Difficulty: Hard
A regression line has x̄=10, ȳ=25, and slope 2. What is its intercept?
A 10
B 20
C 25
D 5
Answer: D. 5 The intercept is ȳ−b x̄=25−2(10)=5.
A line has x̄=5, ȳ=30, and slope −2. Find its intercept.
A 40
B 20
C 30
D −40
Answer: A. 40 The intercept is ȳ−b x̄=30−(−2×5)=40.
An X-on-Y line has x̄=3, ȳ=10, and slope 0.5. Find its intercept.
A 2
B −2
C 3
D 8
Answer: B. −2 The intercept is x̄−b_xyȳ=3−0.5(10)=−2.
For regression of Y on X, which point must lie on the fitted line?
A (0,0)
B (maximum X,maximum Y)
C (x̄,ȳ)
D (median X,median Y)
Answer: C. (x̄,ȳ) A least-squares line with an intercept passes through the sample mean point.
A model predicts sales of 20 while actual sales are 23. What is the residual?
A −3
B 20
C 43
D 3
Answer: D. 3 Residual equals actual minus predicted, so 23−20=3.
A model predicts 15 while actual output is 11. What is the residual?
A −4
B 4
C 11
D 26
Answer: A. −4 Residual equals observed minus fitted, so 11−15=−4.
For least squares with an intercept, what is the average residual?
A The slope
B Zero
C The mean X
D The variance
Answer: B. Zero The residuals sum to zero, so their average is also zero.
Which quantity is minimized in Y-on-X least squares?
A Squared horizontal residuals
B Absolute rank gaps
C Squared vertical residuals
D Squared predictor values
Answer: C. Squared vertical residuals Y-on-X least squares minimizes the squared vertical distances from observed Y values to fitted Y values.
What does a residual plot with a curve suggest?
A The fit is perfect
B The slope is zero
C The means are equal
D The linear form may fail
Answer: D. The linear form may fail A curved residual pattern indicates systematic structure left unexplained by a straight line.
What does a random residual cloud around zero support?
A A reasonable linear fit
B A proven causal effect
C Perfect prediction
D No outcome variation
Answer: A. A reasonable linear fit A structureless residual pattern is consistent with the linear model’s assumptions.
A Y-on-X line has slope 0.8. If X rises by 5, predicted Y changes by how much?
A Falls by 4
B Rises by 4
C Rises by 5
D Stays unchanged
Answer: B. Rises by 4 The predicted change is 0.8×5=4 units upward.
What are the slope units in a sales-on-advertising model?
A Advertising per sale
B Sales squared
C Sales per advertising unit
D No measurement units
Answer: C. Sales per advertising unit The slope expresses predicted change in sales for one unit of advertising.
When is an intercept often hard to interpret?
A The slope is negative
B The sample is small
C The line crosses means
D Zero X is unrealistic
Answer: D. Zero X is unrealistic If zero is outside the realistic predictor range, the intercept may lack practical meaning.
Why should a fitted line be used cautiously beyond observed X values?
A The pattern may change
B Residuals must be zero
C The slope becomes correlation
D The means move
Answer: A. The pattern may change The observed trend may not hold outside the range used to fit the model.
Which statement best distinguishes regression from correlation?
A Correlation predicts Y
B Regression is directional
C Regression is always causal
D Correlation has a slope
Answer: B. Regression is directional Regression specifies an outcome and predictor, while correlation treats the variables symmetrically.
A steep slope alone tells you what about association strength?
A It proves strong fit
B It proves causation
C Not enough by itself
D It fixes correlation at one
Answer: C. Not enough by itself Slope size depends on measurement units and does not alone show how tightly points fit.
Why can an outlier affect a regression line strongly?
A It changes all ranks equally
B It fixes the intercept
C It removes the residuals
D It can pull the line
Answer: D. It can pull the line Least squares squares residuals, so a distant point can have substantial influence.
If Y is doubled while X units stay fixed, what happens to the Y-on-X slope?
A It doubles
B It halves
C It stays unchanged
D It is squared
Answer: A. It doubles Doubling every Y value doubles the fitted change in Y per unit X.
If X is measured in hundreds instead of units, what happens to the slope value?
A It is unchanged always
B It is scaled
C It becomes a correlation
D It changes sign always
Answer: B. It is scaled Changing predictor units changes the numerical slope while preserving the same fitted relationship.
For a standardized Y-on-X regression, what is its slope?
A The intercept
B The residual mean
C Pearson’s r
D The sample size
Answer: C. Pearson’s r When both variables are standardized, their standard deviations are one and the slope equals r.
If r=1, what do the paired observations show?
A A random cloud
B A curved pattern
C No association
D An exact straight line
Answer: D. An exact straight line A correlation of 1 means all paired points lie exactly on an increasing line.
If the Y-on-X slope is zero and both variables vary, what is r?
A Zero
B One
C Minus one
D Undefined
Answer: A. Zero The slope equals rσY/σX, so with positive standard deviations a zero slope gives r=0.
Which predictor value is safest for interpolation?
A One far beyond range
B One within observed range
C Any negative value
D The largest possible value
Answer: B. One within observed range Interpolation estimates within the span of observed predictor data, where the fitted trend is supported.
What does a fitted value represent?
A Observed predictor
B Residual sign
C Model-estimated outcome
D Sample variance
Answer: C. Model-estimated outcome A fitted value is the outcome estimated by the regression model for a given predictor value.
What does a negative residual mean when residual is observed minus fitted?
A Observed is above fitted
B Slope is negative
C Correlation is negative
D Observed is below fitted
Answer: D. Observed is below fitted A negative observed-minus-fitted residual means the observed outcome is below its prediction.