P554 Statistics in Psych, Prof. Kruschke
P554 Statistics in Psychology,
Prof. Kruschke
Homework for Ch. 11. Due at beginning of class, Tu 3 Apr 2007.
General instructions. Please write your full name
at the top of every page you hand in. Please collate and staple
your pages together. Please write clearly and thoroughly explain all
your computations; an unannotated series of formulas that mysteriously
ends up with the correct number will not be given full credit. When
doing the homework, you are encouraged to use all resources at your
disposal to the extent that they help you learn the material;
nevertheless, you must write your own answers in your own words.
- (8 pts.)
Load into SPSS the data from Table
11.5 of the textbook.
- Do an omnibus test on the age factor using repeated measures
ANOVA and the Greenhouse-Geisser correction to interpret
significance. Include the ANOVA table and explain what part of the
table you used to make your conclusion. (You did this in lab; just do
it again here.)
- To every score from subject 2, add 1000. From every score from
subject 12, subtract 100. Include a print out of the resulting data
table. Do an omnibus test on the age factor using repeated measures
ANOVA and the Greenhouse-Geisser correction to interpret
significance. How does the result compare to the previous part?
- Explain why adding a constant to every score from a subject
makes no difference to a within-subject ANOVA.
- (12 pts.)
Load into SPSS the data from #19, p. 571. Read
the paragraph that explains the experiment from which these data were
generated.
- Make a boxplot of the data from the four conditions. Are there
outliers?
- Make a graph of the means from the four conditions.
- Run an omnibus repeated-measures ANOVA, using the
Greenhouse-Geisser correction for non-sphericity. Include the table
and your conclusion.
- The experiment was run using infants, who often are impatient
during experiments and cannot be run through several conditions
because they fuss and cry. So it might be necessary to run the
experiment using a between-subject design. Analyze the data as if they
were from a between subject design. Compare the results with the
repeated-measures analysis.
- It is often difficult to get large numbers of infants to
participate in an experiment; there are only so many babies in a
community with parents who are willing to go to the effort to
participate in an experiment. So there is a big incentive to use
within-subject designs if possible, to reduce the number of subjects
needed. Briefly discuss how the treatment effects themselves in this
experiment might differ when the conditions are run between-subject or
within-subject. As a separate issue, also discuss possible
differential carryover effects for the specific conditions in a
within-subject design.