298

SPLIT-PLOT DESIGN

V a III — ——€—<—

8.14

TABLE 8.13-2 Analysis of Variance Table for
Type SPF-pru.g Design
a vv z v zi nmv zv — o —J—
F(A,B,C,
and D
" Fixed
Eflects,
Unweighted- Subjects
Source Regular df means df Random)
vv idi anni z ———

1 Between subjects npru — 1 N-1

2 A p—-l p-l [5]
3 C r-il r-l [3]
4 OD u-l u—1 [š]
5 AC (p- 1Xr— b) (pP- kr — ib) [5]
6 AD (p- Mu —- 1) (p — Xu — 1) š]
7 CD (r—- iXu—- l) (r — lXu — 1) [5]
8 ACD (p — 1Xr — lXu — 1) (p — 1Xr — lXu- b) [š]
9 Subj w.groups pru(n — 1) N — pru

10 Within subjects npru(g — 1) Ni(g —- lb)

li OB a-l g-l [iš
1) AB (p - lXa- D (p-Wa-b O (B]
13 — BC (a - Wr- 1) (a- IXYr-D O [B]
14 — BD (ga - lMu- b) (a — lXu- b) [is]
15. ABC (p — lXa —- 1X — 1) (p— lMa- Wr-1 O [i8]
16  ABD (p — lXg — lXu — 1) (Ip - Ma —- lMu-1). [18]
I7.'— BCD (a — lXr — lXu— l) (g — IXr — IMu- 1). [5]
18. -ABCD (p — iXg - Mr - Mu- 1) (p- iXa- Wr- Mu—-1). [iš]

19. — Bx subj w.groups prujn — lXa — lb) (N — prua — |)
a a z vv ze oo

20 Total . npyru — | Ng—-1
La z |] mnv —O—O—€—€6€6€——— — — —

COMPUTATIONAL PROCEDURES
FOR TYPE SPF-p.gr DESIGN

A split-plot design can be used in research situations reguiring
repeated measures on two or more treatments. One variation of this design,
called a type SPF-p.gr design, is described here. This design has one
between-block treatment (A) and two within-block treatments (B, C).
The design reguires np samples of subjects who are randomly assigned to
treatment A, with n subjects (blocks) in each level. The seguence of adminis-
tration of the BC treatment combinations within an np block is randomized
independently for each subject. An alternative to using repeated measures

ai o on

dna EM me o ji av de a o oi o an bla bb mate A sobane katka dka nie
oe, et ob