Page 1 of 193
ISYE 6414-REGRESSION ANALYSIS FINAL EXAM 550+
QUESTIONS AND DETAILED SOLUTIONS LATEST
UPDATE THIS YEAR
ISYE 6414-REGRESSION ANALYSIS FINAL EXAM
In the cancer survival example, the survival time is statistically different for those patients
with Breast cancer versus those with Bronchus or Stomach cancer.
True
In regression, the explanatory variables are random variables, while the response is fixed.
False
In simple linear regression, we aim to create a deterministic, linear model between the
explanatory variable and response variable.
False
In a general sense, a fitted model is simply the output of minimizing the errors as measured
by a particular criterion.
True
We define residuals as the difference between the observed response values and the fitted
response values.
True
When the t-value for a beta value is large, we reject the null hypothesis, which by default is
that the beta value is equal to zero.
True
1
,Page 2 of 193
Goodness of fit describes how accurately a model fits the observed data by minimizing its
residuals.
False
Leverage points need to be removed because they influence a regression model's fit
significantly.
False
In simple linear regression, R-squared is the same as the square of the correlation coefficient.
True
If we run a simple linear regression and encounter a bi-modal distribution for an outcome
variable, we should discard the linear model.
False
Your computer broke, but a colleague has asked you to review a brief report they assembled
on a dataset. You see a graph of a square root (constantly increasing, but decreasing in the
rate of increase.) Your colleague is new to coding but ran a Box-Cox test and reported the
suggested value of lambda = 2.
You can't check the test, but does it seem plausible?
True
Q: If x* is one of the observations for the predicting variable, then we use estimation. (T/F)
True
In Box-Cox transformation, if the value of lambda is 1, we:
do not need to transform
2
,Page 3 of 193
R-square is 1 minus the ratio between the sum of squared errors and sum of square total.
True
In ANOVA, when we comparing the means, the alternative hypothesis is that at least two of
the means are different from each other.
True
Which of the following assumptions is not needed in ANOVA
Linearity
If we have independent chi-squared random variables, their sum is also a chi-squared
distribution.
True
In the PPP (purchasing power parity) example, what do we conclude from the boxplot about
the exchange rate change between the developed and developing countries?
they have very different exchange rate change
When we are conducting ANOVA, we cannot use boxplots to see whether group means are
statistically significantly different from each other.
True
The F-test for equal means has a null hypothesis that the means are different: if you obtain a
lower p-value than your threshold (normally 0.05), then you reject the null hypothesis and
accept that the means tested as part of the F-test are equal.
False
3
, Page 4 of 193
One objective in ANOVA is to determine which groups have statistically significantly different
means from each other. We can compare two groups at a time, which is called pairwise
comparison, and the Tukey method (TukeyHSD()) will generate all of these for you.
True
Linearity is not an assumption of ANOVA - a key difference from Simple Linear Regression.
True
To check the constant variance assumption of ANOVA, we ought to plot the residuals by
treatment group to assess whether the groups individually have different variability that
contradicts our assumptions.
True
Professor Serban conducts exploratory analysis of the inflation difference (predictor) and the
average annual change in exchange rate (response). She displays a boxplot showing the
response variables for developed and developing countries.
Choose True if both of these statements are true:
1. This boxplot is generated from the original dataset after recoding the dummy variable
(Developed) to be a categorical variable (with two factors: "Developed" and "Developing".
2. The visual difference in the response variables shown in the boxplots provide strong
evidence that two simple linear regressions (one for Developing and one for Developed
countries) would be better than a single model.
False
The conclusion of the analysis is that purchasing power parity (PPP) is not 'robust' to unusual
economic or political conditions."
This implicitly means that in certain circumstances, the predictive power of PPP becomes
weaker.
4
ISYE 6414-REGRESSION ANALYSIS FINAL EXAM 550+
QUESTIONS AND DETAILED SOLUTIONS LATEST
UPDATE THIS YEAR
ISYE 6414-REGRESSION ANALYSIS FINAL EXAM
In the cancer survival example, the survival time is statistically different for those patients
with Breast cancer versus those with Bronchus or Stomach cancer.
True
In regression, the explanatory variables are random variables, while the response is fixed.
False
In simple linear regression, we aim to create a deterministic, linear model between the
explanatory variable and response variable.
False
In a general sense, a fitted model is simply the output of minimizing the errors as measured
by a particular criterion.
True
We define residuals as the difference between the observed response values and the fitted
response values.
True
When the t-value for a beta value is large, we reject the null hypothesis, which by default is
that the beta value is equal to zero.
True
1
,Page 2 of 193
Goodness of fit describes how accurately a model fits the observed data by minimizing its
residuals.
False
Leverage points need to be removed because they influence a regression model's fit
significantly.
False
In simple linear regression, R-squared is the same as the square of the correlation coefficient.
True
If we run a simple linear regression and encounter a bi-modal distribution for an outcome
variable, we should discard the linear model.
False
Your computer broke, but a colleague has asked you to review a brief report they assembled
on a dataset. You see a graph of a square root (constantly increasing, but decreasing in the
rate of increase.) Your colleague is new to coding but ran a Box-Cox test and reported the
suggested value of lambda = 2.
You can't check the test, but does it seem plausible?
True
Q: If x* is one of the observations for the predicting variable, then we use estimation. (T/F)
True
In Box-Cox transformation, if the value of lambda is 1, we:
do not need to transform
2
,Page 3 of 193
R-square is 1 minus the ratio between the sum of squared errors and sum of square total.
True
In ANOVA, when we comparing the means, the alternative hypothesis is that at least two of
the means are different from each other.
True
Which of the following assumptions is not needed in ANOVA
Linearity
If we have independent chi-squared random variables, their sum is also a chi-squared
distribution.
True
In the PPP (purchasing power parity) example, what do we conclude from the boxplot about
the exchange rate change between the developed and developing countries?
they have very different exchange rate change
When we are conducting ANOVA, we cannot use boxplots to see whether group means are
statistically significantly different from each other.
True
The F-test for equal means has a null hypothesis that the means are different: if you obtain a
lower p-value than your threshold (normally 0.05), then you reject the null hypothesis and
accept that the means tested as part of the F-test are equal.
False
3
, Page 4 of 193
One objective in ANOVA is to determine which groups have statistically significantly different
means from each other. We can compare two groups at a time, which is called pairwise
comparison, and the Tukey method (TukeyHSD()) will generate all of these for you.
True
Linearity is not an assumption of ANOVA - a key difference from Simple Linear Regression.
True
To check the constant variance assumption of ANOVA, we ought to plot the residuals by
treatment group to assess whether the groups individually have different variability that
contradicts our assumptions.
True
Professor Serban conducts exploratory analysis of the inflation difference (predictor) and the
average annual change in exchange rate (response). She displays a boxplot showing the
response variables for developed and developing countries.
Choose True if both of these statements are true:
1. This boxplot is generated from the original dataset after recoding the dummy variable
(Developed) to be a categorical variable (with two factors: "Developed" and "Developing".
2. The visual difference in the response variables shown in the boxplots provide strong
evidence that two simple linear regressions (one for Developing and one for Developed
countries) would be better than a single model.
False
The conclusion of the analysis is that purchasing power parity (PPP) is not 'robust' to unusual
economic or political conditions."
This implicitly means that in certain circumstances, the predictive power of PPP becomes
weaker.
4
