Study guide

10.3 Residuals and Model Appropriateness

New York Regents Algebra II Exam

Track your progress as you read

Finished sections get a check mark.

1 Chapter 1: Regents Exam Format, Scoring & Reference Guide

1.1 Regents Exam Format, Parts, and Score Conversion 1.2 The Next Generation Reference Sheet, Calculator Rules & Pacing

2 Chapter 2: Number and Quantity: Complex Numbers & Exponents

2.1 Rational Exponents and Radicals 2.2 The Complex Number System and Operations 2.3 Complex Solutions of Quadratic Equations

3 Chapter 3: Polynomial Arithmetic & Structure

3.1 Polynomial Structure and Factoring Techniques 3.2 Factoring Cubes, Polynomial Identities, and Binomial Expansions 3.3 Polynomial Division and the Remainder Theorem

4 Chapter 4: Solving Polynomial & Radical Equations

4.1 Zeros, Roots, and Graphing of Polynomial Functions 4.2 Radical Equations and Extraneous Solutions 4.3 Linear-Quadratic Systems: Parabolas and Circles 4.4 Creating Equations from Context and Justifying Each Step 4.5 Graphical Solutions: Reading f(x) = g(x) and f(x) < g(x)

5 Chapter 5: Rational Expressions & Equations

5.1 Simplifying and Operating with Rational Expressions 5.2 Solving Rational Equations and Modeling 5.3 Rational Function Graphs, Asymptotes, and Discontinuities

6 Chapter 6: Function Fundamentals & Transformations

6.1 Function Notation, Domain, Range, and Composition 6.2 Key Features of Functions and Average Rate of Change 6.3 Function Transformations 6.4 Comparing and Combining Functions Across Representations

7 Chapter 7: Inverses, Exponentials & Logarithms

7.1 Inverse Functions and Domain Restrictions 7.2 Exponential Growth, Decay, and Continuous Compounding 7.3 Logarithmic Properties and Solving Exponential/Logarithmic Equations 7.4 Graphing Exponential, Logarithmic, and Cube Root Functions

8 Chapter 8: Sequences and Series

8.1 Arithmetic and Geometric Sequences 8.2 Summation and Finite Series Formulas 8.3 Applications of Series: Financial Models and Annuities

9 Chapter 9: Trigonometric Functions

9.1 Radian Measure, Arc Length, and the Unit Circle 9.2 Graphs of Sine and Cosine & Sinusoidal Modeling 9.3 Trigonometric Identities and Equations

10 Chapter 10: Quantitative Data & Normal Distributions

10.1 The Normal Distribution, Empirical Rule, and Z-Scores 10.2 Bivariate Data, Scatterplots, and Regression Models 10.3 Residuals and Model Appropriateness 10.4 Judging Whether a Normal Model Fits and Choosing a Regression Model

11 Chapter 11: Statistical Inference & Probability

11.1 Statistical Studies: Surveys, Experiments, and Observational Studies 11.2 Sampling Variability, Simulation, and Margin of Error 11.3 Conditional Probability, Two-Way Tables, and Independence 11.4 Compound Events, Rules of Probability, and Expected Value

10.3 Residuals and Model Appropriateness

Key Takeaways Explain This Topic AI

A residual is the vertical deviation between an observed data value and the model's predicted value: residual = observed y - predicted y = y - ŷ.

Positive residuals indicate that the model underpredicted the observed value (point lies above the line), while negative residuals indicate an overprediction (point lies below the line).

A residual plot exhibiting random scatter around the zero line confirms model appropriateness, whereas a distinct curved U-shape pattern indicates that a linear model is inappropriate.

The Pearson correlation coefficient r (-1 ≤ r ≤ 1) measures the direction and strength of a linear relationship, while r² represents the proportion of variance in y explained by x.

A strong correlation coefficient never proves a causal relationship between variables, as confounding or lurking variables may drive the observed association.

Last updated: September 2026

10.3 Residuals and Model Appropriateness

Quick Answer: A residual measures the vertical prediction error of a regression model, calculated as residual = observed y − predicted y = y − y ^ \text{residual} = \text{observed } y - \text{predicted } y = y - \hat{y} residual = observed y − predicted y = y − y ^ ​ (Actual minus Predicted). A positive residual indicates that the model underpredicted the observation, while a negative residual indicates an overprediction . In a residual plot ( x x x vs. residual), a random scatter of points around the horizontal zero line proves that the chosen model is appropriate. A distinct curved pattern (such as a U-shape) proves that a linear model is inappropriate and a non-linear model should be used. The correlation coefficient ( r r r ) measures linear strength and direction ( − 1 ≤ r ≤ 1 -1 \le r \le 1 − 1 ≤ r ≤ 1 ), but correlation never implies causation .

1. Defining and Calculating Residuals

Note

Scope note. Under the Next Generation standards, the residual-plot standard (S-ID.6b) and the correlation-coefficient standard (S-ID.8) sit in Algebra I , not on the Algebra II blueprint, which lists only AII-S.ID.6a for this domain. Residual reasoning is still worth your time in Algebra II for two reasons: it is how you decide which model to fit under AII-S.ID.6a, and interpreting numerical summaries and critiquing statistical claims are exactly AII-S.IC.6a and AII-S.IC.6b . Expect this material to show up as justification rather than as a standalone "make a residual plot" prompt.

Once a regression line or curve is constructed, statisticians must evaluate how accurately the model reflects the observed sample data. The foundational diagnostic tool for this assessment is the residual .

The Residual Formula

For each paired observation ( x i , y i ) (x_i, y_i) ( x i ​ , y i ​ ) in a dataset: Residual = y − y ^ = Observed y − Predicted y \text{Residual} = y - \hat{y} = \text{Observed } y - \text{Predicted } y Residual = y − y ^ ​ = Observed y − Predicted y

Where:

y y y is the actual, observed response measurement from the data table.

y ^ \hat{y} y ^ ​ (pronounced " y y y -hat") is the predicted response value obtained by substituting the explanatory value x x x into the regression equation.

Tip

The "AP" Memory Aid: Remember the letters AP — A ctual minus P redicted ( y − y ^ y - \hat{y} y − y ^ ​ ). Never subtract in reverse order ( y ^ − y \hat{y} - y y ^ ​ − y ), as that inverts the sign of every residual.

Directional Interpretation of Residuals

Residual Sign Mathematical Relationship Graphic Position Relative to Line Model Interpretation

Positive ( y − y ^ > 0 y - \hat{y} > 0 y − y ^ ​ > 0 ) Observed y > Predicted y ^ y > \text{Predicted } \hat{y} y > Predicted y ^ ​ Point lies above the regression line The model underpredicted the true value.

Negative ( y − y ^ < 0 y - \hat{y} < 0 y − y ^ ​ < 0 ) Observed y < Predicted y ^ y < \text{Predicted } \hat{y} y < Predicted y ^ ​ Point lies below the regression line The model overpredicted the true value.

Zero ( y − y ^ = 0 y - \hat{y} = 0 y − y ^ ​ = 0 ) Observed y = Predicted y ^ y = \text{Predicted } \hat{y} y = Predicted y ^ ​ Point lies exactly on the regression line The model predicted the value perfectly .

The Zero-Sum Property of Least-Squares Regression

In any ordinary least-squares linear regression line, the sum of all individual residuals is mathematically guaranteed to equal zero: ∑ i = 1 n ( y i − y ^ i ) = 0 \sum_{i=1}^n (y_i - \hat{y}_i) = 0 i = 1 ∑ n ​ ( y i ​ − y ^ ​ i ​ ) = 0

Positive prediction errors cancel negative prediction errors exactly.

2. Constructing and Interpreting Residual Plots

A residual plot is a specialized diagnostic graph where the horizontal axis represents the explanatory variable x x x (or the predicted values y ^ \hat{y} y ^ ​ ), and the vertical axis represents the computed residuals ( y − y ^ y - \hat{y} y − y ^ ​ ). A horizontal reference line is drawn at residual = 0 \text{residual} = 0 residual = 0 . Appropriate Model (Random Scatter) Inappropriate Model (Curved Pattern)

  • Residual ^ Residual ^
  • | * * | * *
  • + | * + | * *
  • --0-+----+-------+----+---> x ---0-+-------+-------+-----> x
  • | * * - | * *
  • | * | *
  • Diagnostic Decision Rules

Random Scatter ⟹ \implies ⟹ Model is Appropriate: If the points in the residual plot are distributed randomly above and below the zero line across the entire domain—with approximately equal spread and no discernible geometric shape—the chosen regression model is appropriate .

Curved Pattern ⟹ \implies ⟹ Linear Model is Inappropriate: If the residual plot exhibits a distinct curved shape (such as a U-shape, inverted U-shape, or sinusoidal wave), the linear model is inappropriate . A curved residual plot reveals systematic curvature in the original data that the straight line failed to capture, proving that a non-linear model (e.g., quadratic or exponential) is necessary.

Funnel / Fan Shape ⟹ \implies ⟹ Non-Constant Variance: If residuals fan out or pinch inward as x x x increases, the prediction error variability is not constant across the domain (heteroscedasticity).

Residual Plot Visual Pattern Diagnosis of Current Model Recommended Action

Uniform Random Scatter Model is appropriate ; errors are random noise. Retain current regression model for inference and predictions.

U-Shape / Inverted U-Shape Linear model is inappropriate ; underlying data is curved. Fit a quadratic or polynomial regression model.

Monotonic Curve / Exponential Flare Linear model is inappropriate ; rate of change is proportional. Fit an exponential regression model ( y = a ⋅ b x y = a \cdot b^x y = a ⋅ b x ).

3. The Linear Correlation Coefficient ( r r r ) and Determination ( r 2 r^2 r 2 ) (Algebra I standard, used here as supporting evidence)

While residual plots evaluate model appropriateness, the Pearson correlation coefficient ( r r r ) quantifies the strength and direction of a linear relationship between two quantitative variables.

Mathematical Properties of r r r

Bounded Interval: − 1 ≤ r ≤ 1 -1 \le r \le 1 − 1 ≤ r ≤ 1 .

Direction:

If r > 0 r > 0 r > 0 , there is a positive linear association (as x x x increases, y y y tends to increase).

If r < 0 r < 0 r < 0 , there is a negative linear association (as x x x increases, y y y tends to decrease).

Strength:

r = 1 r = 1 r = 1 : Perfect positive linear correlation.

r = − 1 r = -1 r = − 1 : Perfect negative linear correlation.

r = 0 r = 0 r = 0 : No linear relationship between the variables.

General benchmarks: ∣ r ∣ ≥ 0.8 |r| \ge 0.8 ∣ r ∣ ≥ 0.8 represents strong correlation; 0.5 ≤ ∣ r ∣ < 0.8 0.5 \le |r| < 0.8 0.5 ≤ ∣ r ∣ < 0.8 represents moderate correlation; ∣ r ∣ < 0.5 |r| < 0.5 ∣ r ∣ < 0.5 represents weak correlation.

Important

Enabling Diagnostics on TI-84: By default, TI-84 calculators do not display r r r and r 2 r^2 r 2 after running LinReg . You must activate diagnostics once: Press 2nd → \to → 0 ( CATALOG ), scroll down to DiagnosticOn , press ENTER , and press ENTER again until Done appears.

Coefficient of Determination ( r 2 r^2 r 2 )

The coefficient of determination ( r 2 r^2 r 2 ) is the square of the correlation coefficient, satisfying 0 ≤ r 2 ≤ 1 0 \le r^2 \le 1 0 ≤ r 2 ≤ 1 : r 2 = ( r ) 2 r^2 = (r)^2 r 2 = ( r ) 2

Contextual Interpretation: r 2 r^2 r 2 represents the proportion (or percentage) of the total variation in the response variable y y y that is explained by the linear regression relationship with x x x .

For example, if r = − 0.92 r = -0.92 r = − 0.92 , then r 2 = ( − 0.92 ) 2 ≈ 0.8464 r^2 = (-0.92)^2 \approx 0.8464 r 2 = ( − 0.92 ) 2 ≈ 0.8464 . Thus, approximately 84.6 % 84.6\% 84.6% of the variability in the response variable can be explained by the linear model, while 15.4 % 15.4\% 15.4% is due to unexplained random variation or lurking variables.

4. The Critical Principle: Correlation Does Not Imply Causation

A high correlation coefficient ( r ≈ 1 r \approx 1 r ≈ 1 or r ≈ − 1 r \approx -1 r ≈ − 1 ) proves the existence of a strong statistical association, but it never establishes that changes in x x x cause changes in y y y .

Why Association Does Not Mean Causation

Lurking / Confounding Variables: A third unmeasured variable may simultaneously drive changes in both x x x and y y y .

Classic Example: Ice cream sales and swimming pool drownings are strongly positively correlated ( r ≈ 0.85 r \approx 0.85 r ≈ 0.85 ). Buying ice cream does not cause drownings; both variables increase during hot summer weather (the lurking variable: ambient outdoor temperature).

Reverse Causation: It may be that y y y influences x x x , rather than x x x causing y y y .

Establishing Causality: Causation can only be established through a randomized controlled experiment where subjects are randomly assigned to treatment and control groups to isolate the explanatory variable. Observational regression data cannot prove cause and effect.

  • 5. Worked Examples
  • Worked Problem 1: Tabular Residual Calculation

Problem: A researcher models the relationship between hours of flight training ( x x x ) and simulated flight errors ( y y y ) using the linear equation y ^ = − 1.8 x + 24.5 \hat{y} = -1.8x + 24.5 y ^ ​ = − 1.8 x + 24.5 . The data table includes an observation of a pilot with 5 hours 5\text{ hours} 5 hours of training who committed 13 errors 13\text{ errors} 13 errors . Calculate the residual and interpret its meaning.

Step 1: Calculate the predicted value y ^ \hat{y} y ^ ​ for x = 5 x = 5 x = 5 . y ^ = − 1.8 ( 5 ) + 24.5 = − 9.0 + 24.5 = 15.5 errors \hat{y} = -1.8(5) + 24.5 = -9.0 + 24.5 = 15.5\text{ errors} y ^ ​ = − 1.8 ( 5 ) + 24.5 = − 9.0 + 24.5 = 15.5 errors

Step 2: Calculate the residual using y − y ^ y - \hat{y} y − y ^ ​ . Residual = 13 − 15.5 = − 2.5 errors \text{Residual} = 13 - 15.5 = -2.5\text{ errors} Residual = 13 − 15.5 = − 2.5 errors

Step 3: Interpret the residual in context. The residual is − 2.5 -2.5 − 2.5 . Because the residual is negative, the pilot committed 2.5 2.5 2.5 fewer errors than the model predicted. The linear regression equation overpredicted the number of flight errors for this pilot.

Worked Problem 2: Evaluating Residual Plots for Model Appropriateness

Problem: A student fits a linear regression model to the trajectory of a rocket, obtaining r = 0.94 r = 0.94 r = 0.94 . However, the residual plot displays points below the zero line for small and large values of x x x , and points above the zero line for intermediate values of x x x , forming an inverted U-shape. Is the linear model appropriate? Explain your reasoning.

Step 1: Inspect the residual plot for pattern. The residuals follow a distinct non-random, inverted U-shaped curved pattern.

Step 2: Formulate the verdict. No, the linear model is inappropriate .

Step 3: Provide mathematical justification. Even though the correlation coefficient r = 0.94 r = 0.94 r = 0.94 indicates a strong linear association, model appropriateness is determined strictly by the residual plot. The presence of a clear curved pattern indicates that the underlying relationship is non-linear (specifically quadratic), and a non-linear regression model must be used instead.

Worked Problem 3: Interpreting r r r and r 2 r^2 r 2 in Context

Problem: A study of 200 200 200 high school seniors reports a correlation of r = 0.80 r = 0.80 r = 0.80 between hours of physical exercise per week and hours of sleep per night. Calculate r 2 r^2 r 2 , interpret it in context, and determine whether the study proves that exercising more causes students to sleep longer.

Step 1: Calculate the coefficient of determination r 2 r^2 r 2 . r 2 = ( 0.80 ) 2 = 0.64 = 64 % r^2 = (0.80)^2 = 0.64 = 64\% r 2 = ( 0.80 ) 2 = 0.64 = 64%

Step 2: Interpret r 2 r^2 r 2 in context. Approximately 64 % 64\% 64% of the variation in nightly sleep duration can be explained by the linear relationship with weekly exercise hours.

Step 3: Evaluate the causal claim. The study does not prove causation. Because this was an observational study, confounding variables (such as overall stress levels, extracurricular workload, or dietary habits) may influence both exercise and sleep. Correlation alone cannot establish causality.

6. Common Regents Pitfalls & Exam Strategies

Pitfall 1: Relying Solely on r r r to Judge Linearity. A high correlation coefficient ( r = 0.95 r = 0.95 r = 0.95 ) does not guarantee that a line is appropriate. Curved datasets often yield high r r r values. Always examine the residual plot .

Pitfall 2: Confusing Overprediction and Underprediction. If residual = − 4 \text{residual} = -4 residual = − 4 , the actual value is 4 4 4 units lower than predicted, meaning the model overpredicted the outcome.

Pitfall 3: Claiming r 2 r^2 r 2 is a Probability. The coefficient of determination r 2 r^2 r 2 is the proportion of explained variance in the response variable, never the probability that a prediction is correct.