File: Ch17, Chapter 17: Analyzing Data using Nonparametric Statistics

True/False

1. Statistical techniques based on assumptions about the population from which the sample data
are selected are called parametric statistics.

Ans: True
Response: See section 17.0
Difficulty: Easy

2. The methods of parametric statistics can be applied to nominal or ordinal data.

Ans: False
Response: See section 17.0
Difficulty: Medium

3. Nonparametric statistical techniques are based on fewer assumptions about the population and
the parameters compared to parametric statistical techniques.

Ans: True
Response: See section 17.0
Difficulty: Easy

4. Nonparametric statistics are sometimes called distribution-dependent statistics.

Ans: False
Response: See section 17.0
Difficulty: Easy

5. An advantage of nonparametric statistics is that some nonparametric tests can be used to

analyze situations in which data are available at only the nominal or the ordinal level.

Ans: True
Response: See section 17.0
Difficulty: Hard

6. A disadvantage of nonparametric statistics is that the probability statements obtained from

most nonparametric tests are not exact probabilities.

Ans: False
Response: See section 17.0
Difficulty: Hard

7. The one-sample runs test is a nonparametric test for sequential independence

Ans: False
Response: See section 17.1 Runs Test
Difficulty: Easy

8. The one-sample runs test is a nonparametric test of randomness in the sample data.

Ans: True
Response: See section 17.1 Runs Test
Difficulty: Easy

9. In the one-sample runs test for randomness of the observations in a large sample (i.e., the
number of observations with each of two possible characteristics is greater than 20) the sampling
distribution of R, the number of runs, is approximately binomial.

Ans: False
Response: See section 17.1 Runs Test
Difficulty: Medium

10. The sampling distribution of R, the number of runs, in the one-sample runs test for
randomness of the observations in a large sample (i.e., the number of observations with each of
two possible characteristics is greater than 20) is approximately normal.

Ans: True
Response: See section 17.1 Runs Test
Difficulty: Medium

11. The appropriate test for comparing the means of two populations using ordinal-level data
from two independent samples is the Mann-Whitney U test.

Ans: True
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

12. To compare the means of two populations which cannot be assumed to be normally
distributed and only ordinal-level data is available from two independent samples, we should use
the t-test for independent samples rather than the Mann-Whitney U test.

Ans: False
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

13. To compare the means of two populations which cannot be assumed to be normally
distributed and only ordinal-level data is available from two independent samples, instead of the
t-test for independent samples we should use the Mann-Whitney U test.

Ans: True
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

14. The appropriate test for comparing the means of two populations using ordinal-level data
from two related samples is the Wilcoxon test and not the Mann-Whitney U test.

Ans: True
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

15. The nonparametric counterpart of the t test to compare the means of two independent
populations is the Mann-Whitney U test.

Ans: True
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

16. The nonparametric counterpart of the t test to compare the means of two related samples is
the Mann-Whitney U test.

Ans: False
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

17. The nonparametric alternative to the one-way analysis of variance is the Kruskal-Wallis test.

Ans: True
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Medium

18. The nonparametric alternative to analysis of variance for a randomized block design is the
Friedman test.

Ans: True
Response: See section 17.5 Friedman Test
Difficulty: Medium

19. Correlation coefficient cannot be used to analyze the association between two variables when
only ordinal-level data are available.

Ans: False
Response: See section 17.6 Spearman’s Rank Correlation
Difficulty: Medium

20 When only ordinal-level data are available, Spearman’s rank correlation rather than the
Pearson product-moment correlation coefficient must be used to analyze the association between
two variables.

Ans: True
Response: See section 17.6 Spearman’s Rank Correlation
Difficulty: Medium
Multiple Choice

21. Statistical techniques based on assumptions about the population from which the sample data
are selected are called _______.
a) population statistics
b) parametric statistics
c) nonparametric statistics
d) chi-square statistics
e) correlation statistics

Ans: b
Response: See section 17.0
Difficulty: Easy

22. The methods of parametric statistics require ________________.

a) interval or ratio data
b) nominal or ordinal data
c) large samples
d) small samples
e) qualitative data

Ans: a
Response: See section 17.0
Difficulty: Easy

23. Statistical techniques based on fewer assumptions about the population and the parameters
are called _______.
a) population statistics
b) parametric statistics
c) nonparametric statistics
d) chi-square statistics
e) correlation statistics

Ans: c
Response: See section 17.0
Difficulty: Easy
24. Nonparametric statistics are sometimes called _______________.
a) nominal statistics
b) interval statistics
c) distribution-dependent statistics
d) distribution-free statistics
e) qualitative statistics

Ans: d
Response: See section 17.0 Difficulty: Easy

25. The one-sample runs test is a ______________________.

a) nonparametric test for statistical independence
b) parametric test for statistical independence
c) nonparametric test of randomness
d) nonparametric test for correlation
e) parametric test of sequences

Ans: c
Response: See section 17.1 Runs Test
Difficulty: Medium

26. The nonparametric counterpart of the t test to compare the means of two independent
populations is the _______.
a) chi-square goodness of fit test
b) chi-square test of independence
c) Mann-Whitney U test
d) Wilcoxon test
e) Friedman test

Ans: c
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Hard

27. The nonparametric counterpart of the t test to compare the means of two related samples is
the _______.
a) chi-square goodness of fit test
b) chi-square test of independence
c) Mann-Whitney U test
d) Wilcoxon test
e) Friedman test

Ans: d
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Hard

28. The nonparametric alternative to the one-way analysis of variance is the _______.
a) chi-square goodness of fit test
b) Kruskal-Wallis test
c) Mann-Whitney U test
d) Wilcoxon test
e) Friedman test

Ans: b
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Hard

29. The nonparametric alternative to analysis of variance for a randomized block design is the
_______.
a) chi-square test
b) Kruskal-Wallis test
c) Mann-Whitney U test
d) Wilcoxon test
e) Friedman test

Ans: e
Response: See section 17.5 Friedman Test
Difficulty: Hard

30. Which of the following tests should be used to compare the means of two populations if the
samples are related?
a) Mann-Whitney test
b) Wilcoxon test
c) Runs test
d) Spearman’s test
e) Kruskal-Wallis test

Ans: b
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Hard

31. Which of the following tests should be used to compare the means of two populations if the
samples are independent?
a) Mann-Whitney test
b) Wilcoxon test
c) Runs test
d) Spearman’s test
e) Kruskal-Wallis test

Ans: a
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Hard

32. The null hypothesis for a one-sample runs test is __________________.

a) “the observations in the sample are randomly generated”
b) “the observations in the sample are not correlated”
c) “the observations in the sample are statistically independent”
d) “the observations in the sample are cross-linked”
e) “the observations are systematically generated”

Ans: a
Response: See section 17.1 Runs Test
Difficulty: Easy

33. The alternate hypothesis for a one-sample runs test is __________________.

a) “the observations in the sample are not cross-linked”
b) “the observations in the sample are correlated”
c) “the observations in the sample are not statistically independent”
d) “the observations in the sample are not randomly generated”
e) “the observations are not systematically generated”

Ans: d
Response: See section 17.1 Runs Test
Difficulty: Easy

34. Charles Clayton monitors the daily performance of his investment portfolio by recording a
“+” or a “-“sign to indicate whether the portfolio’s value increased or decreased from the
previous day. His record for the last eighteen business days is “- + + - - - + - - + + + - + + + + -“.
The number of runs in this sample is _________.
a) uncertain
b) four
c) five
d) nine
e) one

Ans: d
Response: See section 17.1 Runs Test
Difficulty: Easy

35. Charles Clayton monitors the daily performance of his investment portfolio by recording a
“+” or a “-“sign to indicate whether the portfolio’s value increased or decreased from the
previous day. His record for the last eighteen business days is “- + + - - - + - - + + + + + + + +
-“. The number of runs in this sample is _________.
a) seven
b) six
c) four
d) three
e) one

Ans: a
Response: See section 17.1 Runs Test
Difficulty: Easy

36. Which of the following tests should be used to compare the means of three populations if the
sample data is ordinal?
a) one-way analysis of variance
b) Kruskal-Wallis test
c) Wilcoxon test
d) Mann-Whitney test
e) Friedman test

Ans: b
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Medium

37. In a Mann-Whitney U test, U statistic was calculated to be 38.78 based on sample sizes of 24
and 20. What is the z value for this test?
a) 0.133
b) -4.74
c) 240
d) 42.43
e) 8.75

Ans: b
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

38. In a Mann-Whitney U test, the U statistic was calculated to be 58.0 based on sample sizes of
22 and 28. What is the z value for this test?
a) 51.17
b) 308
c) 0.117
d) -4.88
e) -2.44

Ans: d
Response: See section 17.2 Mann-Whitney U Test
Difficulty: Medium

39. The Wilcoxon test was used on 18 pairs of data. The total of the ranks (T) were computed to
be 111 (for + ranks) and 60 (for - ranks). The z value for this test is ____.
a) -1.11
b) -0.05
c) -0.07
d) 0.033
e) 2.22

Ans: a
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

40. The Wilcoxon test was used on 16 pairs of data. The total of the ranks (T) were computed to
be 76 (for + ranks) and 60 (for - ranks). The z value for this test is _____.
a) -0.41
b) -0.02
c) 0.02
d) 16
e) -0.041.

Ans: a
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

41. In the Wilcoxon test of the differences between two populations, the value z statistic was
calculated to be 1.80. If the level of significance is 0.05, which of the following decisions is
appropriate?
a) Reject the null hypothesis
b) Do not reject the null hypothesis
c) Indeterminate without the sample size
d) Indeterminate without all of the data
e) Inconclusive

Ans: b
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

42. In the Wilcoxon test of the differences between two populations, the value z statistic was
calculated to be 1.80. If the level of significance is 0.10, which of the following decisions is
appropriate?
a) Reject the null hypothesis
b) Do not reject the null hypothesis
c) Indeterminate without the sample size
d) Indeterminate without all of the data
e) Inconclusive

Ans: a
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

43. Most "Before and after" types of experiments should be analyzed using _______.
a) chi-square goodness of fit test
b) Kruskal-Wallis test
c) Mann-Whitney U test
d) Wilcoxon test
e) Friedman test

Ans: d
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

44. In a Wilcoxon matched-pairs signed rank test with 20 matched-pairs of observations, the
observed value of the T statistic based on sample data is 76.33. The corresponding observed z-
value is ___________.
a) −1.79
b) −2.07
c) −1.70
d) −1.59
e) −1.07

Ans: e
Response: See section 17.3 Wilcoxon Matched-Pairs Signed Rank Test
Difficulty: Medium

45. The Kruskal-Wallis test is to be used to determine whether there is a significant difference
(alpha = 0.05) between the three groups using the following data:

Group 1
Group 2
Group 3

For this test, how many degrees of freedom should be used?

a) 3
b) 2
c) 4
d) 8
e) 1

Ans: b
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Easy

46. The Kruskal-Wallis test is to be used to determine whether there is a significant difference
(alpha = 0.05) between the three groups using the following data:
For this situation, the critical (table) chi-square value is _______.
a) 15.507
b) 7.815
c) 9.488
d) 5.991
e) 3.991

Ans: d
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Medium

47. The null hypothesis in the Kruskal-Wallis test is _______.

a) all populations are identical
b) all sample means are different
c) x and y are not correlated
d) the mean difference is zero
e) all populations are not identical

Ans: a
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Medium

48. The Spearman correlation coefficient is calculated for a set of data on two variables, x and y.
It appears that as the rank of x increases, the rank of y is decreasing. We would expect the
Spearman correlation coefficient to be ________.
a) equal to zero
b) positive
c) negative
d) greater than 5
e) greater than 1

Ans: c
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Easy

49. Correlation coefficients may be computed for parametric and nonparametric data. If the data
are nonparametric, which of the following should be used?
a) Pearson correlation coefficient
b) Spearman correlation coefficient
c) Gaussian correlation coefficient
d) De Moivre correlation coefficient
e) Gossett correlation coefficient

Ans: b
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Medium

50. Correlation coefficients may be computed for parametric and nonparametric data. If the data
are interval data, which of the following should be used?
a) Pearson correlation coefficient
b) Spearman correlation coefficient
c) Gaussian correlation coefficient
d) De Moivre correlation coefficient
e) Gossett correlation coefficient

Ans: a
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Medium

51. What is the Spearman rank correlation coefficient for the following set of data?

a) -10.2
b) -2.35
c) 0.65
d) 0.50
e) 0.05

Ans: d
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Hard

52. What is the Spearman rank correlation coefficient for the following set of data?

a) -0.20
b) 1.00
c) 0.20
d) 0.80
e) -1.20

Ans: d
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Hard

53. Personnel specialist, Steve Satterfield, is assessing a new supervisor's ability to follow
company standards for evaluating employees. Steve has the new supervisor rate five
hypothetical employees on a scale of one to ten. He is interested in how the new supervisor's
ratings correlate with company norms for these benchmark cases.

Employee
1 2 3 4 5
New Supervisor 8 8 9 7 5
Company Norm 8 6 10 4 4

The Spearman rank correlation coefficient is ___________.

a) 0.80
b) 0.85
c) 0.90
d) 0.95
e) 1.00

Ans: d
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Hard

54. Two stock analysts rank five investment portfolios for overall performance and risk.

Portfolio
A B C D E
Broker 1 - Rankings 4 5 2 1 3
Broker 2 - Rankings 1 4 3 2 5

The Spearman rank correlation coefficient is ___________.

a) 0.80
b) 0.20
c) 0.05
d) 0.95
e) 1.00

Ans: b
Response: See section 17.6 Spearman’s Rank Coefficient
Difficulty: Hard

55. A Kruskal-Wallis test is to be performed. There will be four categories, and alpha is chosen
to be 0.10. The critical chi-square value is _______.
a) 6.251
b) 2.706
c) 7.779
d) 4.605
e) 3.234

Ans: a
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Easy

56. A Kruskal-Wallis test is to be performed. There will be five categories, and alpha is chosen
to be 0.01. The critical chi-square value is _______.
a) 15.086
b) 13.277
c) 7.779
d) 9.236
e) 8.987

Ans: b
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Easy

57. Performance records for 18 salespersons are selected to investigate whether compensation
methods are a significant motivational factor.

Compensation Method Sales

Straight Salary 18 12 22 28 28
Straight Commission 27 34 34 27 20 16 24
Salary plus Commission 11 17 27 14 30 22
A Kruskal-Wallis test is to be performed with a= 0.01. The null hypothesis is _______.
a) Group 1 = Group 2 = Group 3
b) Group 1  Group 2  Group 3
c) Group 1  Group 2  Group 3
d) Group 1  Group 2  Group 3
e) Group 1  Group 2 ≥ Group 3

Ans: a
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Easy

58. Performance records for 18 salespersons are selected to investigate whether compensation
methods are a significant motivational factor.
Compensation Method Sales
Straight Salary 18 12 22 28 28
Straight Commission 27 34 34 27 20 16 24
Salary plus Commission 11 17 27 14 30 22
A Kruskal-Wallis test is to be performed with a = 0.01. The critical chi-square value is
_______.
a) 15.086
b) 13.277
c) 7.779
d) 9.210
e) 8.657

Ans: d
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Medium

59. Performance records for 18 salespersons are selected to investigate whether compensation
methods are a significant motivational factor.

Compensation Method Sales

Straight Salary 18 12 22 28 28
Straight Commission 27 34 34 27 20 16 24
Salary plus Commission 11 17 27 14 30 22

A Kruskal-Wallis test is to be performed with a = 0.01. The calculated K value is _______.

a) 15.086
b) 1.1.715
c) 7.779
d) 9.210
e) 8.657
Ans: b
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Hard

60. Performance records for 18 salespersons are selected to investigate whether compensation
methods are a significant motivational factor.

Compensation Method Sales

Straight Salary 18 12 22 28 28
Straight Commission 27 34 34 27 20 16 24
Salary plus Commission 11 17 27 14 30 22

A Kruskal-Wallis test performed with a = 0.01 will result in a decision to _____.

a) reject the null hypothesis
b) reject the alternate hypothesis
c) do not reject the null hypothesis
d) do no reject the alternate hypothesis
e) do nothing

Ans: c
Response: See section 17.4 Kruskal-Wallis Test
Difficulty: Easy

61. Which of the following is NOT an assumption of the Friedman test?

a) The blocks are independent.
b) The population has a normal distribution.
c) There is no interaction between blocks and treatments.
d) Observations within each block can be ranked.

Ans: b
Response: See section 17.5 Friedman Test
Difficulty: Medium

62. In the Friedman test of c treatment levels, the degrees of freedom will be _____.
a) n - 1
b) (r – 1)(c – 1)
c) c - 1
d) b – k - 1
e) c + 1
Ans: c
Response: See section 17.5 Friedman Test
Difficulty: Medium

63. In a Friedman test with 8 treatment levels and 5 blocks, df = _________.

a) 28
b) 40
c) 8
d) 7
e) 6

Ans: d
Response: See section 17.5 Friedman Test
Difficulty: Medium

64. In a Friedman test with 4 treatment levels and 5 blocks, the df = ________.
a) 2
b) 3
c) 4
d) 5
e) 6

Ans: b
Response: See section 17.5 Friedman Test
Difficulty: Medium

65. In a Friedman test with 7 treatment levels and 5 blocks, df = _________.

a) 16
b) 34
c) 4
d) 35
e) 6

Ans: e
Response: See section 17.5 Friedman Test
Difficulty: Medium
66. In a Friedman test with 6 treatment levels and 7 blocks, the df = ________.
a) 6
b) 5
c) 42
d) 41
e) 7

Ans: b
Response: See section 17.5 Friedman Test
Difficulty: Medium

67. A randomized block design has 7 blocking levels and 5 treatment levels. Given the level of
data, a nonparametric alternative to a randomized block design must be used to analyze the data.
The number of degrees of freedom for the appropriate test statistic in this test of the null
hypothesis that the treatment populations are equal is ___________.
a) 35
b) 34
c) 24
d) 6
e) 4

Ans: e
Response: See section 17.5 Friedman Test
Difficulty: Medium

68. A randomized block design has 7 blocking levels and 5 treatment levels. Given the level of
data, a nonparametric alternative to a randomized block design must be used to analyze the data.
For the test of the null hypothesis that the treatment populations are equal at a 1% level of
significance, the critical value of the test statistic is ___________.
a) 16.81
b) 15.09
c) 13.28
d) 9.49
e) 3.48

Ans: c
Response: See section 17.55 Friedman Test
Difficulty: Medium
69. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels, the df = ________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 6
b) 5
c) 4
d) 3
e) 1

Ans: d
Response: See section 17.5 Friedman Test
Difficulty: Medium

70. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels with α = 0.5, the critical value of chi-square is
________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 7.81
b) 5.99
c) 11.07
d) 9.49
e) 6.81

Ans: a
Response: See section 17.5 Friedman Test
Difficulty: Medium
71. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels with α = 0.01, the critical value of chi-square is
________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 13.28
b) 11.34
c) 15.09
d) 9.21
e) 9.18

Ans: b
Response: See section 17.5 Friedman Test
Difficulty: Medium

72. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels, the ranks for block 1 are ________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 2, 3, 4, 1
b) 3, 4, 1, 2
c) 4, 1, 2, 3
d) 4, 2, 3, 1
e) 1, 2, 3, 4

Ans: c
Response: See section 17.5 Friedman Test
Difficulty: Medium
73. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels, the ranks for block 2 are ________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 2, 3, 4, 1
b) 1, 2, 3, 4
c) 4, 1, 2, 3
d) 4, 2, 3, 1
e) 3, 4, 1, 2

Ans: e
Response: See section 17.5 Friedman Test
Difficulty: Medium

74. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels with α = 0.5, the calculated value of chi-square is
________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) 4.51
b) 5.27
c) 2.83
d) 3.48
e) 1.48

Ans: d
Response: See section 17.5 Friedman Test
Difficulty: Hard
75. When using the Friedman test to test the following data to determine whether there is a
significant difference between treatment levels with α = .05, the appropriate conclusion is
________.

Bloc Treatment 1 Treatment 2 Treatment 3 Treatment 4

k
1 97 55 79 80
2 94 98 50 71
3 93 57 62 53
4 52 58 61 51
5 64 67 77 63
a) reject H0: the block populations are equal
b) do not reject H0: the block populations are equal
c) reject H0: the treatment populations are equal
d) do not reject H0: the treatment populations are equal
e) that the test is inconclusive

Ans: d
Response: See section 17.5 Friedman Test
Difficulty: Hard