Chapter_11_Correlational Designs

Question 1

Correlation Coefficient (r)

Answer 1

an indicator of the degree of
relationship between two variables

Question 2

The square of the correlation coefficient (r^2)

Answer 2

the proportion of the variance on one variable that can be accounted for by the other variable

Question 3

Differences in Correlations

Answer 3

the relationship differs between groups in the population

Question 4

2 assumptions underlying correlation

Answer 4

1. Linearity
2. Additivity

Question 5

Factors Affecting the Correlation Coefficient

Answer 5

1. Reliability of the Measures
2. Range Restriction
3. Outliers
4. Subgroup Differences

Question 6

3 Circumstances Should Keep Facets Separate

Answer 6

1. when the facets are theoretically or empirically RELATED to different dependent variables
2. when the theory of the construct predicts an INTERACTION among the facets
3. facets should not be combined simply as a CONVENIENCE

Question 7

2 Circumstances Should Combine Facets

Answer 7

1. when a researcher is INTERESTED in the latent variable
2. when the latent variable is more IMPORTANT in theory

Question 8

2 Forms of Multiple Regression Analysis

Answer 8

1. Simultaneous MRA
2. Sequential MRA

Question 9

Simultaneous MRA

Answer 9

to derive the equation that most accurately predicts a criterion variable

Question 10

Sequential MRA

Answer 10

control other variable for the effect of target variable

Question 11

Information Provided by MRA

Answer 11

1. multiple correlation coefficient (R != r)
2. regression coefficient (b)
3. change in R^2

Question 12

Multicollinearity

Answer 12

two or more predictor variables are highly correlated with each other

Question 13

Effects of Multicollinearity

Answer 13

1. inflation of the SE of regression coefficients
2. nonsignificant statistical test
3. misleading conclusions about changes in R^2

Question 14

Causes of Multicollinearity

Answer 14

1. inclusion of multiple measures of the same construct
   solution
   - a latent variables analysis
2. measures of conceptually different constructs have highly correlated scores
3. sampling error can lead to artificially high correlations

Question 15

The Empirical Problems of Dichotomization or median split

Answer 15

1. can’t detect CURVILINEAR relationship
2. reduces the RELIABILITY of the measure
3. participants who are in DIFFERENT SAMPLES but who have the same scores would be classified differently

Question 16

The Statistical Problems of Dichotomization

Answer 16

1. loss of statistical­ power
2. spurious statistical significance