Learn · Grade 8 Mathematics

Scatterplots and trend lines

Describe associations and make limited predictions.

Lesson 1 of 3About 40 minutesTEKS 8.5(C)TEKS 8.5(D)TEKS 8.11(A)

Original curriculum. These lessons and questions have not been independently reviewed.

01 · Learn the idea

Paired data preserve the pairing

Each point in a scatterplot pairs two measurements on the same case. Invented practice data A are (1,4),(2,5),(3,8),(4,9),(5,12); B are (−2,4),(−1,1),(0,0),(1,1),(2,4); C are (1,6),(2,2),(3,9),(4,4),(5,7),(6,1),(7,8),(8,5). Plot each on its own labeled axes.

A suggests a positive roughly linear association; B makes a curved U pattern; C has no clear overall association in this small example. A curved pattern is not “no association” merely because it is not a line. Do not connect points in collection order as though the joining segments were a measured path.

02 · Build understanding

Use a trend as an approximation

For A, a reasonable approximate line is y = 2x + 2: observed points lie on or near it. At x = 3.5 it predicts y ≈ 9. The line need not pass through every point; residual differences are expected. Predictions within the observed range are usually better supported than far beyond it.

A relationship between practice time and performance in invented data would not by itself prove that time caused every difference. Other factors and sampling choices matter. Describe the observed pattern before making a causal claim.

03 · Build understanding

Worked example: from a table to a trend line

These numbers are made up for practice. A student measures a sunflower seedling once a week. Week, x: 1, 2, 3, 4, 5, 6 and height in centimeters, y: 5, 9, 10, 14, 18, 20.

Step 1: Set up the axes. Put the week on the x-axis and the height on the y-axis, because the height is the thing you want to predict. Label both axes with their units.

Step 2: Plot one point for each week. The pair for week 2 is (2, 9): go right 2, then up 9. Keep each pair together. You get six separate points, and you leave them unconnected.

Step 3: Describe the pattern. As the week goes up, the height goes up, and the points sit close to a straight line. That is a positive linear association.

Step 4: Draw a trend line through the middle of the points and pick two points on it. A line through (1, 5) and (4, 14) fits well. Its slope is the change in y divided by the change in x: (14 − 5)/(4 − 1) = 9/3 = 3 centimeters per week.

Step 5: Find where the line starts. Put (1, 5) into y = 3x + b to get 5 = 3 + b, so b = 2. The trend line is y = 3x + 2.

Step 6: Predict, then check. At 3.5 weeks the line gives y = 3(3.5) + 2 = 12.5 centimeters. Week 3.5 is between week 3 and week 4, and 12.5 is between the heights measured then, 10 and 14, so the prediction makes sense. Three of the six points are on the line, two are 1 centimeter above it and one is 1 centimeter below it.

  • y = 3x + 2
Six sunflower points rise from (1, 5) to (6, 20), close to the trend line y = 3x + 2.

04 · Build understanding

Worked example: a falling trend at a cocoa stand

More made-up data. A cocoa stand writes down the afternoon temperature and the cups it sells. Temperature in °F, x: 30, 35, 40, 45, 50 and cups sold, y: 61, 49, 40, 31, 19.

Step 1: Describe the pattern before you do any arithmetic. As the temperature goes up, the cups sold go down, and the points fall close to a straight line. That is a negative linear association.

Step 2: Use the trend line. A line drawn through the middle of these points is y = −2x + 120. The slope −2 has units: the stand sells about 2 fewer cups for each degree the afternoon warms up.

Step 3: Predict inside the data. For a 42°F afternoon, y = −2(42) + 120 = −84 + 120 = 36 cups.

Step 4: Check against the table. 42°F is between 40°F and 45°F, and 36 cups is between the 40 and 31 cups sold on those afternoons. The prediction fits.

Step 5: Try a temperature far outside the data. For 70°F the line gives −2(70) + 120 = −20 cups. A stand can't sell a negative number of cups. The data stop at 50°F, so the line has no support out at 70°F.

  • y = −2x + 120
Five cocoa points fall from (30, 61) to (50, 19), close to the trend line y = −2x + 120.

05 · Build understanding

Linear or not: look at how y changes

Two invented sets show the difference. Set P has x: 1, 2, 3, 4, 5, 6 and y: 4, 9, 12, 18, 21, 26. From one point to the next, y changes by 5, 3, 6, 3, 5. The jumps aren't equal, but y always goes up and the jumps stay near the same size. The points suggest a linear relationship.

Set Q is a made-up toy drone hop, with seconds x: 0, 1, 2, 3, 4, 5, 6 and height in feet y: 2, 12, 18, 20, 18, 12, 2. Here y changes by 10, 6, 2, −2, −6, −10. The height climbs, slows down, then drops. The points make an arch, so they do not suggest a linear relationship.

Set Q still has an association. Knowing the time tells you a lot about the height. It is a non-linear association, and a straight trend line would be the wrong tool for it.

No association is a third case. The points are spread out like scattered seeds, and knowing x tells you almost nothing about y.

Seven drone hop points rise from (0, 2) to a top at (3, 20), then fall to (6, 2), making an arch.

06 · Build understanding

Three mistakes to watch for

Swapping the pair. In the sunflower data, week 2 goes with 9 centimeters, so the point is (2, 9). Plotting (9, 2) puts the point far to the right and near the bottom, away from all the others.

Calling a curve "no association". The drone hop points do not follow a line, but they follow a clear arch. Say non-linear association. Save "no association" for points with no pattern at all.

Trusting the line far past the data. The cocoa line gives 36 cups at 42°F, which fits the table. The same line gives −20 cups at 70°F, which is impossible. Stay inside the x-values you measured, or say plainly that the prediction is a stretch.

Apply and explain

Your turn

Construct all three scatterplots, label their patterns, and draw a reasonable trend line for A. Use it to predict at x = 2.5 and explain why a prediction at x = 50 is less supported. Create a paired-data question you could study, describing both variables and a possible confounding factor.

Use paper or explain your thinking aloud before opening the solution.

Check your reasoning

A is roughly increasing linear; B is nonlinear; C lacks a clear overall association. Using y = 2x+2 predicts 7 at x = 2.5; other reasonable lines can yield nearby values. Fifty is far outside the observed 1–5 interval. Questions must preserve paired observations and distinguish association from cause.

  • Plot every supplied pair correctly.
  • Contrast linear, nonlinear, and unclear patterns.
  • Use an approximate line and qualify predictions.
  • Avoid causal conclusions from association alone.

Ready to practice?

Try the questions, get feedback, and review your reasoning.

Scatterplots and trend lines practice →