What is regression?
1. A scatter plot of wage and education shows an upward cloud of points. What does simple regression add to the graph? Hint onlyA fitted line that summarizes the average relationship between wage and education. A proof that education causes wage changes for every worker. A rule that every data point must lie exactly on one line. A way to remove the need for economic interpretation.
Dependent and explanatory variables
2. A student estimates voteA on shareA. Which interpretation of the variable roles is correct? Hint onlyvoteA is the dependent variable and shareA is the explanatory variable. shareA is the dependent variable because it appears after the word on. Both variables are residuals until the model is estimated. The variable roles do not affect coefficient interpretation.
The simple regression equation
3. In y = beta0 + beta1x + u, which part captures influences on y that are not included in x? Hint onlyu beta1 x beta0
What does beta0 mean?
4. Why can an intercept be useful even when x = 0 is not realistic in the sample? Hint onlyIt helps position the fitted line over the observed range of x. It guarantees the slope is causal. It makes every residual equal to zero. It replaces the need for a dependent variable.
What does beta1 mean?
5. In the teaching wage sample, the education slope is about 2.33. Which sentence is best? Hint onlyOne more year of education is associated with about $2.33 higher predicted hourly wage in this sample. Every worker's actual wage rises by exactly $2.33 after one more year of education. Education explains 2.33 percent of wage variation. The model predicts $2.33 for workers with zero education.
What is the error term?
6. Why can a simple wage-on-education regression be descriptive but not automatically causal? Hint onlyOther factors such as experience, occupation, or ability may affect wages and be related to education. Regression cannot estimate a slope from numeric variables. The error term removes all omitted variables from concern. A positive coefficient is always a causal coefficient.
Fitted values and residuals
7. A worker's actual wage is 26.90 and the fitted wage is 24.85. What is the residual? Hint only2.05 -2.05 51.75 24.85
Ordinary Least Squares intuition
8. Which line does Ordinary Least Squares choose? Hint onlyThe line with the smallest sum of squared residuals. The steepest possible line through the scatter plot. The line that ignores all points below the mean. The line where the intercept is always zero.
R-squared
9. If a regression has R-squared = 0.72, what is the best interpretation? Hint onlyThe model explains 72 percent of the sample variation in the dependent variable. The slope is 0.72 units of y per unit of x. There is a 72 percent probability that x causes y. The residual for every observation is 0.72.
Regression in Python
10. In statsmodels, why do we usually add a constant column before estimating a simple regression? Hint onlyTo estimate an intercept along with the slope. To sort the dataset by the dependent variable. To make the slope equal to zero. To convert the CSV into a chart.
Interpreting regression output
11. Which output-reading sequence is strongest for a beginner? Hint onlyRead the slope in units, check R-squared, then add uncertainty and interpretation cautions. Report only the p-value and ignore the coefficient's units. Use R-squared as proof of causality. Interpret the intercept as the only important result.
Practice regression project
12. A student uses SLEEP75 to regress sleep on totwrk. What is a careful final claim? Hint onlyIn the sample, more total work is associated with a different predicted amount of sleep; causal claims need more support. Total work has been proven to be the only cause of sleep differences. The slope can be interpreted without knowing the units of sleep or totwrk. The scatter plot is unnecessary once Python prints output.
Datasets
13. Which dataset-question pair is well matched for simple regression practice? Hint onlyCEOSAL2: CEO salary as y and return on equity or firm sales as x. WAGE1: residual as y before estimating any regression. VOTE1: use two dependent variables and no explanatory variable. MEAP93: choose a chart color as the outcome.
Formula review
14. What does y-hat represent in the fitted equation y-hat = beta0-hat + beta1-hat x? Hint onlyThe model's predicted value of y for a given x. The observed value of x. The residual after subtracting y from x. The sample size.
Interpretation review
15. What makes a regression interpretation high quality? Hint onlyIt names the variables, uses the correct units, explains fit or uncertainty, and avoids unsupported causal claims. It reports the biggest number in the output and stops there. It uses causal language whenever the coefficient is positive. It ignores residuals because they are not part of the fitted equation.