Quiz (Chapter 17)
Inferences on Percentages, Proportions,
- If concerned with group membership, researchers compute summaries
such as frequencies, _____, or _____.
- What level of measurement is contained in the data if a researcher
measures people in terms of gender?
The Sign Test, the Binomial Test, and Fisher's Exact Test
- If, in Excerpt 17.2, a reduced caesarean section rate had been observed
in 12 of the intervention hospitals (rather than 13 of them), the
Ho most likely _____ (would/wouldn't)
have been rejected.
- Can a sign test be used with 1 group measured twice? With 2 matched
groups? With 2 independent groups?
- The binomial test is exactly like a sign test, except the _____ (former/latter)
has a more "flexible" Ho.
- Answer the 3 parts of Question 4, but this time think of the binomial
test rather than the sign test.
- How many comparison groups are involved in a Fisher's Exact Test?
Must these groups be matched?
- The null hypothesis in a Fisher's Exact Test says that the population
proportions are ____ (equal/equal to .50).
- In Excerpt 17.4, the sample data would have been most "in line with" the
null hypothesis if ___ of the children from professional families had shown
low VIQ scores.
Chi-Square Tests: An Introduction
- (T/F) Most chi-square tests (but not all of them) deal with nominal
- What Greek symbol is used to indicate that a chi-square tests has
- (T/F) If a chi-square test is used with frequencies, it most likely
is Pearson's approximation to chi square.
Three Main Types of Chi-Square Tests
- (T/F) A 4-category 1-sample c2
test could yield a "p<.05" result if 25% of the sample falls in each
- If 30 men in the male sample exercise but 10 don't, how many of the
60 women in the female sample would have to exercise in order for the
sample data to be in full agreement with the chi-square null hypothesis?
- In Excerpt 17.10, the df of the chi square test ___ (was/wasn't)
influenced by how many of the respondents in each group were or were not interested
in serving children with disabilities.
- (T/F) When chi square is used with a contingency table, the researcher
must specify the values of P in Ho.
- In Excerpt 17.11, the sample data would have been ____ (more/less) "in line with"
the null hypothesis if there had been 8 girls and 13 boys in the music group.
- (T/F) In Excerpt 17.12, the null hypothesis stated that each of
the gender group populations had one-third of its members in each of
the three drinking categories.
Chi Square as a Correlational Probe
- When chi-square is used in a correlational sense, researchers sometimes
call it a "chi square test of ____."
- For 2x2 contingency tables, chi-square can be converted into a correlational
index called ____.
- For contingency tables with 3 or more rows/columns, c2
can be converted into what correlational index?
- In Excerpt 17.17, if the calculated value for chi square had turned out larger than 70.94, the numerical value of phi
would have been ____ (larger/smaller) than .57.
Issues Related to Chi-Square Tests
- With chi-square tests, the df are determined by ____. (a) the sample
size (b) the number of categories
- If a researcher used chi-square to see if a relationship exists between
the political affiliation of bumper stickers (Republican vs. Democrat)
and driver's age (<30, 30-50, >50), the df would be equal to __.
- If a chi-square test comparing the % of men and women getting one
or more traffic tickets in the past year leads to a significant result,
a post hoc test ____ (would/would not) be needed to clarify the results.
- Chi-square should not be used if the _____ (expected/observed) frequencies
are too small.
- Researchers sometimes use the "Yates correction" when their chi square's
df is ___ (1/more than 1).
McNemar's Chi Square
- McNemar's chi square is appropriate when the 2 samples of data are
- McNemar's chi-square test is very much like a
- sign test
- paired t-test
- two-way ANOVA
- hierarchical multiple regression
- McNemar's chi-square test ____ (can/can't) be used with matched
The Cochran Q Test
- Cochran's test is used when there are ___ (2/more than 2) _____ (independent/correlated)
samples giving data on a dependent variable having ___ (2/more than
- If Cochran's test leads to a rejection of the omnibus null hypothesis,
what test will probably be used to make pairwise post hoc comparisons?
The Use of z-Tests When Dealing With Proportions
- Researchers sometimes use z-tests rather than sign, binomial, and
McNemar tests when:
- underlying assumptions are not met
- the null hypothesis can't be rejected
- the sample sizes are big
A Few Final Thoughts
- (T/F) Statistical tests that focus on frequencies, percentages, or proportions
have a built-in feature that prevents Type I errors from occurring.
- The Bonferroni technique ___ (can/can't) be used with tests on frequencies,
percentages, and proportions.
- To judge whether a chi-square-based effect is small, medium, or
large, researchers often convert their chi-square calculated value
into a ___ or ___ correlation.
- In Excerpt 17.29, the researchers conducted a(n) _______ (a priori/post hoc)
Click here for answers.