Study Questions for Quiz #3
1. Describe the dimensions/attributes used to organize
the variety of single-factor research designs and explain under what conditions
results from which of the resulting designs may be causally interpreted.
2. Describe the advantages of multiple condition designs,
the different kinds of conditions that may be included and how a researcher
decides which conditions to include in a research project.
3. Tell the
different types of variables involved in a design and how one decides the role
of a variable in a particular study.
4. Differentiate between ways of controlling
subject and procedural variables and tell why “equivalent on the average” is
sufficient to allow causal interpretability of research results.
5. What is the H0: for ANOVA when applied to a multiple
condition design? What is the limitation
of this procedure, and what do we do to accompany this initial test? How do we choose which follow-up procedure to
use?
6. What is "alpha inflation" and how does one
calculate the extent of it for a given set of ANOVA follow-up analyses?
Describe the three ways to assess Experiment-wise Type I error. Which of these do you prefer? Explain your answer carefully.
7. Be prepared to describe each of the following pairwise
comparison methods, or to compare and contrast any pair.
Fisher's LSD Fisher's
protected t-test Tukey's
HSD Scheffe's
Bonferronni/Dunn's
Analytic Comparisons & Trend Analyses
8. Distinguish between simple comparisons, complex
comparisons, trend analyses and pairwise comparisons and tell when the results
from each of these can be causally interpreted.
9. Distinguish the
information obtained from a trend analysis versus from a series of pairwise
comparisons and give a carefully explained example of each being used when the
other should have been.
k-group Confidence Intervals & Power Analyses
10Describe the different effect size estimates and how
each can be applied to omnibus and various follow-up analyses. (Be sure to include the differences of the
treatment of between and within-groups models).
11. Describe how to define S, N & n when using Friedman’s
power table for BG & WG k-group designs
12. Describe how to combine the information from pairwise
NHST, CIs and effect size estimates to better
understand your results.
BG & WG ANOVA
13. Describe “total variation” and explain how it is
portioned differently by BG and WG ANOVA models. Describe how the F-test is computed differently
for these two types of ANOVA and explain the effect this has on the F-value and
effect size for the two models.
14, Tell how the “more realistic” ANOVA model partitions
“total variation” for BG and WG ANOVA models.
Why is this called “more realistic” and what
warnings does it give us about applying ANOVA to our research data.
15, Describe how we can use the “more realistic” ANOVA model to
anticipate how certain aspects of the research design will influence F.
k-group X²
16. What is the H0: for X² when applied to a multiple
condition design? What is the limitation
of this procedure, and what do we do to accompany this initial test? How do we choose which follow-up procedure to
use?
17. What is
"alpha inflation" and how does one calculate the extent of it for a given
set of X² follow-up analyses? Describe the three ways to assess Experiment-wise
Type I error. Which of these do you
prefer? Explain your answer carefully.