506 Chapter 14 Repeated-Measures Designs
Exercises
14.1 It is at least part of the folklore that repeated experience with any standardized test leads
to better scores, even without any intervening study. Suppose that we obtain eight subjects
and give them a standardized admissions exam every Saturday morning for 3 weeks. The
data follow:S First Second Third
1 550 570 580
2 440 440 470
3 610 630 610
4 650 670 670
5 400 460 450
6 700 680 710
7 490 510 510
8 580 550 590a. Write the statistical model for these data.
b. Run the analysis of variance.
c. What, if anything, would you conclude about practice effects on the GRE?
14.2 Using the data from Exercise 14.1,
a. Delete the data for the third session and run a (matched-sample) t test between Sessions
1 and 2.
b. Now run a repeated-measures analysis of variance on the two columns you used in part
(a) and compare this Fwith the preceding t.
14.3 To demonstrate the practical uses of basic learning principles, a psychologist with an inter-
est in behavior modification collected data on a study designed to teach self-care skills to
severely developmentally handicapped children. An experimental group received reinforce-
ment for activities related to self-care. A second group received an equivalent amount of
attention, but no reinforcement. The children were scored (blind) by a rater on a 10-pointPartition (Introduction)
Partialling out (Introduction)
Repeated-measures designs
(Introduction)
()(Introduction)
(Introduction)
Main diagonal (14.3)
Off-diagonal elements (14.3)
Compound symmetry (14.3)
Covariance matrix ( ) (14.3)
Sphericity (14.3)
Multivariate analysis of variance
(MANOVA) (14.3)
Multivariate procedure (14.3)
(14.7)
(14.7)
Intraclass correlation (14.10)
Sequence effects (14.11)
Carryover effects (14.11)
Latin square (14.11)Randomized blocks designs (14.11)
Matched samples (14.11)
Mixed models (14.12)
Multilevel modeling (14.12)
Hierarchical modeling (14.12)
Maximum likelihood (14.12)
Restricted maximum likelihood (REML)
(14.12)
SPSS Mixed (14.12)
SAS Proc Mixed (14.12)errorwithinerrorbetweengSSwithin subj
SSbetween subj SSS
Key Terms