Vous êtes sur la page 1sur 9

This article was downloaded by: [Northwestern University]

On: 12 January 2015, At: 10:12


Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41
Mortimer Street, London W1T 3JH, UK

Journal of the American Statistical Association


Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/uasa20

One-at-a-Time Plans
Cuthbert Daniel

R. D. 2, Rhinebeck , N.Y. , 12572 , USA


Published online: 05 Apr 2012.

To cite this article: Cuthbert Daniel (1973) One-at-a-Time Plans, Journal of the American Statistical Association, 68:342, 353-360
To link to this article: http://dx.doi.org/10.1080/01621459.1973.10482433

PLEASE SCROLL DOWN FOR ARTICLE


Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained in the
publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or
warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed
by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly
in connection with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions

One-at-a-Time Plans
CUTHBERT DANIEL *

5. Nested or curved o.a.t, plans produce a subset of results by


variation of one easy-to-vary factor, with all others held
constant.

Downloaded by [Northwestern University] at 10:12 12 January 2015

One-et-a-time experiments are always done when the experimental


system is set up to produce single results or pairs of results. When
random error is small compared to main effects expected, such experiments are economical, but may give biased estimates. These biases
can usually be described by two-factor interactions (2fi). Minimal augmentations of standard one-at-a-time sequences are given, first to
separate main effects from 2fi, then to estimate each 2fi separately.
Each new datum produces one or more new estimates.

It almost, but not quite, goes without saying that we


must give up most randomization if we are to do o.a. t.
plans. Randomization is a form of insurance against bias
from selection, planned or accidental, of experimental
material or of treatment. Some situations require more
1. INTRODUCTiON
insurance than others. The two commonest sources of
Some scientists do their experimental work in single bias are, I believe, model error and uncontrollable timesteps. They hope to learn something from each run, or trends. (Unconscious selection of favorable allocation of
trial. They see and react to their data more rapidly than experimental units to treatments seems to me to be rare
experimental agronomists or clinical investigators whose' in serious scientific experimentation.) The plans proposed
endpoints may take months or years to materialize. The here express and detect model error as two-factor
statistician who tells such an experimenter that he can interactions.
Uncontrollable time-trends are mainly dangerous to
secure much greater precision or validity by doing 16 or
more runs in a balanced set, may be listened to with o.a.t. plans when they are long compared to the time
required to do a single run (trial). The two natural routes
courtesy, but rarely with enthusiasm.
If a one-finding-at-a-time experimenter has in fact to take to minimize the effects of such trends are: (1)
found out a good deal by his methods, it must be true to use blocks of two trials, as close together as possible,
that the effects he has found are at least three or four and (2) to use a trend-free or trend-robust design. We will
times his average random error per trial. The plans take both routes, but not simultaneously.
The factorial representation and its nomenclature have
proposed in this article are directed to such experimenters. The plans are arranged to produce data of been used unchanged for nearly forty years. Both are
greater validity, i.e., of less bias, than the usual sequences well expounded in many texts, e.g., [7, 4, 2]. For our
of one-at-a-time trials. The improvement comes in re- present purposes, only the elementary logic and symmoving two-factor interaction (henceforth 2fi) biases bolism of the 2 n series will be required. One level of each
from main effect estimates. Minimal augmentations are factor is indicated by the presence of the corresponding
given, both for separating effects from 2fi, and for esti- lower-case letter in a symbol; the other level is indicated
by the absence of that letter. Thus in a three-factor exmation of single 2fi components.
periment,
ac means that a trial is to be made at "high"
The one-finding-at-a-time scientist usually varies one
levels
of
the
two factors A and C (the assignment of the
factor (independent variable) at a time. It is convenient
term
high
is
arbitrary) and at the "low" level of B. The
to divide one-at-a-time plans into five categories, since
symbol
(1)
means
a trial with all factors at their low
their informative augmentations follow different patterns.
levels.
The
average
effect, measured from the mean, of
I name these five types: strict, standard, paired, free, and
varying
factor
A
over
all combinations of levels of the
nested or curved.
other factors is indicated by A, its estimate by A. If a
1. Strict o.a.t. plans vary one factor from the condition of the
2 2 factorial experiment shows that factor A has different
last preceding trial.
effects
at the two levels of factor B, we call this form of
2. Standard o.a.t, plans vary one factor from some standard
non-additivity
"the two-factor interaction AB" and meacondition.
sure
it
by
the
obvious difference of the two "simple"
3. Paired o.a.t, plans produce two observations and hence one
A-effects:
.
simple comparison at a time.
4. Free o.a.t.vplans are for those who can make each new run
under any conditions that appear useful, after study of all
earlier runs.

[Cab - b) - (a - (1))J/4 = Cab - a - b + (1)J/4.

* Cuthbert Daniel is a private consultant, R. D. 2, Rhinebeck, N.Y. 12572.


The substance of this article was read as the Fisher Memorial Lecture at the
annual meeting of the American Statistical Association at Colorado State University, on AugllBt 25, 1971. The author is indebted to J.T. Daniel, O. Kempthome,
B.H. Margolin, H. Scheff~, J.W. Tukey, M. Zelen, and to a referee for helpful
comments.

Conforming to tradition, we use the same symbol to


denote the conditions for a trial and the result of that

353

Journal of the American Statistical Association


June 1973, Volume 68, Number 342
Theory and Methods Section

354

Journal of the American Statistical Association, June 1973

Al and A2. STRICT ONE-AT-A-TIME MINIMAL


23//4 .. 8; TIME-TRENDS NEGLIGIBLE
Run

Spec.

(1)

,,,

Estimable

A - AB - AC

ab

B+AB-BC

abc

C + AC + BC

(jJ'

be

a c

8 a

~J""\ A2

-.
A

~7/
I
I

A; AB + AC

...
\.1'7--

B; C; AB - Be

"
I

AB; AC; BC
ABC;

8
A~

V
2

J. W. Gorman has shown me, the easiest way to find this


expression-and most others in this article-is to enter
+ 1 or -1 in the corresponding places in Yates' standard
order [7,p.15J, or [2,p.264] for the 2", and then to carry
through the familar addition and subtraction algorithm.
The non-zero entries in the nth column (here n = 3)
identify the aliases. Thus:

(0) (1)

,
3

All other effects with double precision.

Downloaded by [Northwestern University] at 10:12 12 January 2015

Not strict o.a. t.

trial. If a full 2 3 is done, then the expression for AB just


given must be averaged with the corresponding quantity
for the four runs at high C. The difference between these
two measures of AB gives (twice) the three-factor interaction, ABC.
For a 2" factorial experiment, a treatment combination
may be represented by an n-tuple, x = (Xl, X2, .. x,,)
where each x is 1. Thus the factorial representation of
the result of a single run in a 2 3 may be written:
y(x)

+ AXI + BX2 + CX3 + (AB)xlx2


+ (AC)XIX3 + (BC)X2X3 + (ABC)XIXZX3 + e(x)

where A, B, C, (AB), etc. are parameter values,


M is the average of all eight error-free responses,
= -1 at the low level of A, B, or C for i = 1, 2,3
= +1 at the high level of the corresponding factor.

Xi

The e(x) represents the random independent error in


this particular trial, and is assumed to have expected
value zero and constant variance, o-Z. The symbols (A B) ,
etc. are inseparable, not products. Their parentheses will
usually be omitted. As an example, for run ac we write
E{ae} = M

+A

- B

+C -

AB

+ AC -

BC - ABC.

It is common experience that 2fi's are rarer than main


effects. Three-factor interactions (3fi) are generally rarer
than 2fi. We propose in this article to assume that all 3fi
are zero or negligible, but to retain symbols for all 2fi. We
may drop this assumption about 2fi if evidence turns up
that sums of two or more are small.
2. STRICT ONE-AT-A-TIME PLANS

-1

(3)

(1)
a

+1

0
0

0
0

0
2

b
ab

0
0

0
0

2
0

0
-2

B
AB

e
ac

0
0

2
0

0
0

0
-2

C
AC

be
abc

0
0

0
0

-2
0

0
2

BC
ABC.

It is worth emphasizing that this computation is not


done for the purpose of finding effects from data, but for
getting aliases from designs.
It is easy to see that the expected value of the
A-effect, measured at high Band C, i.e., by (abc - be), is
2(A + AB + AC + ABC), and hence that we can estimate the effect of A free of 2fi by adding to the o.a.t.
plan of Fig. A1, the single run be, since
E{a - (1)

+ abc -

beJ = 2(A

+ ABC).

Similarly, by going to e, again by strict o.a.t. variation,


we can estimate B free of 2fi bias from the two pairs of
runs in which B was the only factor varied. The effect of
C can also be estimated from (e - (1) + abc - ab)/4.
It should be clear from Figure A2 that the augmenting
runs must lower the levels of the factors in the same order
in which they were raised. The six-run plan permits the
estimation of the three main effects unaliased with any
2fi and so is, in familiar terminology, of Resolution IV.
This plan, with its obvious extensions to n > 3, requires
the minimum number of runs [6J for a Resolution IV
2", namely 2n. They all have the additional property of
A3. STRICT O.A.T. MINIMAL 25//10 .. 16;
TIME-TRENDS NEGLIGIBLE
Run

Spec.

Estimable

J!wL

(1)

2.1 The 2 3 in Six to Eight Runs, or 2 3//6 ... 8


We start with the simplest instructive case. The
experimenter has three factors, A, B, and C under study.
He has made four runs, each varying a factor from the
conditions of the preceding run, thus: (1), a, ab, abc, as
diagrammed in factor space in Figure AI. He now wants
safer, more generalizable, estimates of the average effects
of A, B, and C than are given by the obvious comparisons
between the results of consecutive runs.
The simple A-effect from the first two runs, (1) and
a, has expected value (A - AB - AC + ABC)/2. As

(2)

AE

de

AD

A -ints. with A

12

ab

B "

13

a cde

AC;AB

abc

14

a cd

BE

abed

D "

15

a c

BC;BD

abcde

16

ace

CD;CE;DE

cde

de

10

Estimable

bede

Spec.

11

II

II

A; AB- + AC

B; AB - BC -

C; AC

AD

BD -

AE
BE

BC - CD - CE

D; AD + BD + CD - DE
E; AE + BE + CE + DE

355

One-at-a-Time Plans

being connected so that runs can be made in strict o.a.t.


order.
Estimates of pairs of 2fi can be made from the same
data. Thus

E[(abc - be) - (a -

(lJ =

4(AB

B. STRICT-ONE-AT-A-TlME; 23 / / 4 .. 10; LINEAR (L)


AND QUADRA TIC (Q) TIME-TRENDS ASSUMED
Estimable
(1)

+ AC),

A-AD-AC+L-3Q

C+AC+BC+L-

abc

and

be

El (abc - ab) - (c - (I}

Downloaded by [Northwestern University] at 10:12 12 January 2015

B + AD - BC + L - 2Q

ab

El(be - c) - (ab - a)} = 4(BC - AB),

4(AC

+ BC).

If these estimates come out small, the experimenter


probably need not worry about 2fi. It is conceivable, but
hardly likely, that there is in fact a large positive AB
and an equally large negative AC. If the experimenter feels
this a contingency worth guarding against, he has only
to read further.
If any of the three contrasts measuring the sum or
difference of two 2fi turns out to be large, it will surely be
important to separate its components. The single run ac
will accomplish this. Inspection of the 23 sketched in
Figure A2 shows why this is so. The run ac completes
three faces of the cube, each one of which permits the
estimation of a 2fi. The statistician-reader who is used to
seeing data with large error may well need to be reminded
that we are dealing with systems whose effects are large
compared to their errors. The fact that the 2fi estimates
given here have twice the variance of those from the full
2 3 is then not seriously adverse and is more than balanced
by the advantage of optional stopping.
We now extend this simple example in three directions:
to get better elimination of time-trends, to study more
than three two-level factors, and to study factors at more
than two levels.
2.2 The 2 3//6 ... 8 with Elimination of Linear and Quadratic
Time-Trends

We start with the 2 3//6 of the preceding section. Since


two more requirements are to be met, namely the elimination of linear (L) and quadratic (Q) time trends, two
additional runs will be needed. The runs must repeat
earlier experimental conditions. We choose the conditions
(1) and a in that order, to maintain our strict o.a.t.
regime. This will give the A-effect with half the variance
of the Band C estimates. We could of course choose b
or e instead of a.
It is assumed that trials can be made in nearly equallyspaced periods of time. If they cannot, then the simple
integer coefficients given later will not be correct. Adding
linear and quadratic time terms to the usual factorial
representation and solving the resulting equations for
parameter estimates will then usually require some computer aid.
Figure B shows the two extra runs added as numbers
7 and 8 to the former 23//6 of Figure A2. Each estimate
before run number 8 is inevitably biased by L or Q or
by both. It is of some interest to see that early estimates
are not equally trend-biased, and that the L-bias is

(1)

AD + AC - L + 3Q/2

B - 3Q/2;

AD - BC - 2L + Q

C - 3Q/2;

AC + BC + L + Q/2

A;

a c
10

AB;
ABC;

A - 3Q/2;

B;

C;

AC;

AD + AC;

AD - BC;

L;

BC

All other effects with nearly doubled precision.

eliminated for the three main effects at runs 5, 6, and 7,


respectively.
The orderly solution of the eight factorial representations of the single responses, with Land Q added, bringing
in one new result at a time, gives all the expected values
and corresponding estimators shown in lines 2 to 8 of
Figure B. The estimators after run S are shown in Table
1 as regression coefficients, like all parameter-estimates in
this article. These are half the overall effects used by
Yates, Kempthorne and Davies for the two-level factorials. The Land Q coefficients are scaled by the familiar
Fisher-Chebyshev orthogonal polynomial integer coefficients [3].
The reader will notice that the Land Q regression
coefficients are estimated by the familiar contrasts for
four equally-spaced or -timed observations, even though
these are not so spaced. The contrast for sA is a weighted
average of the three intuitive estimators of A, needing no
further trend correction. The coefficient - 3 for a in 813
is "really" (-2-1). The -2 gives with the other three
2's an estimate of (8B - 12Q). The four l's appear
in the right places to estimate + 12Q.
I conjecture that a similar partition of any contrast
whose coefficients are not all 1 (but which are small
integers) is always feasible and will often be illuminating.
This plan gets all its information on quadratic trend

1. ESTIMA TlON MATRIX FOR STRICT o.a.t. 23 / / 8 WITH


LINEAR AND QUADRATIC CORRECTIONS
Run

Spec.
Effect

(1)

ab

-1

+1

II

+1

-3

+2

+1

-1

-2

12 (AC+BC)

+1

-2

12 (AC+BC)

+1

-3

24

-1

12

II

+1

abc

+2

Var(X)

Eff.

3/16

2/3

(1)

be
~2

-1

+1

-2

-1

+1

3/8

1/3

+2

-3

+1

3/8

1/3

+2

-1

7/36

9/14

+3

-1

19/72

9/19

-1

+1

+1

-1

-1

+l

+2
+2
+3

+3

-3
-3

1/144
1/36

6/7
3/14

Journal of the American Statistical Association, June 1973

356

ct.

STRICT ONE-AT-A-T1ME 33//20 .. 24.


TIME-TRENDS NEGLIGIBLE. MAIN
EFFECTS CLEARED FIRST
9

2.4 The 3 3/ / 12 .. 24

19

I
+-_
_

11 .....

I
I

1~
2

~~

.-

17

~6

1c

118
I

Downloaded by [Northwestern University] at 10:12 12 January 2015

~
20
2

The symbol in the caption means that the effects of


three factors (the exponent), each at three levels (the
base), are to be explored by o.a.t. rules in from 12 to
24 runs depending on the degree of aliasing-and hence
biasing-that is discovered as the trials proceed.
Assuming all 3fi and error to be negligible, the factorial
representation for the 3 3 may be written as a regression
equation:

Y = M

I
I

effect and 2fi estimates in 16 runs. This limitation can


perhaps be made useful as a reminder to the experimenter
that if he commits himself to strict o.a.t. plans, then only
a few paths of minimal length are possible.

A...
from pairs of runs at the ends of the sequence, and therefore involves a risky correction to the results of the inter. mediate runs. If the experimenter can break the strict
o.a.t. rule for one run, it is natural to ask for a "standard"
run, (1), in the middle of the sequence. This produces a
slightly more precise design. It is a safer plan since it
takes its measure of quadratic trend partly from the
middle of the sequence. As before, the runs ae and b will
give 2fi estimability and improved precision, respectively.
2.3 2 Plans for Four or More Factors

Strict o.a.t. plans for the 2 4//8 and the 25//10 are
exactly analogous to the 2 3//6 already discussed. If simple
trends are likely, then 2n + 2 runs will suffice. But
augmentation to estimate 2fi brings minor surprises.
The 2 4//8 can only be extended in one way, namely to
ad (permitting estimation of AD). Then there are two
alternatives for the next run, abd (to estimate BD and
CD), or aed (to estimate AB and AC). The final run in
each sequence is then forced. It is bd for the former and
ae for the latter (to estimate all remaining 2fi).
The 25//10 .. 16 is shown in Figure A3, for ease of
comparison with Figure A2. It has been assumed that the
experimenter wants the interactions with A to be estimable first. Starting with ae it is also possible to continue
with abe, abee, bee, ce.: and ace and so find first the interactions with E, then those with D.
The only other starting point (at run 11) is ee, which
permits the interactions with D to be estimated first.
There are a few minor permutations possible after ee, but
it is almost disappointing to find that 60 very few paths
(six in all) through the five dimensional factorial grid are
admissible in the sense that they provide at least one new
estimate after each run, and that they yield all main

+ ALXl + B Lx2 + CLX3

+ AQx~ + BQx~ + CQx~


+ (A LBL)XIX2 + (A LCL)XIX3 + (B LCL)X2X3
+ (ALBQ)XIX~ +... + (BQCL)X~X3 (six terms)
+ (AQBQ)x~x~ + (AQCQ)x~x~ + (BQCQ)x~x~,
In the interests of typographical economy, carets are
omitted over the regression coefficients on the right. It is
anticipated that a small minority of the terms in this
equation will dominate. All Xi are scaled to have levels
-1, 0, +1. This scaling has the familiar meaning for
factors with quantitative equally spaced levels. It can
also be used for "qualitative" factors when two new
varieties (or unordered treatments) are being compared
with a standard. In this case the "linear" term estimates
the difference between the two new varieties and the
"quadratic" term measures the difference between the
standard and the average response of the two just
contrasted.
The equation is printed on five lines to separate sets of
terms of decreasing expected importance. The plans
which follow produce parameter estimates in the same
order.
.
3//20.
We give two strict o.a.t. paths through the 3
The
first (Figure 01) gives early estimates of the A, B, and
C effects, biased as shown; the second (Figure 03) clears
the A-estimates in 11 runs, then those of Band C.
Figure 01 shows in its first 12 runs an edge-path
analogous to the 2 3//6 of Figure A2. We see in Figure 02
that the main effect estimates are cleared of linear-bylinear (L X L) 2fi by the opposite-edge augmentations,
but that the L X Q and Q X Q components are still
attached. To get full separation we must of course take
observations at the axial or internal points. The figure
starts by going to the A-axis points but any of the three
factors (B by strict o.a.t., C not) could be chosen. Presumably the experimenter would prefer to clear first the
factor showing the largest effects, or largest interactions.
After run 15 has been done, we have nine points in the
sloping plane from runs 1, 2, 3, through 13, 15, 14, to
9, 8, 7, and these suffice to estimate A L and A Q free of
all 2fi. Table 2 gives the estimation matrix for the A
effects and for the separated pairs of 2fi with A. Two

357

One-at-a-Time Plans

more runs (16 and 17) permit corresponding separation


for B and three more complete the design for C and its
interaction-pairs.
Just as for the 2 n series, we can derive the alias structure of each estimate by use of the extension of Yates'
algorithm to the 3n series which is shown in Davies
[2,p.366].
The aliasing patterns are complex at first sight, but the
computations of effects and 2fi are all simple and should
be done by hand. After run 3:

"A L "

"A Q"

(Ya - Yl)/2;

(Yl - 2Y2

Downloaded by [Northwestern University] at 10:12 12 January 2015

+ALCL =
AQBL +AQCL =

ALBL

+ Y7 - Y9)/4,
(Yl - 2Y2 + Ya + Y7 - 2ys + Y9)/12,
(-Yl + Ya + Y7 - Y9)/4,
(-Yl + 2Y2 - Ys + Y7 - 2ys + Y9)/12,

Estimable

(1)

1
a
a
a

' \ - A ed gh t Zfi with A


Q

'\ four Zfi ' \X;

Zb l

B - B eight Zf1 with B


L
Q

A
Q

four Zf1 AQX

B four 2f1 BQX


Q

aZb Z

BL four 2ft BLX;

C - C eight 2f1 with C


L
Q

ZbZc1

aZbZc Z

S.

lb Zc Z

'\ + AQ + four 2f1

~;

bZc Z

'\ + ,\B Q + '\CQ;

AQ

four 21;

'\BL
10

c/

6~

Spec.

a
a
13

4(-\~)C

-1

(1)

lSA
Q

12(A~~)

36(A~XQ)

-1

-Z

2
b c
1 1
c

15

14

10

11

aZbZc Z

-1

-1

-Z

lb l l

-2
-1

-Z

-2

c
Zb l l
bZc Z
lb Zc Z

-1

-1

-Z

-2

-Z

AX means "interactions with A."


b Run numbers for the two paths of Figure. Cl and C3.
e ALXL means "estimate of ALBL + ALCL."

(Ys - Yl

C2. ESTIMA TES AND ALIASES AFTER EACH RUN


OF THE 3s//20 24
...&lm.

c2

+ Ys)/6.

The inverted commas are merely intended to remind the


reader that these are early and hence maximally-biased
estimates.
After run 9 we have:

"A L " =
"A Q" =

2. ESTIMATION MATRIX FOR A AND AXe in the 33//20

+ ,\CL;

four 21 CQX

four Zfi

AQB L

~BQ

all as would be intuitively guessed. Each divisor is the


sum of the squares of the coefficients of the observations
in that estimator. The symbol
is used to abbreviate:
"The sum of the two parameters on the left is estimated
by the statistic on the right."
Just as for the 2 3//6, we must now add runs at the
vacant corner (ac in Figure AI, a2C2 in Figure 01) and
along one or more edges adjacent to that corner, to
separate one or more 2fi pairs. It is clear from the figure
that the addition of two runs will complete one face
(a 3 2) and so permit estimation of the separate elements
of one pair of 2fi. The four runs a2C2, alC2, a2Cl, and a2blc2,
will allow estimation of all 2fi.
Another path through the 3 3//20 responds to the experimenter who requires early information of the main effects
of A. It is shown in Figure 03. The full Resolution IV

C3. STRICT ONE-AT-A-TiME 33//20 .. 24. TIMETRENDS NEGLIGIBLE. A CLEARED FIRST, THEN B.

+ AQC Q;

~~

10

11

-.Cz

11

az

z
~

13

CQ

+ ~CL + BQCL;

~CQ

+ BQCQ;

'\B L - BL~;

' \CQ

+ BL CQ

'\~

+ BLCL

17

lcl

14

1S

16

17

rs

a l -. c l

19

ZO

(Zl)

(ZZ)

C2J)

(Z4)

Zb l l

,"'-

lb l l
lb Zc l

Zb l

16

15

CL; CQ;
C

,\B L;

,\C Q
'\~;

+ BLCQ;

AQC Q

+ BQC Q

z
z

,\B Q;

~BL;

AQB Q

,\CQ;

AQC L;

AQCQ

BLCQ;

BQC L;

BQCQ

I
I

~.\

BLCL

16

I
I

Zb1

14

ii- -

lb 1

;!t-I
I //

18

20
1

A.-

13

1c

358

Journal of the American Statistical Association, June 1973

01. STRICT ONE-AT-A-TiME. 72//21

02. ROW SWEEP. 72//21

--

1234567

13

14

,
I

5
6

15

If>

21

20

9-

17

11
I

I
I

18

7
3

- 1~
8- -

7
11
13

16

2.5 A 72//21

10

plan can now be done in 23 runs, since we are not compelled to repeat any point by this route.
Downloaded by [Northwestern University] at 10:12 12 January 2015

234

16

12

-- -

12

14

15
17 18

20 21

19

through all levels of one factor at fixed level of the other,


will be more interested in Figure D2. Here the same
BIB plan is arranged and numbered for this purpose.
Such a plan is obviously not a strict o.a.t. plan. It could
be rendered more robust to linear and quadratic timetrends by repeating its first row, or even one observation
from its first row, near the middle of the sequence and
again at the end.

We take the case of two seven-level factors and assume


that the experimenter prefers to sample the 49 cells,
getting early information on row and column effects,
together with moderate, patchy information on nonadditivity. If the 21 observations can be made changing
3. STANDARD ONE-AT-A-TIME PLANS
only one factor's level at a time (three in each row and
column) we will have eight degrees of freedom for samStarting simply, we contemplate the four trials specipling the 2fi.
fied by (1), a, b, c as diagrammed in Figure E. This plan
The standard balanced incomplete block (BIB) design has all the defects of 2fi bias of its strict o.a.t. counterpart,
for seven varieties in seven blocks of three [2, Appendix plus another which the diagram makes obvious. All the
6Aj 4,p.528J, can be written as a "partially replicated runs are done in one corner of the cube. It is then going
two-way layout," using varieties for rows and blocks for to be harder to augment this plan to remove 2fi biases.
As R. De Baun showed me long ago, it is possible to
columns. It can then be rearranged as shown in Figure
D1 so that the maximum number of rows and column get a rather blurred look at non-additivity by appending
can be swept through in sets of three runs. Such a plan, to the first four the single run abc, numbered 5 in Figure
followed in the sequence indicated, will give estimates E. If the response found there matches closely the value
of all row and column parameters by its 17th run, biased we would expect from the first four runs, we might dare to
of course by whatever 2fi exist in those cells. This plan guess that we have near-additivity. The expected value of
should be compared with the classical o.a.t. plan which the difference between the "abc effect," i.e., [abc - (1)J,
goes through one row and one column in 13 runs, but and the sum of the three simple effects of A, B, and C, is
gives no clue to non-additivity. If large residuals appear 4(AB + AC + BC) and so may give some idea of the
when the calculated row and column effects are sub- seriousness of the failure of the factors to operate
tracted from the observed values, the experimenter additively.
knows that he has found serious interaction.
E. STANDARD ONE-AT-A-TIME; 23//TWO
SETS OF FOUR

2.6 Other Partially Replicated Two-Way Layouts

The cells unoccupied in Figure D1 also form a BIB


plan. They are also re-orderable so that no diagonal steps
are taken, and hence are usable as a strict o.a. t. plan.
This can be used as a start if a more thorough look at 2fi
is desired (with 15 degrees of freedom instead of 8), or to
complete the square if that is judged advisable.
Many other BIB and even PBIB (partially balanced)
can be rearranged in this way, but some cannot. For
example, the Regular Group Divisible plan for six
varieties in six blocks of three [1, design R1J does not
appear to be orderable in this sense.
The o.a.t. experimenter who insists on sweeping

Run

Spec.

Estimable

(1)

A-AB-AC

B - AB -

C - AC - BC

abc

AB

bc

A'

AB

AC

a c

B;

AB

BC

ab

C;

Ac

BC

BC

..~'

t~
,~
-

I~

AB' AC; BC.


(Main effects with Eff.

~;

2fi with Eff. 1)

359

Downloaded by [Northwestern University] at 10:12 12 January 2015

One-at-a-Time Plans
Another run must be added to get a single interaction
isolated. If two can be added to the six already done,
then all three 2fi can be estimated. The completion of
the cube is then well-nigh irresistible, since it permits
double precision on all main effects and 2fi as well as a
first look at the 3fi, ABC.
Suppose now that a standard set of four trials has been
done but that some drift or blocking effect must be
allo~ed for if new data are needed. The full complementary set of four, a one-down-at-a-time set from abe
must now be carried out. Some information does appear
after each run. Thus if the upper horizontal pair (runs
5 and 6, or abe and be) are completed, we have .a new
estimate of the effect A, which may be averaged with the
earlier one to get a less-biased estimate of the effect of A
over the whole cube. The two A effects may also be
differenced to get a measure of the sum of the 2fi with A.
namely (AB + AC).
When the second set of four is complete, one surprising
and one disappointing result emerge. The three 2fi are
separately estimated with full efficiency; the blocking
bias has not interfered. But the three main effects can
still only be estimated within blocks and so only with
efficiency factor t.
So much for this fractured 2 3 All of this generalizes
easily to larger numbers of factors, provided that the
word easily is interpreted loosely. Margolin [5J has given
a useful table of weighing designs that can be used for
this purpose.

4. ONE-PAIR-AT-A-TIME PLANS
Much routine testing is controlled by repeated comparison with a standard. Sometimes the comparisons are
made in sequence in one test set-up; sometimes two
parallel set-ups measure standard and unknown simultaneously. Blocks of two will be more useful when parallel
trials are technically reasonable.
It may be a mistake to rush greedily forward and use
those blocks of two that give early main-effect estimates,
cleared of 2fi. Thus, if we consider the two blocks
abed- (1), and a- bed, we can indeed find the A effect
unaliased with any 2fi, but we have aliased all 2fi with
block means and so can only estimate them biased by the
block-to-block variation, which is expected to be large.
If instead we start as in Figure F, with the philistine
simple comparisons of Blocks I to IV, we are in the
framework of the earlier plans. We get each main effect
with all its 2fi after the first round, but we can now separate
main effects from 2fi strings by doing the complementary
sets V to VIII.
There is no logical necessity to carry out these eight
blocks in the Roman order. Thus if the very first block
showed a very large effect, the experimenter might want
to go after that immediately. The first move wo~ld
surely be V, to separate A from its 2fi. If the latter strmg
appeared large, the two blocks IX and X (n~w number~d
3 and 4) would suffice to give separate estimates of Its
three components.

F. ONE-PAIR-AT-A-TIME PLAN. 24 / / 4 .. 12
BLOCKS OF TWO

Computation

Estimable

Spec.

(1)

A-AB-AC-AD

II

b -

(1)

B - AB - BC - BD

III

c - (1)

C - AC - BC - CD

IV

d -

(1)

D-AD-BD- CD

v+

I;

V - I

abed - bed

A;

AB+AC+AD

VI

abed - acd

B;

AB

10

VII

abed - abd

C;

AC+BC+CD

VII - III;

VII - III

11

VIII

abed - abc

D;

AD+BD+CD

VIII

+ IV;

VIII - IV

12

BC + BD

VI

+ II;

VI - II

IX

ab - b

AB;

AC

IX - I;

V - IX

abc - be

AC;

AD

X - IX;

V - X

XI

be - c

BC;

BD

XI - II;

XII

cd - d

CD

AD

I+VI-IX-XI

XIII - III

Suggested order if all 2fi with A, then with B, etc. are desired.

The experimenter may now want to return to the


factor B. If II and VI (numbered 5 and 6) are done, he
can estimate B and the string of 2fi containing B. Since
he has already found AB, only BC and BD require
separation. One more block, XI (number 7) will give BC
with good precision, but BD with less good, since four
block-differences as indicated must be used in its
estimation.
We have thus far done seven blocks and we have seven
effect estimates clear, namely, A, B, and all their 2fi. We
can get the usual preliminary estimates of C and D-each
main effect with all its 2fi-from Blocks III and IV
(numbered 8 and 9). We then separate C and D from their
respective 2fi by Blocks VII and VIII (numbered 10 and
11). A final block, XII, is necessary to get an unencumbered good estimate of CD.
In summary, we have shown one way to get good
estimates of four main effects and six 2fi's in twelve blocks
of two, learning something from each block, and responding in some way to each outcome.
5. FREE ONE-AT-A-TIME PLANS

We have had a simple example of a free o.a.t. plan in


Figure E at run 5, since we there added the run abc to
the standard set (1), a, b, e.
A different sort of example is given here, partly because
its analysis produces some surprises. We imagine that
some enlightened experimenter has completed the halfreplicate of a 2 3 , a 2 3- 1 then, with the experimental
conditions (1), ab, ae, and be, and that only one contrast
appears large. Call it the one that estimates (A - BC),
the usual A-contrast. We "therefore" assume that all
other effects, B, C, AB, and AC are negligible.
Every run in the half-replicate not yet done has an
expected value that includes (A + BC). We choose one
run, a, and write down its conditional expected value,
dropping all the terms just assumed negligible:

E'la} = M+A+BC.

360

Journal of the American Statistical Association, June 1973

The prime (') is a reminder that this expectation is conditional on the assumption that four parameters are zero.
We can then estimate (A + BC) by simply correcting
the observed response to trial a by the estimated mean
for the whole 2 3 which we can find from the 2 3- 1 already
completed.

-+ BC = a - M
4(A -+ BC) = 4(a - M)

(5.1)

= 4a - (1) - ab - ac - be (5.2)
4(A .: BC)

(1)

+ ab + ac -

be.

(5.3)

Therefore,

4A = 2a -

be

(5.4)

4BC = 2a - ab - ac.

(5.5)

(1) -

Downloaded by [Northwestern University] at 10:12 12 January 2015

and
Equations (5.2) and (5.3) are solvable by inspection
for A and BC. It is rather counterintuitive to see that
even though four observations are needed to estimate
(A - BC), only three are required to estimate A. The
estimator of BC (5.5) is also unexpected in that some of
us have not seen an interaction estimated from three
observations. Even this contrast is redundant, being as
it is the sum of two simple comparisons, each of which
estimates BC without bias. This is only possible, of course,
because we have already made a judgment that Band C
are negligible.

6. ONE-CURVE-AT-A-TIME PLANS
The common practice of sweeping through the levels
of a single multi-level factor, holding all other factors
constant, can be justified (or discredited, as the case may
be) by augmentation in the manner that must by now be
familiar, i.e., by changing all the factors held constant
to the other ends of their ranges, and again sweeping
through the levels of the easy-to-vary factor. If the two
curves produced have nearly the same shape, most
experimenters will take it as proven that the "easy"
factor does not interact with the others. If the two
curves are considerably different, then intermediate
curves-at intermediate levels of the other factors-must
be taken.
When the experimental situation requires, say, a
5 X 2 2, the primitive o.a.t.-er would do ao, aI, a2, a3, a4.
I suggest augmentation by another set of five at high B
and C. (To keep in the strict o.a. t. regime we should
interpolate the run a 4 b between the two sweeps.) We can
now separate the A-effects from their 2fi with Band C.
We do this, of course, by comparing the two five-point
curves (if A is continuous) or the five pairs of observations
(if A is discrete). If the two sets of five do not show similar
spacing of their responses then another set of five should
be run, either at high B and low C, or at low B and high
C. We will then have 16 runs and can separate and estimate the four degree of freedom interaction AB, the

4 d.f. AC, and the single value for BC. We are still
assuming that all 3fi are negligible.
The reader will see that more runs, patience and
ingenuity will be required if there are more than two
levels of B and of C, or if more than two "hard" factors
must be studied. But the general ideas are clear and the
examples given should suffice to guide the reader through
these more complex situations.

7. CONCLUSIONS
The laboratory researcher's practice of doing one trial
at a time, and of making some judgment after each trial,
is justifiedwhen the effects are expected to be of magnitude 40- or more. Methods have been given for refinement
of the first estimates produced by this practice, isolating
two-factor interaction biases, first in strings, then
individually.
These augmentations appear to work best for the
experimenter who can do "strict" one-at-a-time sets,
i.e., each trial varying only one factor from the previous
trial. They work well too for the researcher who is set up
to carry through simultaneous or closely related pairs of
runs. The "standard" one-at-a-time plans-each trial
once-removed from some standard condition-have the
limitations of their conservatism. They are harder to
augment, but the best method of augmentation is clear.
When long trends are likely to be present, and are
approximable by linear plus quadratic terms in time, the
addition of two runs may give sufficient information to
correct adequately for such drifts.
The two basic requirements (small error, quick results)
are stringent and, of course, cannot always be met. The
plans proposed here cannot conceivably be used in
agricultural field trials, in long-term clinical trials, in
full-scale plant experiments, or in studies of consumerproduct shelf-life. For these and for many other systems,
the classical plans of Fisher, Yates, Box, and their
associates seem irreplaceable and will probably continue
to dominate the field of multifactor experimental design.
[Received March 1972. Revised November 1972.J

REFERENCES
[lJ Bose, R.C. et al., Tables of Partially Balanced Designs, Technical
Bulletin 107, North Carolina Agricultural Station, Raleigh,
N.C., 1954.
[2J Davies, O.L., ed., Design and Analysis of Industrial Experiments, New York: Hafner Publishing Co., 1956.
[3J Fisher, R.A. and Yates, F., Statistical Tables, New York:
Hafner Publishing Co., 1949, Table XXIII.
[4J Kempthorne, 0., Design and Analysis of Experiments, New
York: John Wiley and Sons, Inc., 1952.
[5J Margolin, B.H., "Results on Factorial Designs of Resolution IV
for 2" and 2"3'" Series," Technometrics, 11 (August 1969),
431-44.
[6J - - , "Resolution IV Fractional Factorial Designs," Journal
of the Royal Statistical Society, Ser. B, 31, No.3 (1969), 514-23.
[7J Yates, F., Design and Analysis of Factorial Experiments, Harpenden, England: Imperial Bureau of Soil Science, 1937.

Vous aimerez peut-être aussi