Deck 12: Nonparametric Measures of Correlation

Question 1

To test the significance of the rank-order correlation coefficient for large samples (N > 30), we use the following expression:

Answer 1

Question 2

To test the null hypothesis that X and Y are not associated in the population, we convert Our calculated G to a(n):

A)t score.
B)z score.
C)F ratio.
D)analysis of variance.

Answer 2

z score.

Question 3

Which of the following is NOT a requirement for Spearman's rank-order correlation Coefficient?

A)Both X and Y variables must be ranked or ordered.
B)ordinal data
C)random sampling
D)All of the above are required.

Answer 3

All of the above are required.

Question 4

When ordinal data measurement produces a large number of tied ranks, we should use:

A)Pearson's r.
B)Spearman's r.
C)Cramer's V.
D)Goodman's and Kruskal's gamma.

Question 5

A perfect positive correlations is expressed as +.5

Question 6

If our obtained chi-square value for a 2 x 2 cross-tabulation is equal to 19.4 and we have 10 respondents in each cell, the phi coefficient is:

A).696
B).571
C)2.14
D)none of the above

Question 7

The contingency coefficient (C) will never exceed:

A).1
B).5
C)1
D)2

Question 8

As a measure of association, the phi coefficient can only be used:

A)for nominal data.
B)for tables that are 2 x 2.
C)with samples that have been randomly drawn from a larger population.
D)all of the above

Question 9

Goodman and Kruskal's gamma (G) is used for ordinal data.

Question 10

Agreements and inversions can be understood as expressing the __________ of correlation Between X and Y variables.

A)correction
B)deviation
C)movement
D)di rection

Question 11

Goodman's and Kruskal's gamma is best used when comparing:

A)variables at the nominal level with no tied ranks.
B)variables at the ordinal level with many tied ranks.
C)one nominal with one ordinal variable.
D)one ordinal and one interval variable with no tied ranks.

Question 12

Which of the following is taken into consideration when using Goodman and Kruskal's Gamma?

A)the number of agreements
B)the number of disagreements
C)the number of inversions
D)a and c

Question 13

Which of the following is a requirement when computing Spearman's rank-order Correlation coefficient?

A)nominal level data
B)interval level data
C)ordinal level data
D)b and c

Question 14

To determine the relationship between rank-ordered or ordinal data, we compute:

A)Pearson's r.
B)Spearman's rank-order correlation coefficient.
C)contingency coefficient.
D)Cram er's V.

Question 15

To calculate the phi coefficient we require:

A)the chi-square value.
B)total number of cases.
C)nominal data.
D)all of the above

Question 16

What value below is missing from the following definition of Spearman's rank-order correlation coefficient (r)

A)
B)
C)
D)

Question 17

When nominal data are presented in a 3 x 3 cross-tabulation, the correlation is computed Using:

A)Pearson's r.
B)Spearman's r.
C)the phi coefficient.
D)the contingency coefficient.

Question 18

Cramer's V is preferable to the contingency coefficient when:

A)the table is non-square (has an unequal number of rows and columns).
B)the table is 2 x 2.
C)the table has an equal number of rows and columns.
D )there is a large sample.

Question 19

Perfect inversion indicates a perfect negative correlation expressed as:

A)
B)
C)
D)

Question 20

Which test is an alternative to the contingency coefficient?

A)The phi coefficient
B)Goodman's and Kruskal's gamma
C)Spearman's r
D)Cram er's V

Question 21

The number of variables that rs seeks to find a correlation between is __________.

Question 22

The phi coefficient ( s only used for a _____ row × _____ column size table.

Question 23

The question below refers to the following 3 x 4 contingency table: Calculate the contingency coefficient.

Question 24

In Goodman's and Kruskal's gamma (G), perfect __________ indicates a perfect negative correlation.

Question 25

Requirements for using gamma and phi include __________ sampling.

Question 26

The question below refers to the following data depicting the relationship between political affiliation and age range:
Calculate Goodman's and Kruskal's gamma.

Question 27

Compute G and a z-score to test for significance for the data below:

Question 28

The question below refers to the following data depicting the relationship between political affiliation and age range:
Calculate the sum frequency of agreements.

Question 29

Goodman's and Kruskal's gamma (G) is designed for large numbers of __________ ranks.

Question 30

The contingency coefficient (C) is used for nominal tables larger than 2 x 2.

Question 31

The question below refers to the following data depicting the relationship between political affiliation and age range:
Calculate the sum frequency of inversions.

Question 32

The phi coefficient ( is only used for __________ level data.

Question 33

For rs to determine the exact position in the case of ties, one must add the tied ranks and __________ by the number of ties.

Question 34

Calculate G using the following data:

Question 35

The question below refers to the following observations between work group size and satisfaction with work:
Using the gamma test of significance, calculate the z-score. Can we reject the null hypothesis?

Question 36

The question below refers to the following observations between work group size and satisfaction with work:
Test the association between work group size and satisfaction.

Question 37

In Goodman's and Kruskal's gamma (G), perfect agreement indicates a perfect __________ correlation.

Question 38

The contingency coefficient (C) and Cramer's V both use the calculated __________ value.

Question 39

Spearman's rank-order correlation is only used for __________ level data.

Question 40

The contingency coefficient is NOT based on chi-square.

Question 41

Work Problems
The question below refers to the following 3 x 4 contingency table:
Calculate Cramer's V.