Deck 3: Statistical Analytics With R Part 2

Question 1

What is the most common data format encountered in statistics with R, from among the following?

A) .dta
B) .sas7bdat
C) .csv
D) .sav

Answer 1

C
The comma-separated values format allows easy transfer between R and various other statistical packages, spreadsheets, and other applications.

Question 2

What could be an "object" in R?

A) The output from a statistical procedure
B) A vector of values of a variable
C) A data frame
D) Any of the above

Answer 2

D
An object is a data structure with (1) attributes and (2) suitability for methods which act on its attributes. All objects are of some "class". View this by typing class(name_of_object). A class may be thought of as a type of "blueprint" for an object. Any given statistical method will only work on certain classes of objects for which it was designed.

Question 3

If race is a numerically coded variable, what does the command racef <- factor(race) do?

A) Recodes race so 0 is the first value
B) Converts race into a categorical variable
C) Prints the level labels for race
D) Prints the codebook for race

Answer 3

B
is the correct answer. Factors are categorical variables.
racef <- as.factor(survey$race)
class(survey$race)
[1] "integer"
class(racef)
[1] "factor"

Question 4

"Fitted values" are…

A) observed values
B) standardized values
C) predicted values
D) centered values

Question 5

What command in R is used to implement a binary logistic regression model?

A) logistic
B) blogistic
C) binomial
D) glm

Question 6

In logistic regression and elsewhere, deviance is an important measure of model fit (and model error). Which is not synonymous with deviance?

A) the regressor
B) -2LL
C) model chi-square
D) likelihood ratio

Question 7

What does the relevel command do?

A) Sets the reference level for a categorical variable
B) Standardizes a variable
C) Centers a variable
D) Converts a variable to a flat distribution

Question 8

In a well-fitting model what kind of distribution should this R command display?: hist(olsmodel$residuals)

A) Dots around a 45-degree line
B) A bi-modal distribution
C) A normal distribution
D) A random distribution

Question 9

To compute an estimated marginal mean, sum the observed values and divide by n.

Question 10

For what type of data is gamma regression used?

A) heteroskedastic
B) skewed
C) missing
D) count

Question 11

What model is used for overdispersed count data?

A) gamma
B) poisson
C) negative binomial
D) cloglog

Question 12

What R command mainly used in the textbook to implements multilevel (mixed linear) models?

A) lm
B) lmer
C) hsb
D) mixed

Question 13

What R command is mainly used in the textbook to implement panel data regression models?

A) plm
B) lmer
C) lm
D) glm

Question 14

In panel data regression using the leading R package for this procedure, when asking for a fixed effects model of subject-level effects, what are the model= and effect= specifications?

A) model="within", effect="individual"
B) model="within", effect="year"
C) model="between", effect="individual"
D) model="between", effect="year"

Question 15

In panel data regression using the leading R package for this procedure, what do idiosyncratic effects refer to?

A) random effects variance at the individual level
B) fixed effects variance at the individual level
C) random effects variance not explained by the subject (individual) variable or fixed effects
D) fixed effects, as opposed to variance components

Question 16

In a structural model with standardized structural (path) coefficients, what is the value of a compound (indirect) path?

A) the sum of the structural coefficients
B) the product of the structural coefficients
C) 1.0
D) This value is always indeterminate.

Question 17

In the text supplement for Chapter 3, which R command is primarily used to implement structural equation models?

A) structural
B) structure
C) lavaan
D) path