answer: ŷ = 4.6x + 57.73; about 78.4Use technology; 4.5 is inside the data range. - In the study model, interpret the slope and the intercept in context.
Check your answer
Answer: Each extra hour of study is associated with about 4.6 more points, on average; a student who studies 0 hours is predicted to score about 57.7. - A model for Missouri counties (illustrative) is predicted median rent = 12.5 × (median income in thousands) + 600. Interpret 12.5.
Check your answer
Answer: Each $1,000 more in median income is associated with about $12.50 more in predicted median rent.
Challenge
Stretch problems. Expect to think before you write.
- The arm-span data have mean height 167 cm and mean arm span 166.75 cm. Show that the least-squares line passes through (mean height, mean arm span).
Check your answer
Answer: 1.021 × 167 + (−3.7566) = 166.75Every least-squares line passes through the point of means.
Review
Keep earlier skills sharp.
- Write the equation, in slope-intercept form, of the line through (−6, 9) and (0, 3).
Check your answer
Answer: y = −x + 3Slope −1; substitute the point to find the intercept 3. - Solve: −4x + 6 = 9x + 32
Check your answer
Answer: x = −2Move the x-terms to one side: −13x = 26.
Where the Line Comes From: Least Squares Worked Without Technology
SLS worked computation of the arm-span example in this lesson, plus paired July average maximum temperatures from NOAA NCEI Climate at a Glance for San Diego (USW00023188) and Fresno (USW00093193), 2019–2025; retrieved September 23, 2026 · Public domain (works of the United States government); worked computation SLS original
Why this reading. The lesson says “technology gives” the least-squares line, which is true and also leaves a student without a device with nothing to do. It is two columns of arithmetic. Here it is, worked twice: once on the lesson’s own arm-span data, so the numbers you have already been given come out, and once on a real pair of NOAA temperature series.
The two formulas
Let x̄ and ȳ be the two means. Build three sums:
Sxx = ∑(x − x̄)² Sxy = ∑(x − x̄)(y − ȳ) Syy = ∑(y − ȳ)²
Then the least-squares slope is m = Sxy / Sxx, the intercept is b = ȳ − m x̄, and the correlation coefficient is r = Sxy / √(Sxx Syy). The line always passes through (x̄, ȳ), which is a free check: substitute the two means and the equation must balance.
Worked on the arm-span data
Heights add to 1336, so x̄ = 167 cm. Arm spans add to 1334, so ȳ = 166.75 cm.
| Height x | Arm span y | x − 167 | y − 166.75 | product | (x − 167)² | (y − 166.75)² |
|---|---|---|---|---|---|---|
| 152 | 150 | −15 | −16.75 | 251.25 | 225 | 280.5625 |
| 158 | 157 | −9 | −9.75 | 87.75 | 81 | 95.0625 |
| 160 | 162 | −7 | −4.75 | 33.25 | 49 | 22.5625 |
| 165 | 163 | −2 | −3.75 | 7.50 | 4 | 14.0625 |
| 168 | 170 | 1 | 3.25 | 3.25 | 1 | 10.5625 |
| 172 | 171 | 5 | 4.25 | 21.25 | 25 | 18.0625 |
| 178 | 180 | 11 | 13.25 | 145.75 | 121 | 175.5625 |
| 183 | 181 | 16 | 14.25 | 228.00 | 256 | 203.0625 |
| Totals | 0 | 0 | 778.0 | 762 | 819.5 |
The two deviation columns must add to zero; if they do not, stop and find the arithmetic error. Then m = 778 ÷ 762 = 1.02100 and b = 166.75 − (778/762)(167) = −3.7566, which is the model the lesson prints as 1.02x − 3.76. And r = 778 ÷ √(762 × 819.5) = 778 ÷ 790.23 = 0.9845, the 0.98 the lesson attributes to technology.
Check the free point: 1.021(167) − 3.7566 = 166.75 = ȳ.
Worked on real paired data
Seven Julys, each one a pair: the July average maximum temperature in San Diego (x) and in Fresno (y) in the same year.
| July of | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|---|---|
| San Diego x (°F) | 73.3 | 74.4 | 76.0 | 72.2 | 74.1 | 76.7 | 73.3 |
| Fresno y (°F) | 99.0 | 99.9 | 103.3 | 100.8 | 101.8 | 105.2 | 97.0 |
x̄ = 520.0 ÷ 7 = 74.286, ȳ = 707.0 ÷ 7 = 101.000. The three sums come out to Sxx = 15.109, Sxy = 20.140 and Syy = 44.820. So the slope is 20.140 ÷ 15.109 = 1.333, the intercept is 101.000 − 1.333(74.286) = 1.98, and
predicted Fresno ≈ 1.333 × (San Diego) + 1.98, with r = 20.140 ÷ √(15.109 × 44.820) = 0.774.
Interpreting the slope. For each extra degree in San Diego’s July, Fresno’s July is predicted to be about 1.33 degrees warmer. Interpreting the intercept: do not. A San Diego July of 0 °F is 72 degrees outside the data and has never happened; the intercept is a number the algebra needs and not a statement about anything.
A prediction. A San Diego July at 75.0 °F predicts a Fresno July at 1.333(75.0) + 1.98 ≈ 101.9 °F. That is an interpolation, inside the 72.2–76.7 range of the data, so it is the kind of prediction the model supports.
Seven points is few, and r = 0.77 with seven points is not strong evidence of anything. Say so in your write-up; a number with four decimal places is not the same as a finding.
Sources: NOAA NCEI, Climate at a Glance: City Time Series (Fresno) (ncei.noaa.gov); NOAA NCEI, Climate at a Glance: City Time Series (San Diego) (ncei.noaa.gov)