248

a A o o A A nn

8.2

SPLIT-PLOT DESIGN

(1963) discuss in detail the problems resulting from unegual correlations
among levels of the repeated treatment. Subseguent sections descrihe
suggested procedures for coping with these problems.

The model underlying the F test for a split-plot repeated measures
design does not include a term for seguence or carry-over effects. Thus
repeated measurements on the same subject should be avoided for treat-
ments in which the administration of one level affects performance on a
subseguent level. An exception to this, of course, is when carry-over effects
are the primary interest of the experimenter. Gaito (1961) has discussed
the problem of order efject when repeated measures are obtained on the
same subjects and has emphasized the importance of randomizing pre-
sentation of treatment levels.

LAYOUT AND COMPUTATIONAL
PROCEDURES FOR TYPE '
SPF-p.g DESIGN

The layout of a type SPF-2.4 design is illustrated in Table 8.2-1.
Let us assume that an experimenter is interested in vigilance performance.
He has designed an experiment to evaluate the relative effectiveness of two
modes of signal presentation during a four-hour monitoring period.
Treatment A, which is designated as mode of signal presentation, has two
levels, a, — auditory signal (tone) and a, < visual signal (light). Treatment
B has four levels corresponding to successive monitoring periods: b, < |
hour, b, — 2 hours, b, — 3 hours, and b, — 4 hours. The research hypoth-
eses leading to this experiment can be evaluated by means of statistical
tests of the following null hypotheses:

H,:x;,50 for all i

H,:x,£0 for some i

H,:B, GO for some j

Ho: 4aB,;, <0 for all ij
H,:aB, £0 for some ij.

The level of significance adopted for all tests is .05.

A total of eight subjects representing two random samples of
four subjects each has been obtained from a common population. The
two samples of subjects are randomly assigned to the p — 2 levels of A
and observed under all g — 4 levels of B. The dependent variable is desig-
nated as response latency to the auditory and visual signals. Response