Professional Documents
Culture Documents
3, 195-202
A methodology for evaluating a single chair, rather than making a comparison among
chairs, was developed from previous chair studies. The methodology was found to be
rapid and effective when applied to a prototype chair, giving information to the
manufacturer on overall comfort and good and bad points in the design. Testing took
place on three tasks and showed that chair comfort is influenced by the task as well
as the chair.
Introduction subjects to fit the chair to their body and workplace; and
finally, (3) evaluate comfort directly at the workplace using
Is it possible to do a rapid but comprehensive ergonomic
a known methodology so that comparison with previously
evaluation of a single chair? This question arose when the published data was possible. Evaluation at the workplace
authors were asked to evaluate a prototype chair for the can be accomplished by measuring postures adopted (eg,
manufacturers. The aim was not to compare it directly with Oxford, 1969, Grandjean, 1980), by comfort ratings OVotzka,
other chairs but to judge its adequacy as a sitting machine et al, 1969, Shackel, et al, 1969) or by measuring performance
and to discover its strengths and weaknesses. (McLeod, et al, 1980).
Most chair evaluations compare different chairs directly,
Branton (1969) makes the point that seating is only a
but in the current study an absolute evaluation rather than
means to an end rather than an end in itself. The motivation
a relative evaluation was required. This meant that published
for sitting is the task performed in the seat rather than the
standards had to be available for the evaluation methodology
chosen. seat itself so the best that can be hoped for in seating is to
achieve a 'state of non-awareness' of the seat. Despite this,
Any evaluation procedure must be both valid and reliable. performance changes in the task as a function of seating are
For chair evaluation, the use to which the chair is put is the notoriously difficult to detect. McLeod, et al, only found
overriding criterion in choosing a method for chair evaluation. a performance decrement when the seat was both maladjusted
Chair evaluations reported in the literature have assessed chairs and perceived as uncomfortable. Similarly, recording of
used for auditoria 0Votzka, et al, 1969), reading typing and postures or movements on the seat (Rieck, 1969) have shown
eating (Shackel, et al, 1969), school work (Oxford, 1969), no correlation with direct comfort measures. Hence the
easy chairs (LeCarpentier, 1969) and 'just sitting' (Grandjean, practical alternative for workplace evaluation is a comfort
et al, 1973). measure or measures.
There are three basic methods of evaluating chairs, apart
from 'expert judgement' which Shackel et al found to be
unreliable. These methods use measures which compare the Choice of evaluation measures
chair against anthropometric data and chair design principles Comparison with anthropometric data and principles
(eg, Oxford, 1969; British Standards 3044, 3079, 3893); Ever since Akerblom's 'Standing and Sitting Posture'
use fitting trials to adjust the chair to the operator (eg, monograph in 1948, principles from anthropometry,
Jones, 1969; LeCarpentier, 1969) and f'mally to have users biomechanics and orthopaedic surgery have been applied
evaluate the chairs experimentally, either in a laboratory to seating design. The basic principles are that the seat should
setting (Shvartz, et al, 1980, Grandjean, et al, 1973, Shackel, allow a variety of natural postures to be adopted, with good
et al, 1969) or at the real workplace (eg, McLeod, et al, 1980; support for the spine, especially the lumbar spine. Branton
Wotzka, et al, 1969; Jones, 1969; Shackel, et al, 1969). (1969) notes that up to 75% of the body weight is supported
Many evaluations combine two or more of these techniques on the area of the ischial tuberosities (the bones on the
(eg, Wotzka, et al, 1969; Oxford, 1969) and most bottom of the pelvic girdle), and on reasonably hard seats
recommendations on seating design (such as Diffrient, et al, this can mean support over only a few square inchesof area
1974; Panero and Zelnik, 1979) combine recommendations (approximately 1130sq cm). He shows that biomechanically
derived from all three methods. the seated person is unstable, even with a backrest, thus
requiring some static muscular effort to achieve stability.
For this study, a three-stage evaluation procedure was This static effort is fatiguing and must be relieved from time
employed: (1) evaluate the prototype chair against published to time by adopting what Dempster (1955) called 'temporary
dimensional recommendations; (2) use fitting trims to allow closed chains' of body segments. These chains, such as
0003-6870/82/03 0195-08 $03.00 © 1982 Butterworth & Co (Publishers) Ltd Applied Ergonomics September 1982 195
leaning on a desk, create temporarily stable structures to They quote Akerblom (1948) and Darcus and Weddell (1947)
relieve the static tension. that upper back support is unnecessary if a correct lumbar
These principles have been turned into design criteria by support is provided.
many authorities. Grandjean (1980) gave a succinct but All authors are in agreement that padding should be thin,
comprehensive set of criteria: with Diffrient, et al, recommending a maximum compression
1. every work seat should be adjustable for height with a 78 kg occupant of 3-8 cm and Grandjean recommending
(between 38 and 53 cm if it is for office workers); 2"0 cm. The seat, backrest and armrest material should be
slip resistant and able to 'breathe' to remove body-generated
2. the seat should be safe against tipping or slipping. It
moisture. There is also general agreement that the front edge
should have five feet, set in a circle at least as big as
of the seat should curve downwards to prevent pinching Of
the seat itself (40-45 cm diameter);
the underside of the thighs between the seat edge and the
3. the seat should allow the occupant sufficient freedom femur.
of movement. If the occupant needs to get up often,
or to move the seat sideways, then the five feet should For this evaluation, detailed dimensional recommendations
have casters; were gathered from a variety of sources and are compared
with the prototype chair in Table 1. It should be noted that
4. hollow spaces under the desk top help to reduce aches most of these are based on the same studies and thus are
in the legs, because these can be moved about; closely comparable. An example is Panero and Zelnik's list
which is derived from a comparison of five earlier sets of
5. the seat surface should be 4 0 - 4 5 cm across and
recommendations (page 127).
3 8 - 4 2 cm from back to front. A slight hollow in the
seat, with the front edge turned upwards about 4 ° - 6 ° It is important to note, with Shackel, et al, that there is
will prevent the buttocks from sliding forwards. Of no guarantee that a dimensionally correct chair will be
course the front edge should be rounded off; perceived as comfortable.
6. a light padding with 2 cm of latex, covered with non-
slip permeable material (wool or Dralon) is a great Fitting trials
aid to comfort; The technique of fitting trials has been found to produce
7. a work seat that is either fixed or tilting, but with a results comparable to anthropometric recommendations
high, non-adjustable backrest, gives good opportunities for automobile seats (Jones, 1969; Drury and Searle, 1965).
to relax the back muscles occasionally. Such a backrest The method allows the subject to adjust the dimensions of
should extend 4 8 - 5 0 cm vertically above the seat the work place until subjective comfort is maximised. These
surface and have a breadth of 3 2 - 3 6 cm. The backrest 'comfort settings' are either taken as the design data (eg,
should have a lumbar pad, which should give good LeCarpentier, 1969) or used as reference points about
support to the spine and the sacrum at a point which to vary each dimension so that a tolerable range is
10-20 cm above the lowest point of the seat surface. found for each dimension (Drury and Searle, 1965). In a
study of easy chair seating, LeCarpentier found that the
The upper part of the backrest should be concave
mean differences between comfort positions chosen by the
forwards. The backrest may, with advantage, be
same subject on different days were:
concave in all horizontal planes, with a radius of
4 0 - 5 0 cm; Seat height 2-8 cm
Seat length 2.0 cm
8. a workseat with adjustable backrest should give as Seat angle 1.7 °
much support as possible to the lumbar vertebrae. Backrest angle 2.1 o
The backrest should be 30 cm high and 38 cm broad, Seat-to-backrest angle 2.5 °
and be adjustable in both horizontal and vertical These values show a lower bound on the accuracy with which
planes; dimensions and angles should be specified: any tighter
9. foot rests are important so that small people can specifications will be largely imperceptible to users of chairs.
avoid sitting badly, or to meet similar problems in Fitting trials were used in this study only as a means of
people of other sizes. They should be an integral part structuring the user's first interaction with the prototype
of each work table; chair. Subjects were allowed 5 min to experiment with the
10. a work seat should be designed in conjunction with three adjustments available (seat height, seat length and
the work place at which it is to be used. The most backrest height). Chosen values were measured at the end of
important consideration is the distance from seat this period. Five minutes is considered long enough for a
height to work height. Assuming that the elbows are user to become familiar with a chair (Grandjean, et al, 1973;
held downwards and the arms bent at 90 ° , this distance Shackel, et al, 1969).
should be between 27 and 30 cm.
Perhaps the only arguable recommendation in this list is User comfort evaluation
number 7, the high backrest, which comes from Grandjean's Oxford's comment that the criterion for a good chair is
own work and is not seen in other recommendations. that the occupant can 'sit for long periods without becoming
There is a feeling by other authors that a high backrest aware of the chair' agrees closely with Branton's comment
is undesirable. For example, Floyd and Roberts, writing quoted earlier. Branton (1969) reviews the literature on
in British Standard 3044, state that seat comfort assessment and concludes that
'The backrest must not restrict the movement of the 1. Avoidance of discomfort should be the criterion
spinal column or of the arms; thus complete support 2. Simple relatively unstructured scales should be used
for the whole of the back is excluded'. 3. Subjects should indicate discomfort by body part
Panero
BS 3 0 7 9 Diffrient and Z e l n i k Grandjean Prototype
Feature a n d BS 3 8 9 3 e t al 1 9 7 4 1979 1980 chair
(if fixed) 43 cm 43 - 45 -
. Check lists should be used to have subjects evaluate Body part discomfort scales used in the past have included
specific features of the chair. a ranking of '3 most comfortable' and '3 least comfortable'
areas (Shackel, etal, 1969), through three point scales used by
He notes that it is seat comfort we want to assess, not user
comfort, and the user is only a channel of information from
the seat to the evaluator. Scales are thus needed of overall General c o m f o r t rating
comfort, body part comfort and specific chair features. Please rate the chair on your feelings now
General comfort scales have ranged from two point
using adjective pairs (eg, comfortable/uncomfortable) by feel completely relaxed
McLeod, et al, 1980, through three point scales
(uncomfortable, medium, comfortable) used by Grandjean's feel perfectly comfortable
team (Grandjean, et al, 1973; Wotzka, et al, 1969), and the
seven point scale used by Shvartz, et al, 1980 (1 = extreme feel quite comfortable
discomfort to 7 = extreme comfort) to the eleven point
scale used by Shackel, et al, (1969). feel barely comfortable
This latter scale, reproduced in Fig. 1, was constructed
feel uncomfortable
by having chair users rank order 20 statements about comfort
and choose 11 responses which gave the most consistent
equal interval scale. The authors used this to compare feel restless and fidgety
ten general purpose chairs, either for reading in a laboratory,
for typing at the work place or in a cafeteria for eating. This feel cramped
scale was chosen for the current study as it is fine enough
to evaluate even very comfortable seating. It is approximately feel s t i f f
an equal interval scale and the results of the current study
can be compared directly with the data for ten chairs feel numb ( o r pins and needles)
published by Shackel, et al.
feel sore and tender
In this study, the scale was administered every half hour
after the initial 5 rain adjustment period, for a total of 2~ h. feel unbearable pain
Subjects marked a blank scale each time, with measurements
later being made in half-scale-step intervals. Fig. 1 General comfort scale
Too Too 1 h * * * NS NS NS
narrow Correct wide
Seat width q> i 90 m i n NS * * NS NS *
Slopes too far Slopes too for 2 h NS * * NS NS *
towards hack Correct towards front
Slope of seat q)
150 min * * * NS * *
The armrests were appreciated but were considered too Rolls very quickly 3 2
hard and likely to interfere with drawer-opening in typing. Feet interfere with body 1 -
Padded armrests are to be introduced on production models,
I
introduction to the science of ergonomics.
The APPLIED ERGONOMICS HANDBOOK
has been adopted worldwide by universities
and other educational establishments as a
course text and has been reprinted five
times since its initial publication.