draw a line of best fit for a scatter plot with a linear association by judging how close the points are to the line, and find its slope.
Phase 4: Build
Do Now. Scientists have measured carbon dioxide in the air on a Hawaiian mountain since 1958. Read about the record. Then look at the yearly numbers in the table and predict.
The annual mean went from about 401 ppm in 2015 to about 427 ppm in 2025. About how much is that each year, on average?
The reading opens over the lesson; close it to come back to exactly where you were.
A line can tell the story of many points. When a scatter plot shows a linear association, we can draw a line of best fit, also called a trend line. It does not have to touch any point. It follows the direction of the cloud.
How to judge a good fit. (1) The line follows the trend of the whole data set. (2) About as many points are above the line as below it. (3) The points are as close to the line as possible all along it, not just at one end. (4) Ignore an outlier when you place the line, and say so on your panel.
| Year | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Years after 2015 (x) | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| Annual mean CO2 (ppm, rounded) | 401 | 404 | 407 | 409 | 412 | 414 | 416 | 419 | 421 | 425 | 427 |
Worked example 1: Nia fits the sleep data. On the Sleep and Scores sample plot, Nia lays a clear ruler across the points. She settles on a line through (6, 70) and (9, 88). Checking: 3 points are above the line, 4 are below and 3 are on it, and none is more than 3 points away. She finds the slope: (88 − 70) ÷ (9 − 6) = 18 ÷ 3 = 6 points per hour.
Worked example 2: Diego fits the CO2 data. Diego uses x = years after 2015, so 2015 is x = 0. His line passes through (0, 401) and (10, 427). Slope: (427 − 401) ÷ (10 − 0) = 26 ÷ 10 = 2.6 ppm per year. He checks the middle: at x = 5 the line gives 401 + 2.6 × 5 = 414, and the table says 414. The line stays within 1 ppm of every value.
Two lines, one choice. If two classmates draw different lines, compare them with the four checks. Count points above and below, and look for the line with the smaller gaps overall.
Use a clear ruler or a strand of spaghetti on your plots.
Add a line of best fit to your panel’s plot.
Draw a line of best fit on your Screens and Sleep sample plot. Name two points on your line and find its slope. Is the slope positive or negative, and why does that make sense?
Write a note for your panel explaining how you chose the line, using at least two of the four checks.
A line of best fit is a summary. Good summaries leave out no part of the story.
Exit ticket: why should you ignore an outlier when you place a line of best fit, and what should you write on the panel when you do?
Pacing: warm-up and reading 7 minutes, mini lecture 8, career note 2, guided practice with sim 15, independent 6, closure 2.
Likely misconceptions: connecting the first and last points no matter where the others lie; forcing the line through the origin; counting only points above and below without looking at distance.
Supports: clear rulers or spaghetti strands; a check card with the four fit checks. Stretch: compare the 2.6 ppm per year slope with the average of NOAA’s yearly growth rates for 2016-2025 (2.564 ppm per year, from the growth rate page used in 1.25).
Data: the annual means come from the Mauna Loa annual mean file linked on NOAA GML’s data page (opened 2026-11-20): 401.01, 404.41, 406.76, 408.72, 411.65, 414.21, 416.41, 418.53, 421.08, 424.61 and 427.35 ppm for 2015-2025, rounded to whole ppm in the table. The monthly values in the reading change each month; the ones quoted are from the update of November 19, 2026.
Answer guide: warm-up 2.6 ppm per year; Line B; slope 6; false; 2.6; the same, 2.6. Farthest from Diego’s line: the line gives 401, 403.6, 406.2, 408.8, 411.4, 414, 416.6, 419.2, 421.8, 424.4 and 427, so the gaps are 0, 0.4, 0.8, 0.2, 0.6, 0, 0.6, 0.2, 0.8, 0.6 and 0 ppm; 2017 (0.8 above) and 2023 (0.8 below) tie for farthest. Nia’s line: residuals 0, −1, −1, +2, +1, 0, +3, −1, 0, −1 (3 above, 4 below, 3 on). Screens and Sleep: a line through about (0.5, 9.5) and (3.5, 6.5) has slope −3 ÷ 3 = −1 hour of sleep per hour of screen time; accept slopes from about −0.9 to −1.1. Career facts come from the BLS Occupational Outlook Handbook page for data scientists (opened 2026-11-20).
NOAA Global Monitoring Laboratory, Trends in Atmospheric Carbon Dioxide · U.S. government work, public domain
The Global Monitoring Laboratory is part of NOAA, the National Oceanic and Atmospheric Administration. Its page on carbon dioxide trends shows measurements taken at Mauna Loa Observatory in Hawaii. The page says this is the longest record of direct measurements of carbon dioxide, or CO2, in the atmosphere.
The record was started in March 1958 by C. David Keeling of the Scripps Institution of Oceanography, at a NOAA facility. NOAA began its own measurements in May 1974, and the two sets have run side by side since then.
CO2 is reported in parts per million, or ppm. The page gives an example: the fraction 0.000400 is written as 400 ppm. Each monthly value is an average of daily values, which are themselves built from hourly averages.
When the page was last updated on November 19, 2026, it listed a monthly average of 426.69 ppm for October 2026 and 424.87 ppm for October 2025.
The page also notes a challenge. After the Mauna Loa volcano erupted, measurements stopped on November 29, 2022. From December 2022 to July 4, 2023, they came from a site about 21 miles north, until Mauna Loa observations resumed.
Source: NOAA Global Monitoring Laboratory, Trends in Atmospheric Carbon Dioxide (gml.noaa.gov)