Deck 17: Doing Real Research: Elementary Multivariate Relationships

Question 1

Multiple R is used when:

A) there are multiple dependent variables
B) there are two or more predictor variables
C) all distributions are heteroscedastic
D) all of the above

Answer 1

B

Question 2

A bivariate correlation is also known as a(n)

A) zero-order correlation
B) single-order correlation
C) double-order correlation
D) basic correlation

Answer 2

A

Question 3

Partial regression differs from multiple regression in that

A) it is only a single step away from bivariate regression
B) only one independent variable is used at a time
C) it uses only one dependent variable
D) it is restricted to no more than 5 variables

Answer 3

B

Question 4

One of the problems of layered tables is the need for a large sample size.

Question 5

A layer tabular analysis would take place when

A) a crosstabulation is expanded to include more than 8 cells
B) the original bivariate crosstabulation is reconstructed within each category of a third variable
C) the original bivariate crosstabulation is reconstructed within each category of the dependent variable
D) all of the above

Question 6

Which answer below best characterizes multiple regression?

A) it adds variables to see the effect of the combined variable set on a dependent variable
B) it multiplies the effect of a bivariate relationship by the number of added control variables
C) it regresses each variable on the dependent variable and then sums their bivariate correlations
D) none of the above

Question 7

An R-square of .20 means that

A) 20% of the variance in the dependent variable has been explained by the set of independent variables
B) 40% of the variance in the dependent variable has been explained by the set of independent variables
C) 4% of the variance in the dependent variable has been explained by the set of independent variables
D) 80% of the variance in the dependent variable has been explained by the set of independent variables

Question 8

The name for partial regression comes from partitioning variance in the dependent variable by the effects of independent and control variables.

Question 9

Bivariate statistics are closer to a beginning step in research rather than an ending step.

Question 10

Multiple regression can be used to

A) determine which predictor variable is the best predictor
B) the effect of all variables on a dependent variable
C) use variables as a block, like a single predictor
D) all of the above

Question 11

The best way to find a total multiple correlation is to sum the individual bivariate correlations.

Question 12

Multiple regression

A) has more assumptions than regression
B) has the same assumptions as regression
C) has fewer assumptions than regression
D) the two cannot be compared because of the difference in their equations

Question 13

If three independent variables constitute a multiple R of .5, they are each positively related to the dependent variable.

Question 14

Conceptual property space refers to

A) the amount of measurement present in any one concept
B) the use of unstandardized slopes to refer to the amount of property inherent in a concept
C) the amount of measurement present in any one variable
D) independent and shared measurement

Question 15

A major reason for using statistics is to produce an objective set of findings on which to base policy.

Question 16

The basic difference between multivariate and bivariate analyses is that there are at least two dependent variables to interpret.

Question 17

A bivariate correlation of .40 is larger than a multiple correlation of .40 because of the number of variables in the multiple correlation.

Question 18

Because of their serious data limitations, multiple and partial regression are seldom seen in the actual research literature.

Question 19

An R-square of .80 means that

A) a set of variables has a high correlation with the dependent variable
B) the original bivariate correlation, when controlled for multiple variables, has a high correlation with the dependent variable
C) the original bivariate correlation has an R-square of .20 (1.0 minus .80)
D) a and b above

Question 20

A third-order partial relationship is represented by:

A) a dependent variable, a control variable and an independent variable
B) a dependent variable, two control variables and an independent variable
C) a dependent variable, three control variables and an independent variable
D) a dependent variable and two independent variables

Question 21

Multivariate analyses, because of all of their assumptions, are less like the "real world" than bivariate analyses.