Deck 14: Nonparametric Statistics

Question 1

A student answered 50 of 80 true/false questions correctly. Does this performance indicate that the student knew something, or was he just lucky? (Hint: Use chi-squared and α = .05.)

A) Chi-square =2.5; he knew something.
B) Chi-square =2.5; he was lucky.
C) Chi-square =5.0; he knew something.
D) Chi-square =5.0; he was lucky.

Answer 1

C

Question 2

What is the expected value for the upper left-hand cell of the following 2x2 table? $\begin{array}{ll}15 & 21 \\25 & 20\end{array}$

A) 18.0
B) 17.8
C) 15.0
D) 8.3

Answer 2

17.8

Question 3

What is the nonparametric equivalent to a repeated-measures ANOVA?

A) Chi-square
B) Friedman two-way by ranks
C) Kruskal-Wallis
D) Mann-Whitney U

Answer 3

B

Question 4

If N is 20 or greater, which table would you use to determine the significance of a Mann-Whitney U?

A) Chi-square table
B) Mann-Whitney table
C) Normal distribution table
D) t-test table

Question 5

A basketball coach taught three groups of players to shoot free throws, using a different method for each group. After six weeks of practice, she ranked all of the players according to their ability to make free throws. Which test should she use to analyze these data?

A) Mann-Whitney U
B) Kruskal-Wallis
C) Friedman two-way by ranks
D) Chi-square

Question 6

A researcher had 25 subjects rank four types of workouts (running, swimming, biking, and aerobic dance) from hardest (1) to easiest (4). Which test should he use to determine if there is a significant difference among the ranking of the difficulty of the types of workouts?

A) Friedman two-way by ranks
B) ANOVA
C) Spearman rank order correlation
D) Mann-Whitney U

Question 7

Which inferential procedure is appropriate when a study has one dependent variable measured on a nominal level and one independent variable measured on a nominal level?

A) Chi-square
B) t-test
C) One-way ANOVA
D) Factor analysis

Question 8

What is the most important "conclusion-making" piece of information reported as the result of a chi-square, t-test, or ANOVA?

A) The null hypothesis
B) The independent variable(s)
C) The dependent variable(s)
D) The probability
E) The absolute size of the difference between measured values

Question 9

What do we use to calculate the expected cell frequencies when applying chi-square to a data table to test independence?

A) Marginal totals
B) Binomial distribution
C) Experimental sampling distribution of chi-squared
D) Normal distribution

Question 10

Given that the number of winners from each starting position at the racetrack was as follows, what chi-square value is obtained? Does a relationship exist between winning and starting position? $\begin{array}{cc}\text { Position}& \text {Winners }\\1 & 35 \\2 & 30 \\3 & 25 \\4 & 20 \\5 & 15\\&125 \text { races }\end{array}$

A) 6, no
B) 8, no
C) 8, yes
D) 10, yes

Question 11

When would the use of a nonparametric statistical test be most appropriate?

A) When only interval-level measurements are available
B) When an increase in power is desired
C) When the Ns are very small
D) When normality and equal variances are not problematic

Question 12

What nonparametric statistical test is analogous to a one-way ANOVA with four levels of the independent variable?

A) Friedman two-way ANOVA
B) Phi coefficient
C) Kruskal-Wallis
D) Wilcoxon

Question 13

The percentages of scores falling below the following z-score values for a particular raw score distribution are:
$\begin{array}{ll}z \text {-score of }-2.04 & 5 \% \\z \text {-score of }-1.04 & 20 \% \\\text { z-score of 0 } & 55 \% \\\text { z-score of } 1.04 & 90 \% \\\text { z-score of } 2.04 & 100 \%\end{array}$ Does this distribution differ significantly from a normal distribution? (Use ? = .05.)

A) No, statistic = 11.8
B) No, statistic = 6.7
C) Yes, statistic = 11.8
D) Yes, statistic = 6.7

Question 14

If the level of measurement for the dependent variable was ordinal and a nonparametric statistical test resulted in retaining the null hypothesis, which conclusion is LEAST defensible?

A) There is no significant difference between the sample statistics.
B) A type II error has occurred.
C) The samples come from populations with the same mean.
D) The dependent variable is not normally distributed in the population.

Question 15

When is a nonparametric test of significance most appropriate?

A) When the population scores are normally distributed
B) When group variances are about equal
C) When the scores are derived from a measurement with equal intervals
D) When the scores are in the form of ranks

Question 16

Which of the following nonparametric tests is NOT used to assess association among variables?

A) Phi coefficient
B) Kappa coefficient of agreement
C) Wilcoxon signed rank test
D) Kendall coefficient of concordance

Question 17

Which nonparametric test would be the most powerful for testing for possible significant differences between two correlated samples?

A) Median test
B) Mann-Whitney U test
C) Sign test
D) Wilcoxon signed rank test

Question 18

Which nonparametric statistical test is appropriate for a repeated-measure one-way ANOVA, if the dependent variable is measured at a dichotomous level?

A) Chi-square
B) Kruskal-Wallis
C) Cochran's Q
D) Friedman two-way ANOVA

Question 19

The Spearman rank order coefficient is MOST like which of the following?

A) Percent agreement
B) Kappa
C) PPM
D) Chi-square

Question 20

Which of the following is LEAST like the others?

A) Chi-square
B) Phi coefficient
C) Spearman rank
D) Kruskal-Wallis