MGSC 291 Exam 3 Final Study Guide
Exam Questions and Answers GradedA+
Log-log model - Correct answer-a regression where both Y and X are logged and
modeled as log(Y) ~ log(X); used when variables move multiplicatively.
Multiplicative relationship - Correct answer-when X causes proportional or percent
changes in Y.
Logistic regression - Correct answer-regression for binary outcomes using glm(y ~
x, family="binomial").
Binary response - Correct answer-a variable with two outcomes (1/0, yes/no).
exp() transformation - Correct answer-converts predictions from log scale back to
original scale because exp(log(y)) = y.
Fitting a linear model in R - Correct answer-lm(y ~ x, data = df).
Fitting a log-log model in R - Correct answer-lm(log(y) ~ log(x), data = df).
Fitting logistic regression in R - Correct answer-glm(y ~ x, family="binomial",
data = df).
©COPYRIGHT 2025, ALL RIGHTS RESERVED 1
, Predicting from a model in R - Correct answer-predict(model, newdata).
Predicting with SE in R - Correct answer-predict(model, newdata, se.fit = TRUE).
Frequentist statistics - Correct answer-probability defined by long-run frequencies
from repeated samples.
Bayesian statistics - Correct answer-probability represents beliefs; updates priors
with data to get posterior.
Sampling distribution - Correct answer-the distribution of a statistic (like sample
mean) across repeated samples.
Shape of sampling distribution - Correct answer-approximately normal if n > 30
(CLT).
Center of sampling distribution - Correct answer-equal to the population mean.
Spread of sampling distribution - Correct answer-standard error; SE = s/√n.
Standard error (SE) - Correct answer-the standard deviation of a sampling
distribution.
Dependence effect on SE - Correct answer-dependence makes SE too small and
underestimates uncertainty.
©COPYRIGHT 2025, ALL RIGHTS RESERVED 2
Exam Questions and Answers GradedA+
Log-log model - Correct answer-a regression where both Y and X are logged and
modeled as log(Y) ~ log(X); used when variables move multiplicatively.
Multiplicative relationship - Correct answer-when X causes proportional or percent
changes in Y.
Logistic regression - Correct answer-regression for binary outcomes using glm(y ~
x, family="binomial").
Binary response - Correct answer-a variable with two outcomes (1/0, yes/no).
exp() transformation - Correct answer-converts predictions from log scale back to
original scale because exp(log(y)) = y.
Fitting a linear model in R - Correct answer-lm(y ~ x, data = df).
Fitting a log-log model in R - Correct answer-lm(log(y) ~ log(x), data = df).
Fitting logistic regression in R - Correct answer-glm(y ~ x, family="binomial",
data = df).
©COPYRIGHT 2025, ALL RIGHTS RESERVED 1
, Predicting from a model in R - Correct answer-predict(model, newdata).
Predicting with SE in R - Correct answer-predict(model, newdata, se.fit = TRUE).
Frequentist statistics - Correct answer-probability defined by long-run frequencies
from repeated samples.
Bayesian statistics - Correct answer-probability represents beliefs; updates priors
with data to get posterior.
Sampling distribution - Correct answer-the distribution of a statistic (like sample
mean) across repeated samples.
Shape of sampling distribution - Correct answer-approximately normal if n > 30
(CLT).
Center of sampling distribution - Correct answer-equal to the population mean.
Spread of sampling distribution - Correct answer-standard error; SE = s/√n.
Standard error (SE) - Correct answer-the standard deviation of a sampling
distribution.
Dependence effect on SE - Correct answer-dependence makes SE too small and
underestimates uncertainty.
©COPYRIGHT 2025, ALL RIGHTS RESERVED 2
