Residuals - AP Statistics

1. What is the best interpretation of this value?
   Answer: The typical distance between an actual y-value and its predicted y-value is about 4.5 units. The standard deviation of the residuals (s) measures the typical size of a prediction error (residual). It represents the approximate average vertical distance that the observed y-values fall from the fitted regression line.
2. What was the predicted first-year college GPA for this student according to the model?
   Answer: 3.2 The formula for a residual is $residual = actual \, y - predicted \, y$. Rearranging for the predicted value gives $predicted \, y = actual \, y - residual$. In this case, the predicted GPA is $3.4 - 0.2 = 3.2$.
3. Which of the following is the best conclusion based on this information?
   Answer: Model B provides a better fit to the data than Model A because its sum of squared residuals is smaller. The sum of squared residuals measures the total prediction error of a model. A model with a smaller sum of squared residuals provides a better fit to the data because the data points are, on average, closer to the model's predictions.
4. Which of the following is NOT a possible set of residuals for the 10 data points?
   Answer: One residual is zero and the other nine residuals are positive. A key property of the least-squares regression line is that the sum of the residuals must be zero. It is impossible for one residual to be zero and the other nine to all be positive, as their sum would be greater than zero. The other options describe scenarios where the residuals could sum to zero.
5. What does this indicate about that specific data point?
   Answer: The point is an outlier because its actual y-value is far from the y-value predicted by the regression line. A large residual means the point lies far vertically from the regression line, which is the definition of an outlier in the context of regression. A point can be an outlier without being high-leverage or influential.
6. What is the most likely effect on the residuals?
   Answer: The residual for that point will increase in magnitude, and the line may shift, causing other residuals to change. Moving a point vertically increases its distance from the original line, thus increasing the magnitude of its residual. Because the least-squares line is calculated based on all points, this change will cause the line itself to recalculate and shift (unless the point is influential in a very specific way), which in turn changes the predicted values and thus the residuals for other points.
7. What was the actual weight of this 6-month-old dog?
   Answer: 16.8 pounds First, find the predicted weight: $\hat{y} = 5.4 + 2.1(6) = 18.0$ pounds. The residual is $y - \hat{y}$. So, $-1.2 = y - 18.0$. Solving for y gives the actual weight: $y = 18.0 - 1.2 = 16.8$ pounds.
8. Which of the following is the most accurate statement based on this residual?
   Answer: The model overestimated the selling price of this house by $12,000. A negative residual means the actual value is less than the predicted value ($y < \hat{y}$). This implies that the model's prediction was higher than the actual selling price, so the model overestimated the price.
9. Which of the following statements about the residuals from this least-squares regression line must be true?
   Answer: The sum of the residuals is equal to zero. A fundamental mathematical property of the least-squares regression line is that the sum of the residuals ($\sum(y_i - \hat{y}_i)$) is always zero.
10. What does this pattern in the residual plot suggest?
    Answer: The variability of the prediction errors is not constant across all levels of advertising spending. A fan-shaped or cone-shaped pattern in a residual plot indicates that the variance of the residuals is not constant (a condition known as heteroscedasticity). This violates a key assumption for regression inference.
11. How is the residual for this car calculated?
    Answer: The residual is calculated as $16,200 - 15,000$. The residual is defined as the difference between the actual observed value (y) and the value predicted by the regression model ($\hat{y}$). In this case, the calculation is $y - \hat{y} = 16,200 - 15,000$.
12. Based on this observation, what should the student conclude?
    Answer: A linear model is not appropriate for describing the relationship in the original data. Any discernible pattern in a residual plot, including a linear pattern, indicates that the linear model does not adequately capture the relationship in the data. A random scatter is desired, so a linear model is not appropriate here.
13. What is the primary purpose of constructing and analyzing a residual plot in a linear regression analysis?
    Answer: To assess the appropriateness of the linear model that was fitted to the data. Residual plots are a key diagnostic tool used to check the assumptions of a linear model, such as the linearity of the relationship and the constant variance of errors, to determine if the model is appropriate.
14. What is the most appropriate conclusion the statistician should draw from this residual plot?
    Answer: A linear model is not appropriate for this data; a nonlinear model would likely be a better fit. Any clear pattern, such as a curve, in a residual plot indicates that the relationship between the variables is not linear. Therefore, a linear model is not appropriate.
15. In what sense is this line considered the 'best fit'?
    Answer: It minimizes the sum of the squared residuals. The least-squares regression line is defined as the line that minimizes the sum of the squared vertical distances (residuals) from the data points to the line.
16. What is the residual for this plant?
    Answer: 0.8 cm First, calculate the predicted height using the regression equation: $\hat{y} = 1.2 + 0.5(10) = 1.2 + 5 = 6.2$ cm. The residual is the actual height minus the predicted height: $residual = y - \hat{y} = 7.0 - 6.2 = 0.8$ cm.
17. Which assumption for regression inference is most likely violated according to this plot?
    Answer: The assumption that the residuals are independent of each other. When residuals are plotted against a time variable, a non-random pattern such as strings of positive or negative values suggests that the error term for one observation is related to the error term of a previous observation. This pattern indicates a lack of independence among the residuals (autocorrelation).
18. Which of the following correctly describes the variables plotted on the axes of a standard residual plot?
    Answer: The residuals are plotted on the vertical axis, and the explanatory variable is plotted on the horizontal axis. A standard residual plot graphs the residuals ($y - \hat{y}$) on the vertical (y) axis against the explanatory variable (x) on the horizontal axis. Sometimes, predicted values ($\hat{y}$) are used on the horizontal axis instead.
19. Which of the following is the best interpretation of this residual?
    Answer: The student's actual final exam score was 5 points higher than the score predicted by the model. A positive residual indicates that the actual value (y) is greater than the predicted value ($\hat{y}$). Therefore, the student's actual final exam score was 5 points above what the model predicted based on their midterm score.
20. Which of the following conclusions is supported by the residual plot?
    Answer: The linear model is an appropriate way to describe the relationship between hours studied and test scores. A residual plot that shows a random scatter of points around the zero line with no obvious pattern indicates that a linear model is appropriate for the data.

Start this quiz