Quiz (Chapter 18)
Statistical Tests on Ranks (Nonparametric
- The simplest kind of quantitative data involve _____.
- Test procedures such as the median test and the Mann-Whitney U test
are called ______ tests.
Obtaining Ranked Data
- In how many different ways can a researcher "produce" data that take
the form of ranks?
- In Excerpt 18.2, what data were used in computing the Spearman correlation?
- The 1st of the 2 numbers recorded by the 2nd researcher.
- The 2nd of the 2 numbers recorded by the 2nd researcher.
Reason for Converting Scores on a Continuous Variable into Ranks
- Test procedures designed for use with ranks involve ____ (fewer/more)
assumptions than procedures designed for use with interval- or ratio-level
- Unequal or small sample sizes prompt many researchers to ____ (use/not
use) tests based on ranks.
- (T/F) The total scores people earn on a Likert-type attitude inventory
are probably only ordinal in nature.
The Median Test
- The median test is used to compare 2 or more groups that are _____
- In comparing 3 groups with a median test, how many medians would
appear in the test's null hypothesis?
- In comparing 3 groups with a median test, how many medians need to
be computed from the sample data?
- The calculated value in a median test is referred to as ___.
- (T/F) If, in Excerpt 18.10, group C's median healing time had been
12 weeks (rather than 14 weeks), the null hypothesis might have been rejected.
The Mann-Whitney U Test
- The Mann-Whitney U test test is used to compare 2 groups that are
- In the study associated with Excerpt 18.11, the sums of ranks for the 2
institutions were quite _____ (similar/different).
- The task of interpreting a significant result from a Mann-Whitney
U test is _____ (easy/difficult).
- (T/F) If the U test leads to a rejection of Ho,
the populations might differ in their means, medians, or modes.
- There is a ______ (direct/indirect) relationship between the size of U, the
calculated value, and the size of p.
The Kruskal-Wallis H Test
- The Kruskal-Wallis H test is used to compare 3 or more groups that
are _____ (independent/correlated).
- (T/F) In both the Kruskal-Wallis and Mann-Whitney tests, a "sum of
ranks" is computed for each sample.
- The information in Excerpt 18.12 shows that there is a(n) _____ (direct/indirect)
relationship between the size of H and the size of p.
- In a Kruskal-Wallis test, Ho is rejected if the calculated
H is ____ (smaller/larger) than the critical value.
- In Excerpt 18.14, how many Mann-Whitney U tests were conducted in the post hoc
investigation following the significant Kruskal-Wallis result?
The Wilcoxon Matched-Pairs Signed-Ranks Test
- Wilcoxon's test is used to compare __ (how many?) sets of scores
that are _____ (independent/correlated).
- The calculated value in a Wilcoxon matched-pairs signed-ranks test
is referred to by what capital letter?
- In Excerpt 18.15, what would have happened to T and p
if the two sums of ranks had been more similar?
- Both T and p would have been smaller.
- T would have been larger but p would have been smaller.
- T would have been smaller but p would have been larger.
- Both T and p would have been larger.
- In a Wilcoxon test, Ho is rejected if the calculated value
is ____ (smaller/larger) than the critical value.
- A significant T suggests that the 2 populations differ in what way,
assuming they have the same shape?
Friedman's Two-Way Analysis of Variance of Ranks
- Friedman's test is used to compare __ (how many?) sets of scores
that are _____ (independent/correlated).
- Which parametric ANOVA is analogous to the Friedman two-way analysis
of variance of ranks?
- If each of 3 people tastes 3 soft drinks and then ranks them, what
should you expect the "sum of ranks" to be for any of the beverages
if the null hypothesis (that the 3 beverages are equally good) is true?
- Read Excerpt 18.16 and the paragraph above it. Based on these
descriptions of the study, how many sets of ranks were there?
- It's impossible to tell based on the information provided.
- If the Friedman test yields a significant result, what test is likely
to be used in a post hoc investigation?
Large-Sample Versions of the Tests on Ranks; Ties
- (T/F) A test on ranks made via a z-test (due to large samples) can
be thought of as a "normal curve test."
- The _____ test can be thought of as the "small-sample" version of
the median test.
- In Excerpt 18.17, which number(s) allow us to know that data came from more
women that men?
- The 3 means (12, 16, and 5)
- The standard deviation (21)
- The calculated value, z (-7.15)
- The probability, p, associated with z (<.001).
- The regular median test is inherently a ______ (small-sample/large-sample)
- (T/F) When ties occur in tests on ranks, the only thing a researcher
can do is discard the tied scores.
A Few Final Comments
- Nonparametric tests ___ (are/are not) based on the assumption that
the data come from random samples.
- In Excerpt 18.11, the means and standard deviations for the two samples
____ (do/don't) suggest the presence of "overlapping distributions."
- The last sentence in the 2nd paragraph on p. 454 urges you not to
use _____ as the criterion for judging the worth of empirical studies.
- how large or small the sample size is
- how "impressive" the data-based p-level is
- what alpha level was used
- whether Ho is or isn't rejected
Click here for answers.