Quiz Over 2nd Half of
Chapter 13 (pp. 292-310) of the
Two-Way Analyses of Variance
NOTE: A different quiz covers the 1st half of Chapter 13 (pages 276-292.)
- (T/F) If none of the two-way ANOVA's Fs turns out to be significant,
no follow-up tests will be conducted.
- Follow-up tests can be used to "probe" significant main effects ___
(and/but not) significant interactions.
Follow-Up Tests to Probe Significant Main Effects
- (T/F) In Excerpt 13.9, the significant main effect of vignette required no
follow-up investigation because there were only 2 levels of the vignette factor.
- In Excerpt 13.10, how many Tukey pairwise tests were conducted?
- If post hoc tests are used to probe the significant main effects
of a 2x3 ANOVA, how many null hypotheses will be tested?
Follow-Up Tests to Probe a Significant Interaction
- (T/F) Researchers usually refrain from interpreting the main effect
F-ratios if the interaction is significant.
- In the middle portion of Figure 13.1, which Source(s) would be "ns"
if the cell mean of 5 is changed to 35?
- In a graph of a cell means, the ordinate and abscissa are usually
labeled with the names of ______ & ______.
- If statistically significant interaction is graphed, the lines in
the graph ___ (will/won't) be parallel.
- Comparing cell means that lie in individual rows and/or individual
columns is a test of _____ _____ _____.
- In Excerpt 13.11, there would have been less interaction if the High/NC
group had ended up with a mean equal to
- If the interaction from Excerpt 13.12 were to be displayed in a graph like that
shown in Excerpt 13.11, how many lines would be in the graph?
- In Excerpt 13.12, the means involved in each Tukey test that was
conducted were ______ (cell/main effect) means.
- If the follow-up strategy used Excerpt 13.13 had been used in Excerpt
13.12, a total of __ pairwise comparisons would have been made.
- In two-way ANOVAs, planned comparisons ___ (can/can't) be done if
the interaction is not significant.
- In Excerpt 13.14, the planned comparison was _____ (pairwise/nonpairwise) in nature.
Assumptions Associated With a Two-Way ANOVA
- (T/F) The assumptions of a two-way ANOVA are the same as those associated
with a one-way ANOVA.
- Violations of the equal variance assumptions of a two-way ANOVA are
more disruptive when the __'s differ.
- (T/F) In testing the normality and equal variance assumptions, researchers
hope for nonsignificant results.
- Hartley's F-max test focuses on which assumption, normality or
homogeneity of variance?
- If the normality or equal variance assumption seems untenable, the
researcher can either ____ or ____.
- The term "____" means that a statistical tests functions as it should
even if the assumptions are violated.
- An F-test is considered to be ____ (positively/negatively) biased
if it turns out larger than it should.
Estimating Effect Size and Conducting Power Analyses in Two-Way ANOVAs
- Two-way ANOVAs ___ (can/can't) yield results that have statistical
but no practical significance.
- (T/F) Excerpt 13.18 illustrates the fact that a single index of "practical significance"
is computed for any two-way ANOVA, not individual indices for each main effect and
- It's better to do a power analysis ____ (before/after) the data are
collected and analyzed.
- (T/F) A strength-of-effect index can be computed for each main effect
F and for the interaction F.
- The two most commonly used strength-of-effect measures are called
_____ and _____.
- Most researchers ___ (do/don't) distinguish between statistical and
practical significance in 2-way ANOVAs.
The Inflated Type I Error Rate in Factorial ANOVAs
- (T/F) Since a 2-way ANOVA with 1 dependent variable yields 3 F-tests,
most researchers divide a by 3.
- If a researcher sets a equal to .05 when
evaluating each of the 3 F-tests, the "______" error rate is equal to
- The chances of a _____ (Type I/Type II) error will be increased if a researcher
conducts 2 or more two-way ANOVAs without using the Bonferroni adjustment
procedure (or some other procedure) to adjust alpha.
A Few Warnings Concerning Two-Way ANOVAs
- (T/F) If the research questions are not worth investigating, there's
a fatal flaw in the study than cannot be salvaged by a two-way
ANOVA . . . even if all assumptions are met and all analyses are properly
- In Excerpt 13.9, the researchers report that "... the participants who read
the weight-related teasing vignette reported significantly more negative affect
(M = 15.75, SD = 4.99) than did those who read the neutral vignette
(M = 13.87, SD = 3.42)." Do the reported SDs suggest the
presence of "overlapping distributions"?
- (T/F) Because a two-way ANOVA is so powerful, its F-tests never lead
to either Type I or Type II errors.
- (T/F) If the main effect means for one of the factors in a
two-way ANOVA turn out to be identical, the null hypothesis associated
with that particular factor should be considered to be true (i.e., fully
Click here for answers.
NOTE: Questions dealing with
the 1st half of Chapter 13 appear on a separate quiz.