This article was downloaded by: [TEI Technology Educational Institute]

On: 01 June 2014, At: 09:41

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
Ergonomics
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/terg20
An analysis of domestic cooker control design
R.D. RAY
a
& W.D. RAY
b
a
Robert Gordon's Institute of Technology , Aberdeen
b
Birkbeck College, University of London
Published online: 24 Oct 2007.
To cite this article: R.D. RAY & W.D. RAY (1979) An analysis of domestic cooker control design, Ergonomics, 22:11, 1243-1248
To link to this article: http://dx.doi.org/10.1080/00140137908924698
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/terms-and-conditions
ERGONOMICS, 1979, VOL. 22, NO II, 1243-1248
An analysis of domestic cooker control design
By R. D. RAY
Robert Gordon's institute of Technology, Aberdeen
and W. D. RAY
Birkbeck College, University of London
This study is an attempt to identify, the particular arrangement of the controls of a domestic
cooker which would be the most natural and efficient to usc. A cooker hob simulator
capable of providing both' in line' and' in quadrature' ring layouts, and different ring-
control arrangements was used. Twenty-eight students carried out trials on the simulator
and their error rates were measured. A new statistical test was used to show that one of the
ring-control arrangements for the' in quadrature' rings, gave significantly less errors than the
other three and hence could be regarded as the easiest to use.
I. Introduction
The efficiency and ease of use of most machines and equipment depends on a
variety of factors ; one of these is fairly clearly the layout and operation of the controls.
In the home, one of the most complicated array ofcontrols to be seen is on the modern
cooker whether gas or electric, where often six or more control knobs on taps are
arranged on the front or rear panels. As a general rule these controls are placed' in
line', one for the oven, one for the grill, and four for the hob rings or burners. The
oven and grill controls which are probably used less frequently are usually placed at
either end of the array or together in the centre and do not normally present a real
problem of identification or operation. The arrangement of the four hob ring controls
however may vary appreciably, and being more frequently used often cause problems
of identification and error, and are the subject of this article.
2. Background
The most common form of design in manufacture seems to be to have the controls
. in line', operating the hob rings arranged in quadrature. With this spatial arrange-
ment, it is clear that four different designs of hob rings and their controls are prob-
able as shown in Figure 1. In fact all four of these arrangements have been manu-
faciured. The fact that the controls and the rings have different configurations often
presents the difficulty of relating a given control with its hob ring. If the control
array is symmetrical with the ring configuration, then it is fairly obvious that the two
right hand knobs are associated with the two right hand rings, and similarly for the
left hand side. It is less clear however which ofany two controls is related to a particular
ring. It could be argued of course that the controls would be identified with words or
symbols, and would therefore present no difficulty in usage. Even with some form of
identification, it is probable that the user will operate the controls automatically
according to some plan established on another or earlier model.
The variation in performance of people operating controls lead Fitts and Seeger
(1953) to suggest the idea of' compatability' of displays (i.e. rings) with their con-
trols. This inferred that there may be certain natural linkages within a persons
thought processes which connected some display with its control. Welford (1968) also
mentioned the possibility that people may form mental models by which they are
1243 $02.00 (d1979 Taylor & Francis Ltd
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4

able to capture the essentials of a system and its controls. Basically, it seems that the
closer the display and control, the more direct and unobstructed will be the mental
linkage and the more easily will be the operation, i.e. more compatible. Although the
rings and controls of a cooker are usually in different planes, direct linkages appear to
exist with the hob rings' in line' also, being straight parallel lines as Figure 2(a).
Even in design where the controls are closed in as in Figure 2(b) the linkages still
appear clear though not' parallel and there may be a limit to the degree of closeness
without loss of compatibility.
0
t! tJ
[o}
@@)@@)
Lev @':J
[01
Figure 2. Cooker control arrangements in line
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4

An analvsis of domestic cooker control design 1245
The question is then, is there a single arrangement which could be regarded as the
most natural design which would lead to easier operation and fewer errors, i.e. is
there a population stereotype? If such a stereotype existed, it should be possible to
recommend that arrangement as a standard. In fact in an edition of the Consumers'
Association magazine 'Which' (1977) following several tests both here and in
America, arrangement A4 was recommended as the best.
3. Method
In order to gain an impression of the frequency of the four designs in existing
cookers, a brief survey of those in the kitchens of this Institute showed that all four
arrangements were evident. These were naturally not new designs but varied in age.
The 35 cookers examined are classified in Table I.
Table I.
AI
9
A2
5
A3
5
A4
16
A review of new gas cookers at present on the market, showed that of the 14
different models examined, 6 were of arrangement AI, and 8 were of arrangement A4,
but none of arrangements A2 and A3. In the case of electric models, all the cookers
observed were from four main manufacturers, and these appeared to divide equally,
two employing arrangement A I and two arrangement A4.
In the experiment, following Chapanis and Lindenbaum (1953), a cooker hob
was simulated using coloured perspex discs to represent the rings, each of which
could be illuminated by bulbs operated from a switching panel. The circuits were
arranged so that anyone of the rings could be switched on at the panel and off by the
subject using the correct control knob. The control knobs were fitted on the front
panel of the hob, with an 'in line' configuration, but without any identification.
Also, by adjusting the wiring, all the hob ring and control arrangements could be
constructed without physically altering the knobs.
Twenty-eight female subjects, all students, took part, and in the first instance
were asked to record the arrangement which they preferred best. Each subject was
presented with a series of ring signals in a random manner, first with the rings in line
as Figure 2 and then with the quadrature arrangements; the presentation of the latter
was also randomized. The signals were given in fairly quick succession but with
slight variations in time between each signal to avoid anticipation. The subject was
instructed to switch off the ring light as quickly as possible, with what they thought
was the correct control. Each subject carried out 20 trials on each arrangement and
their errors, i.e. operating the wrong control, were noted for each test.
4. Analysis of Results
4.1. Preferred Arranqements
As the experiment was concerned with quadrature designs only, from the point of
view of preference, the rings' in line' arrangement was not considered here. The
subjects recorded the preferences, shown in Table 2.
Table 2.
AI
6
A2
7
A3
5
A4
10
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4

1246 R. D. Rayand W. D. Ray
4.2. Hob Rinqs ' /11 Line'
It was found that with the rings' in line', no subject incurred any errors both with
the control knobs extended and in the closed in position.
4.3. Hob Rinqs '/11 Quadrature'
The results of the experiments on the four' in quadrature' arrangements expressed
in terms of subject errors are given in Table 3. It is presumed that the most efficient
or easily operated design would be that involving the least number of errors: the
analysis thus involves the statistical question of selection.
Table 3. Experimental Results
AI A2 A3 A4
Number of Trials 560 560 560 560
Number of Errors 6 ~ 91 107 48
Percentage of Errors 12'2 163 19'2 86
Number or Successes 492 469 453 512
Proportion of Successes ~ 7 9 0838 0809 0-914
None of the classical statistical tests answer the question of which of several
populations is the 'best' in some characteristic. It is only fairly recently that appropriate
statistical tests have been devised and become available. It is therefore proposed to
follow the recent work of Gibbons, Olkin, and Sobel (1977) who do in fact lament
that such existing theory has not been used more extensively on applied problems
of this kind.
In order to be able to make some valid probability statements from the sample
information. certain terms need to be defined. Let II be the number of trials carried
out on each control arrangement. It is then necessary to define a distance measure 0,
which is the difference beteen the largest and second largest probability of success
among the k binomial populations of the experiment. Thus we may write,
0= Plk] - Plk-II
The number of populations is four in this experiment.
We must then specify a value of 0 = 0* E (0, I) which gives an indifference zone
within which we would not wish to discriminate the largest p, and a preference zone
where we would. We finally define P, to be the probability of making a correct
selection of the largest p value.
If P = (PI' P2 ... Pk)' then P = p(CSlp) is the probability of correct selection
(CS), whatever the values of P in the preference zone. This is either very difficult or
impossible to determine. However there may be a certain point or region PLF in the
preference zone whose P = P(CS\Pu-) = P*, is the minimum of P over all p. This
configuration PLF is called the least favourable configuration and we may write
P(CSlp) ;;, P(CSIPLF)
for all P in the preference zone. It is important to note that solving P" in this way,
given II and 0*_gives a conservative answer. Thus we can be certain that the probability
of correct selection is at least P*, and may be a lot greater.
It is apparent that control arrangement A4 has the best performance based on the
experimental information obtained, but with what confidence can this be claimed?
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4

An analysis of domestic cooker control design 1247
The procedure is first to specify a value of <5*, and here 005 has been chosen.
Tables of p* are provided in Gibbons (1977), for given nand <5* which on entry with
n = 560, and <5* = 0,05, gives a value of p* = 088. We can now say that we are
88 %sure that control arrangement A4 is the best, based on the sample size. Further,
we can be 88 %confident that the observed probability of success for A4, differs from
its true value by less than 0,05, and this would seem acceptable from a practical point
of view, considering the conservative nature of the test.
It might be of interest to note that if we wish to be more rigorous and take a value
of <5* = 0,025, then from the tables we find that P" = 0,60, which is a much lower
level of confidence. This is of course to be expected, since if we wish to be more
precise about choosing the arrangement with the largest probability at a smaller
distance from its neighbour, then we can not be so confident with the same sample
information available.
5. Discussion
The experiments showed a variation in performance of the subjects in terms of
errors, depending on the different configurations of the rings and controls. In both
'in line' arrangements, the results showed that no errors were incurred, which was
similar to the findings of Chapanis and Lindenbaum for an American population
and probably confirmed a greater degree of compatability.
In the case of the quadrature designs, it was seen that errors were observed in all
arrangements, but in varying proportions. Arrangement A4 gave the least number of
errors over all trials with AI next in order, which prima facie may suggest the most
efficient design, though an appropriate statistical analysis is still highly desirable,
before a clear conclusion can be drawn. Such an analysis did in fact show, with the
acceptable degree of confidence, that A4 was the best arrangement. The particular
statistical test that we used in this investigation is we maintain the most appropriate
for the experiment considered and is one that deserves wider dissemination among
applied scientists.
Considering subject preferences, it was interesting to see a spread of preferences
across all four designs, with 6 subjects for AI, 7 subjects for A2, 5 subjects for A3,
and 10 subjects for A4. There appears to be no great weighting for any single design,
though A4 had a slight majority of preference votes. This may well of course reflect
any subject's experience with their own home design, though the subject were students
in home economics and thus probably had extensive practice with all arrangements.
It is also interesting to notice that the manufacturers have now mainly reduced the
number of control layouts from four to two, one being A4 and the other AI.
Dans cette etude, on se propose de rechercher le mcilleur arrangement des boutons sur une cuisiniere. A
I'aide d'un dispositif de simulation, on a realise diverses combinaisons de corrcspondance boutons-fcux.
scion une disposition des feux cn ligne' au en carte '. Vingt-huit etudiants ant effectue des manipulations
sur cc dispositif. Leur taux d'erreurs a eli: enregistre et traite selon une procedure statistique recente qui
mantee qu'une des quatre combinaisons en carre ' est netternent superieure aux trois autres.
In dieser Stu die werden Versuche zur gtinstigen Anordnung von Schaltern fUr Herdplatten von Haus-
haltsherden bcschrieben. Die Untersuchungen wurden an einem Simulator mit unterschiedlichen An-
ordnungsrnogfichkeiten flir Schalter und Herdplatten durchgeftihrt. An den Versuchen nahmen 28
Denten tcil. Die registrierten Bedienfehler wurden mit eincrn neuen statistischen Test geprtift. Ais
Ergcbnis konnte die signifikant gcringste Fehleranzahl filr cine bestimmte Schalteranordnung bei quad-
ratischer Herdplattenanordnung festgestellt werden.
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4

1248 R. D. Ray and W. D. Ray
References
CHAPANIS, A., and LINDENBAUM, L. E., 1953, A reaction time study of four control-display linkages.
Human Factors, 1, 1-7.
GIBBONS. J. D., OLKIN, I., and SOBEL, M.. 1977, Selection and Ordering 0/ Populations. (A New Statistical
Methodology). (New York: J. WILEY & SONS).
FITTS, P. M., and SEEGER, C. M., 1953, S-R compatibility: spatial characteristics of stimulus and response
codes. Journal of Experimental Psychology, 46, 199-210.
WELFORD, A. T., 1968, Fundamentals of Skill. (London: METHUEN).
Which? Sept. 1977, Consumers' Association.
Manuscript received 20 February 1979
Revised manuscript received 30 April 1979.
D
o
w
n
l
o
a
d
e
d

b
y

[
T
E
I

T
e
c
h
n
o
l
o
g
y

E
d
u
c
a
t
i
o
n
a
l

I
n
s
t
i
t
u
t
e
]

a
t

0
9
:
4
1

0
1

J
u
n
e

2
0
1
4