Skip to main content
New YorkMathsSyllabus dot point

How do you fit a line to two-variable data, interpret its slope and intercept, and read the correlation and residuals?

Construct and interpret scatter plots; fit a linear (or exponential) model to bivariate data; interpret the slope and intercept in context; compute and interpret residuals; and distinguish the correlation coefficient from causation.

A NY Regents Algebra I answer on bivariate data: scatter plots, fitting a line of best fit, interpreting slope and intercept, computing residuals, reading the correlation coefficient, and the correlation-versus-causation distinction.

Generated by Claude Opus 4.89 min answer

Reviewed by: AI editorial process; not yet individually human-reviewed

Have a quick question? Jump to the Q&A page

Jump to a section
  1. What this topic is asking
  2. Scatter plots: form, direction, strength
  3. The line of best fit and its parameters
  4. Residuals
  5. Correlation versus causation
  6. Try this

What this topic is asking

The Regents Algebra I exam (the Interpreting Categorical and Quantitative Data, S-ID, cluster) wants you to build and read a scatter plot, fit a line of best fit to two-variable data, interpret its slope and intercept in context, compute and interpret residuals, and tell the correlation coefficient apart from causation. Two-variable statistics reliably contributes several questions, including a multi-step constructed-response item.

Scatter plots: form, direction, strength

A scatter plot plots each data pair (x,y)(x, y) as a point. You describe it three ways: form (does it follow a line or a curve?), direction (as xx rises, does yy rise, a positive association, or fall, a negative one?), and strength (how tightly do the points cluster around the trend?). A roughly straight, tight, upward cloud is a strong positive linear association.

The line of best fit and its parameters

The line of best fit (least-squares regression line) is the line that best models the trend, usually found with a graphing calculator. In context its parameters carry meaning:

The hat on y^\hat{y} marks it as a prediction, not an observed value. Interpreting the slope and intercept in the situation's units is a frequent constructed-response task: for y^=3.2x+14\hat{y} = 3.2x + 14 relating study hours to score, the slope means about 3.2 more points per hour, and the intercept is the predicted score with no study.

Residuals

A residual measures how far an actual data point lies from the prediction.

Correlation versus causation

The correlation coefficient rr ranges from 1-1 to 11. Values near 11 or 1-1 indicate a strong linear relationship (positive or negative), and values near 00 indicate little linear relationship. A residual plot is a second diagnostic: a patternless scatter of residuals supports a linear model, while a curved pattern suggests a nonlinear one fits better.

The most tested conceptual point is that correlation does not imply causation. Two variables can move together because one causes the other, because a third variable drives both, or by coincidence. Ice cream sales and drowning rates correlate (both rise in summer), but neither causes the other. On the Regents, a strong rr supports prediction within the data range but never proves that changing xx would change yy.

Try this

Q1. For y^=0.5x+30\hat{y} = -0.5x + 30, interpret the slope if xx is days and yy is battery percent. [2 credits]

  • Cue. The battery drops about 0.5 percent per day.

Q2. Actual value 12, predicted 15. Find the residual and state over/underestimate. [2 credits]

  • Cue. 1215=312 - 15 = -3; negative, so the model overestimates.

Exam-style practice questions

Practice questions written in the style of NYSED exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.

Regents (style)2 marksPart I (multiple choice). A line of best fit is y=3.2x+14y = 3.2x + 14, where xx is hours studied and yy is the test score. What does the slope 3.2 represent? (1) the score with no studying (2) the increase in score per additional hour studied (3) the maximum possible score (4) the number of hours studied
Show worked answer →

The correct answer is (2).

In y=mx+by = mx + b, the slope m=3.2m = 3.2 is the change in yy per one-unit change in xx. Here that is the increase in predicted test score for each additional hour studied (about 3.2 points per hour). The intercept 14 is the predicted score with no studying (choice 1 describes the intercept, not the slope).

Regents (style)4 marksPart III (constructed response). A line of best fit for plant height (cm) versus weeks is y^=2.5x+6\hat{y} = 2.5x + 6. (a) Predict the height at week 8. (b) The actual height at week 8 was 24 cm. Compute the residual and state whether the model overestimates or underestimates.
Show worked answer →

A 4-credit question with credits across the parts.

(a) Predicted height: y^=2.5(8)+6=20+6=26\hat{y} = 2.5(8) + 6 = 20 + 6 = 26 cm.
(b) Residual =actualpredicted=2426=2= \text{actual} - \text{predicted} = 24 - 26 = -2 cm. A negative residual means the actual value is below the prediction, so the model overestimates at week 8. Computing predicted minus actual (the wrong order) or omitting the over/underestimate interpretation costs credits.

Related dot points

Sources & how we know this