Pain, 41 (1990) 139-150 139
The Faces Pain Scale for the self-assessment of the severity of pain
experienced by children: development, initial Validation,
and preliminary investigation for ratio scale properties
Daiva Bieri ~,!, Robert A. Reeve b, G. David Champion a, Louise Addicoat a
and John B. Ziegler a
a Division of Paediatrics, The Prince of Wales Children's Hospital, Sydney (Australia), and b School of Behavioural Sciences,
Macquarie University, Sydney, N.S.W. (Australia)
(Received 26 June 1989, revision received and accepted 15 December 1989)
Sununary Altogether 553 children (195 first graders, mean age 6.8 years, and 358 third graders, mean age 8.7 years)
participated in the development of a self-report measure to assess the intensity of children's pain. The first step was the derivation,
from children's drawings of facial expressions of pain, of 5 sets of 7 schematic faces depicting changes in severity of expressed pain
from no pain to the most pain possible. With the set of faces that achieved the highest agreement in pain ordering, additional studies
were conducted to determine whether the set had the properties of a scale. In one study, children rank-ordered the faces on 2
occasions, separated by 1 week. All 7 faces were correctly ranked by 64% (retest 1 week later, 61%) of grade 1 children and by 86%
(reiest 89%) of grade 3 children. In a second study, the faces were presented in all possible paired combinations. All 7 faces were
correctly placed by 62% (retest 86%) of the younger arid by 75% (retest 71%) of the older subjects. A third study asked children to
place faces along a scale: a procedure allowing a check on the equality of intervals. The fourth study checked on whether pain was
acting as an underlying construct for ordering the faces in memory. We asked whether children perceived the set as a scale by asking
if memory for an ordered set of faces was more accurate than for a random set. The final study checked, with 6-ycm'-old children, the
test-retest reliability of ratings for recalled experiences of pain. Overall, the faces pain scale incorporates conventions used by
children, has achieved strong agreement in the rank ordering of pain, has indications that the intervals are close to equal, and is
treated by children as a scale. The test-retest data suggest that it may prove to be a reliable index over time of self-reported pain.
Key words: Pain; Measurement; Children; Facial expression
The aim of the studies presented here was to
develop a simple measure of the intensity of pakl,
to be used with children including those in the late
prelogical Piagetian stage of cognitive develop-
ment before 7 years \[32\]. The research stems first
Current address: Department of Rheumatology, The Royal
North Shore Hospital, St. Leonards, N.S.W, Australia.
Correspondence to: Dr. G.D. Champion, c/o Dr. J.B. Zieg-
!er, Division of Paediatrics, The Prince of Wales Children's
Hospital, High Street, Randwick, N.S.W. 2031, Australia.
of all from the fact that the measurement of pain
in children is a major challenge to health care
professionals. To start with, children present prob-
lems of measurement, compared with adults, be-
cause of their lower level of verbol fluency and the
likelihood that varied developmental !evds alter
their understanding of questions or tests and the
ways in which pain is expressed. In addition, the
measures aeveloped to date for children contain
some specific difficulties.
Physiological indi~es a~-c of IL,'aited value in
paediatric pain assessment, except perhaps in
neonates and infants \[34\]. Measures derived from
observations of pain behaviour correlate with
0304-3959/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
self-report \[1,22,24,27\], are unobtrusive and inval-
uable for infants, but there are doubts about the
validity in distinguishing between pain and other
forms of distress \[31,35\], pain behaviour changes
with development \[28,30\], and inter-observer vari-
abiLty is inevitable. Self-report measures of the
severity of pain have been widely used in children
of 5 years and older. There are several types of
scale including linear visual analogue scales
\[1,2,41\], graphic and numeric rating scales and
thermometer-like derivatives \[36\], verbal rating
scales \[27,31\], projective measures where pain is
inferred from drawings \[38\], selection of colours
\[13\], poker chips \[17\], pain maps with colours to
indicate intensity \[14\], and interpretation of
cartoon pictures \[26\]. It is, however, widely
acknowledged that each of these measures has
shortcomings in concept, validity, reliability, scal-
ing properties, sensitivity, or practical applicabil-
ity for children, particularly preoperational chil-
dren \[28,31,34,35,37\]. The visual and numeric rat-
ing scales and their derivatives as,~ume that the
child, irrespective of age, has the ability to "map"
accurately their own concept of pain onto a sec-
ond dimension. For example, to make a judge-
ment of pain using the visual analogue scale, the
child must translate the severity of his pain experi-
ence to a point on a line where he or she feels that
the distance between either end-point corre-
sponds. Child development literature svggests that
tiffs metaphorical ability follows a developmental
sequence and is not acquired until late in develop-
ment, during the stage of concrete operations be-
ginning at about 7 or 8 years of age \[4,18\].
The difficulties present in the current self-re-
port measures provided the impetus for several
specific goals in the present research. The aim first
of all was to place minimal cognitive demands on
children, allowing use of the measure with very
young children. This aim led us, as it has several
researchers who apparently began work on judge-
ments of faces at much the same time, to con-
centrate on assessments of facial expressions rather
than use verbal or numerical ratings. Face interval
scales measuring pain affect have been considered
preferable for younger (preoperational) children
than more abstract scales. By the age of 4 and 5,
children have the ability to distinguish between
and respond to facial features and facial patterns
exhibiting various basic emotions \[8,11,12,15,19,
39\]. Young infants produce a consistent pattern of
facial responses to pain \[9,21\]. Facial expression
may be the most consistent across-infant indicator
of pain \[23\]. As the infant develops, anger expres-
sion can be perceived \[20\] and the facial reaction
becomes more diverse with learned responses but
the basic pattern is retained in adults \[10\]. It is
expected that children as young as 4 would recog-
nize facial reactions to pain and assess the sever-
ity. A scale of faces should present little cognitive
complexity and metaphoric difficulty and should
be relatively free of adult influences of explana-
tion and interpretation.
In developing the faces scale, we have been
mindful of problems with other paediatric face
interval scales. Several faces scales, including those
by McGrath et al. \[29\], Maunaksda et al. \[27\] and
Whaley and Wont \[40\], have included a sm;2ing
face at the no-pain end. This assumes a continu-
ous pleasure-pain dimension, a physiologically
and philosophically \[33\] untenable assumption. A
well studied sca!e of faces, the Oucher, consists of
a white cra'dboard poster with 6 photographs of a
3-year-old boy in increasing levels of pain on the
right side and a 0-100 scale on the left side \[3,5,6\].
Although several points of validity have been
established, the scaling is arbitrary and, before the
ages of 7-8 years, children experience difficulty in
establishing correspondence between 2 sets or 2
orders (Piaget on space), also older children may
not relate well to this very young face. Children
may have mcwe difficulty relating their own pain
severity to a specific person in the photographs
than to diagrammatic faces.
A second aim was to pay particular attention to
the scale properties of the set of faces. Ideally, a
scale should be as sensitive as feasible (a 7-point
scale, for instance, allows more sensitive measure-
ment than a 5-point scale). It should cover the
whole dimension of pain severity. It should also
contain as far as possible equal intervals (that is,
the perceived change in severity between faces 0
and 1, or 2 and 3, should be the same as between
faces 5 and 6); a property not yet established for
scales such as the Oucher \[3,5,6\] or for the series
developed by Maunuksela et al. \[27\], Le Baron
and Zeltzer \[24\], or by Whaley and Wong \[40\].
The series by McGrath et al. \[29\] has been con-
fi='med to show unequal scale intervals. An im-
portant point about equal intervals is that they
help to ensure discriminability, i.e., that two or
more faces are not measuring roughly the same
quantity. With equal intervals, and the first face
measuring zero pain, we would have the desirable
properties of a ratio scale, enabling meaningful
arithmetic operations on the differences, the use of
geometric mean and coefficient of variation, and
the application of all the common parametric s~a-
Finally, if children's memory for the severity of
painful events is organized by a single cognitive
schema for pain, it would be expected that they
would be more accurate in recalling a set of faces
representing increased pain order than a randomly
ordered set. Such an argument is consistent with
developmental research which shows that
children's recall is often organized around taxo-
nomic or thematic constructs. Conversely, if no
single schema for pain exists, there would be little
or no difference between recall of ordered or
random face arrangements.
"ihese several consider,~tions gave rise to the 5
phases of the study:
Phase I. Children in grades 1 and 3 were asked
to draw 5 faces showing various degrees of pain.
Features used by the children were incorporated
into 5 sets of 7 drawings each.
Phase II. Pilot testing narrowed the 5 sets to 1
(the set sharing the highest degree of agreement on
an ordering for the level of pain.) In the 2 parts of
this phase, additional groups of children were then
asked to rank order the faces, and to make com-
parisons of all possible pair~.
Phase IIL This phase provided a check on the
equality of intervals (children placed faces at a
self-chosen point along a scale, a wedge without
divisions, with the end-points provided).
Phase IV. To test the spontaneous use of the
scale i~, memory, children were presented with
faces in an ordered or random presentation. Mem-
ory for the order of the faces was checked within 2
h or after a delay (overnight). The children later
were asked what emotion was shown by the faces.
Phase V. To examine test-retest reliability,
children rated painful experiences and repeated
the rating 2 weeks later.
Separate groups of chiidren were used for each
phase. Phases i-III included children from grade
1 and from grade 3 ~mean ages 6.7 and 8.7 years)
to allow analysis of effects from changes in gen-
eral cognition capacity. In phase IV only grade 3
and in phase V only grade I children participated.
The samples used for the sever~ phases of the
research are shown in Table I.
Children were drawn from 26 schools: 1 Jewish
school and 25 selected randomly from a listing of
Catholic parish schools in the metropolitan area.
All children were tested individually, at school, by
the one interviewer in phases I-IV, and by a
second interviewer in phase V.
The exact instructions given to the chi!dren in
phases I-IV are presented in the Appendix.
For phase I, children were given a black pencil
and 5 sheets of paper with oval outlires within
which to draw. Half the children in each age
group were simply instructed to dr~w 5 faces
showing degrees of pain, from no pain to a great
deal. The remaining half of each group were asked,
in addition, to create faces changing specifically
the eyes and mouth to indicate degrees of pain.
Drawings were scored for the overall consistency
of facial change including mouth, lips, cheei~s,
eyes, brow, nose ridge area, and the presence of
tears. They were scored also for the ~se of changes
in the eyes, and in the mouth (none, some, con-
sistent). "Some" indicated a just-recognizable pat-
tern of change, wb.~le "consistent" was an un-
equivocal pattern of change. Scoring was by 2
raters, with few disagreements.
The types of facial features drawn by the
children were incorporated into 5 sets of experi-
mental schematic pain faces (n = 7 in each), con-
veying changes in intensity of expressed reaction
to pain. Five adults with artistic skills were in-
PHASES OF THE STUDY WITH AGE AND SEX OF SUBJECTS
Phase School Sex (n)
M Total Mean (Range)
Derivation of faces scale
(I) Children's drawings depicting 1 12
facial reactions to pain 3 14
Studies with faces scale
(IIa) Sequencing of faces in rank of 1 14
displayed pain severity 3 12
(Ilb) Comparison of all faces in pairs 1 10
to determine which displayed more 3 13
(lid Assessment of the interval scale 1 63
properties 3 53
(IV) Test of recall of order:
Correct pain order presentation
Immediate 3 17
Delay 3 20
Random order presentation
Immediate 3 18
Delay 3 18
(V) Test-retest reliability 1 18
14 26 6.7 (6.2-7.2)
13 27 8.7 (8.1-9.6)
14 28 6.8 (6.4-7.3)
16 28 8.6 (8.2-9.4)
11 21 6.8 (6.0-7.3)
11 24 8.5 (7.9-8.9)
57 120 6.7 (6.1-7.7)
67 120 8.8 (7.1-9.8)
25 42 8.7 (7.6-9.7)
20 40 8.8 (8.0-9.7)
20 38 8.7 (7.7-10.0)
21 39 8.8 (7.5-9.9)
19 37 6.7 (6.2-7.4)
structed to create the 5 scales, based on the
children's drawings and the results of the analysis
of their content, to be used as a measure of pain.
For phase II, for each of the 5 sets of faces,
childre~ were asked either to (a) plar,~ the 7 faces
in order, from the most to the least painful, or (b)
using the method of paired comparisons to say,
for each pair of faces shown, which showed the
more pain. The order and position of the faces
presented were randomized. For (a) we de-
termined the percentage of children who agreed
c~n tho rank order of each face and the percentage
of children who agreed on the rank order of all 7
faces. For (b), we determined the percentage of
children who agreed on the order determined from
higher (pain) ranking of each pair. The set with
the highest percentages of agreement in the order-
ing (Fig. 1) was chosen for presentation of the
order agreement data (Table IIl) and for further
study. In Table I we did not include the 80 ag-
ditional children who participated in t~s pl~se
and whose contribution involved the 4 sets of
faces which were discarded because of lower per-
centages of agreement.
For phase III, children were shown a I m long
red wedge, ranging from 0 to 6 cm in width. They
were told that the width of red indicated the
amount of pain and to place the faces along the
wedge according to the pain displayed in the face.
To enable relativity of judgement while minimiz-
ing verbal instructions the "no pain" and the
"most pain possible" faces were placed at either
end of the wedge. These faces had already been
shown to be representative of none and most pain
in phase II. The other 5 faces were presented to
the child singly and in random order. Randomiza-
tion was achieved by using 5 random sequences
and ensuring that each face occurred as frequently
at the beginning as at the end. Once the child had
placed the face on the wedge, the position was
recorded and the face was removed. The investiga-
tor later placed all the faces along the wedge in
the places chosen by the child who was then given
the opportunity to make any further adjustments
to the positioning. The positions on the wedge (for
the single and for the simultaneous placements)
were compared with the expected position if inter-
vals were equal.
For phase IV, children were asked to recon-
struct a presented order of the set of faces either
immediately (within 2 h) or after a delay (over-
night). Further, the faces were presented in the
correct pain sequence or in random sequence (5
different random orders). Four independent
groups of children participated. They were at no
time made aware that the faces were representa-
tive of the construct of pain facial reactions. They
were asked after the task was completed what they
thought the faces meant.
For phase V, children were instructed to rate a
recent painful experience as to the degree of pain
(hurt) felt, using the faces scale. Six examples of
painful childhood experiences were given to aid
his or her choice: falling over while playing, re-
ceiving an injection, burning a hand on the stove,
the pulling off of a band-aid, closing the door on a
finger, and being stung by a bee. Two weeks later
the child was instructed to recall the same per-
sonal pain experience and rate it with the faces
(without prompting). Two children could not be
retested because they were absent on the second
The results for phase I are shown in Table II.
Overall, children easily grasped the idea of draw-
ing faces showing differing degrees of pain and
had no problem with the word "pain." In general,
the children at age 6 and at age 8 used the eyes
and the mouth as the principal signals of change
in pain. Some consistency of change was observed
for furrowing of the forehead, but few children
showed a consistent pattern of change for the
other features used to display pain such as tears,
use of cheeks, chin and nose. The drawings by the
8-year-old children displayed higher overall con-
sistency of change than the drawings by the 6-
year-old group (X 2 -- 9.04, df = 2, P < 0.02).
Otherwise there were no significant differences as
a result of age, the 2 age groups relying similarly
on the eyes and mouth. Instructing the children to
change the eyes and mouth did not result in
significant additional changes, although a trend to
more consistent responses in the eyes was
prompted in the 6-year-old children.
The most frequent pattern of expression with
increasing pain was irregularity of the mouth (lips)
with down-turmng at the ends leading to a square
open mouth reaction, and progressiw closing of
the eyes leading to a tightly shut reaction with
depressed brows. These were the principal features
captured by the illustrator (a professional
cartoonist) in preparing the best of the 5 sets of
schematic faces (Fig. 1). In this set, which showed
the highest levels ~,,f agreement in ordering, pro-
gressive furrowing of the brow was also incorpo-
PHASE I: CONTENT ANALYSIS OF PAIN DRAWINGS
Grade 1 children (mean age 6.7) Grade 3 children (mean age 8.7)
A * (n = 12) B ** (n = 14) A (n = 14) B (n = 13)
(~) (~) (~) (~)
Overall consistency 0 0
of facial change 1 42
Change of eyes none 17
Change of mounth none 0
0 0 0
50 7 0
21 36 31
29 57 69
0 14 0
71 43 62
29 43 38
0 0 0
64 36 31
36 64 69
A * Only degrees of pain asked for.
B * * Instructed to change eyes and mouth.
PHASE If: AGREEMENT AMONG THE CHILDREN IN THE PAIN ORDERING OF FACES IN THE SET ACHIEVING
THE HIGHEST AGREEMENT
Group of children Percentage of children who ranked the faces
Individually in the e~.pec~J.ed order
Face* 0 1 2 3 4 5 6
All 7 in the In the expected
expected order order of pairs
Grade 1 Test 96 89 82 71 75 86 89 64 62
(mean age 6.8) Retest 100 82 82 75 75 89 89 61 86
Grade 3 Test I00 100 96 86 89 100 100 86 75
(mean age 8.6) Retest 100 96 96 96 96 100 100 89 71
* Face 0 represents no pain, 6 the most pain possible.
The results for phase H are shown in Table III.
Th~ data relate to the best of the 5 sets of faces,
that is the set with the highest percentage of
children who agreed about the rank order of faces
as a measure of pain when presented with all 7
faces and with pairs for comparison. The variation
between children concerning the rank ordering for
pain severity in the best set was most evident in
the miOdle face~ (2, 3 and 4), particularly for the
younger children. The incorrect paired compari-
son judgements were not specific to any particular
face(s), that is, no systematic deficiency in the
ordering of the face drawings was evident.
The results for phase III are shown in Table IV
which displays the interval scale properties of th:
faces instrument. The assumption is that in a
perfect interval or ratio scale the faces will be
placed at equal intervals along the wedge whether
placed singly or simultaneously (all faces arranged
together). There was a statistically significant sep-
aration of the faces in the correct pain order for
each presentation (single or simultaneous) in the 2
age groups as demonstrated by the lack of overlap
of the 95~ confidence intervals. The faces scale
interv~s did not coinc,.'de with the expected equal
interval values for faces 2 and 4 and to a lesser
extent face 3, in that these faces were placed
slightly short (mean difference 5.4~) of the ex-
pected positions by the 6- and 8-year-old children.
No significant differences occurred between the
results of the 2 age groups. Significant differences
were found between s;ngle a~d simult0a~eous
placements for faces I (t = -5.89, P < 0.001) and
5 (t = 4.82, P < 0.001). There was a tendency for
these faces to be placed closer to the theoretical
equal scale interval values on simultaneous pre-
sentation. Overall, children displayed a good ap-
proximation to the positions expected on an equal
The results for phase IV are presented in Table
V. When presented in order of pain severity, the
percentages of reconstructions were 50 (im- mediate) and 77.5 (delayed). When the faces were
presented in random order, the correct recall of
the presented order was achieved in 15.8~ (im-
mediate) and 0~ (delayed) of cases. The number
of faces correctly reconstructed was significantly
greater for the pain ordered faces than for random
ordered faces (X 2 = 62.7, df= 6, P < 0.001). It
was observed that, when attempting to reconstruct
Fig. 1. The Faces Pain Scale. Numbers 0-6 left to fight. Each face is 6 cm high in the experiments in phases III-V.
PHASE III: THE EXPECTED AND OBSERVED PLACEMENT OF THE FACES ALONG THE WEDGE
Condition Distance along wedge in cm (mean t95% CI))
Face a 1 2 3 4 5
Expected (assuming equal scale interval) 16.7
Grade I children (mean age 6. 7, n = 120)
Faces placed singly
All faces placed simultaneously
Grade 3 children (mean age 8.8, n = 120)
Faces placed singly
All facez placed simultaneously
33.3 50.0 66.7 83.3
18.7 27.4 * 45.1 59.5 * 79.7
(10-22) (23-31) (40-50) (55-65) (76-84)
14.4 26.8 * 45.4 60.3 * 83.9
(12-17) (23-31) (41-50) (56-65) (81-87)
18.0 29.3 45.2 59.9 * 81.3
(15-21) (26-33) (41-49) (55-65) (78-85)
14.8 28.7 * 45.2 * 62.2 83.9
(12-17) (26-32) (42-49) (58-66) (81-87)
a Faces 0 and 6 placed at the extremes.
* P < 0.05 adjusted for multiple comparisons, t test comparisons of expected and observed place distances.
the random order presented, the mistakes made by
the children showed a tendency to placement in
correct pain order.
In response to the questions as to the meaning
of the faces, the data were not in a form suitable
for formal content analysis. Clear statements of
pain, hurt, ache, being sick, and of emotional pain
such as from teasing were made by 92 (57.9%) of
the children from both conditions. Sadness or
anger combined with hurt were not infrequently
described. Forty-one (25.8%) gave other interpre-
tations such as sadness, anger, boredom, crying
for no reason. Sixteen children (10.1%) did not
know what the faces showed, and 10 (6.3%) could
not be asked because of time limitations of their
classes. Consistent with our hypothesis of organiz-
ing schema in memory, there was a definite theme
of hurting more and more, getting sicker and
sicker or worse and worse in the group given the
faces in correct order, whereas the group given the
faces randomly consistently interpreted the fa~.~es
singly rather than collectively, i.e., lacked a theme.
In phase V, the test-retest reliability data for
6-year-old children yielded a rank correls, tion
coefficient of 0.79, indicating that the score,,, ob-
tained using the faces scale are adequately repro-
ducibie ever time. Even when the rater was varied,
in that, when test and retest data were obtained by
PHASE IV" THE PERCENTAGE OF CHILDREN WHO CORRECTLY RECONSTRUCTED THE FACFS .~N THE ORDER
Condition Number of faces correctly reconstructed
0 1 2 3 4 5 7
Prior presentation of correct order
Immediate (n = 42) 2
Delay (n = 40) 0
Prior presentation of random order
Immediate (n = 38) 8
Delay (n = 39) 13
7 2 5 10 24 50
0 0 5 7.5 10 77.5
10.5 21 26 8 10.5 16
38 18 18 10 3 0
different investigators, a high rank correlation
coefficient (0.82 where n- 35) was obtained.
In phase I, the 6- and 8-year-old children gen-
erally had clear concepts about the term pain and
what it means to experience different levels of
severity. They principally used the eyes and mouth
to display facial changes in pain reactions in their
drawings, although other features including frown
lines on the brow and chin and tears were em-
ployed by some, especially by the older children.
It was possible to incorporate these conventions to
establish a face scale, differentiating various levels
of pain. It is questionable whether the scale is
specifically a measure of the severity dimension of
pain or pain affect, or whether it is a general
measure of reaction to distress. However, support
for the construct validity of the faces scale as a
measure of pain intensity is evident in published
descriptions of facial reactions to pain employing
reliable coding measures. In infants \[16,21,23\], a
discrete facial reaction of pain appears to exist
and is described, in its full expression, as eyes
t~ghtly ~.~osed, brows lowered and together, nasal
root broadened, furrows in forehead, mouth angu-
lar and squarish, tongue taut. These are the same
appearances as in anger except that in the latter
the eyes are open and staring \[21\]. In adults sub-
jected to cold pressor pain, the facial reactions
reported by Craig and Patrick \[10\] comprised a
narrowing of the eye aperture from below, raising
the upper lip, pulling the lip corners, parting of
the lips, or dropping the jaw, and eyes closing or
frequently blinking. Similar changes in acute pain
patients were reported by Le Resche \[25\]. These
descriptions correspond with the facial pattern
which emerged empirically in t.he faces scale.
In phase H, there was good overall agreement
between the children on the rank ordering of the
faces as a measure of pain severity, an indication
of content validity of the scale. The interval scale
assessments in phase III provide strong evidence
that 6- and g-year-old children similarly interpret
and respond to the 1'ace categories as relatively
equally spaced. It was important that initially the
faces were placed singly in order to eliminate any
tendency to equal spacing of the complete set of
faces. The task was thus specific to relating facial
expression to the amount of red representative of
the amount of pain. Reviewing all the faces in
position subsequently was to test if changes would
be made towards equal spacing, thus showing
conceptually insecure judge~,,ei~.ts. In fact, little
adjustment occurred, the: Jiggest change being that
face 5 was moved closer to equ',d spacing position
by the younger children. The children had little
difficulty with the cross-modality matching using
the wedge, as the data from McGrath et al. \[29\]
had led us to expect. The approximately equal
interval scale measurement properties of this scale,
as indicated by the cross-modality matching pro-
cedure, is a pleasing result. Our scaling results
suggest that anomalies of distribution of pain
scores due to measurement bias are unlikely. The
cross-modality wedge experiment provides only
preliminary evidence for approximately equal in-
tervals. Further empirical research is needed, as
with experimental pain such as the cold pressor
The phase IV data add to the scaling/order
properties in that there was sound memory for the
order irrespective of the immediate or delayed
presentation. In particular, the increased recall in
the delayed condition provides support for the-
matic organization in memory It is thus inferred
that pain depiction may be acting as an organizing
construct in memory. We did not consider that
these young children would determine such ap-
propriate ordering if they were reliant only on
change in salient features of the face without an
underlying theme such as pain.
In phase V, the test-retest reliability study
indicated that the scores obtained using the faces
scale are relatively consistent over time, at least
when rating a recalled pain episode. The results
were minimally influenced by the person adminis-
tering the measure or by the sex and age (6
compared with 8 years) of the child.
Our first aim was to derive a scale with minimal
cognitive demands that could be reliably and
validly used in young children. The faces scale has
shown initial indications of meeting these criteria.
The scale represents the conventions used by
children to display pain. It is simple, quick to use,
non-verbal, and requires minimal instruction. The
child is told to point to the face that shows how
much pain he or she feels. The response is scored
by applying the scale number to the face chosen.
Administration is possible by a health care profes-
sional, parent or older child. The scale has been
used satisfactorily, indeed with confidence and
alacrity, by preoperational children with acute pain
and with juvenile arthritis. Children as young as 3
years have used it with apparently good compre-
We used the faces scale in a pilot study in 20
children aged 2-12 years undergoing therapeutic
procedures for acute lymphoblastic leukaemia \[7\].
The scale enabled clear discernment of different
severity of pain from needle insertion compared
with injection of volume. The median face score
(0-6) for intramuscular needle insertion was 3,
injection of volume 1, local anaesthetic for lumbar
puncture 4, venipu,~cture 2, injection of volume
into vein 0. Median values were presented as the
interval properties were still under investigation at
The second major aim was to pay particular
attention to the scale properties. This face scale
has superior scaling properties to other published
faces scales. Although we have focused on the
dimension intensity of pain, later research may
show the measure to include other dimensions
especially affect-related reaction. The scale does
not confuse the pain severity dimension with a
happiness dimension, evident in several other
scales. Further research into additional aspects of
reliability, practicability and validity is ongoing.
Additional evidence of equal intervals would be
particularly valuable to establish its legitimacy as
a ratio scale and hence suitability for arithmetic
operations on differences, the use of geometric
means and the application of all parametric statis-
This research was supported by funds from the
National Health and Medical Research Council,
We gratefully acknowledge the statistical assis-
tance of Dr. R.J. Fisher, Medical Statistician, The
Prince of Wales Hospital. We would like to thank
the schools of the Catholic Educatien Co_m__m_issio_n_
of Sydney and Moriah College for allowing the
participation of their children in this research. Dr.
L. White and the medical and nursing staff of the
Department of Paediatric Haematology/Oncology
at The Prince of Wales Children's Hospital kindly
participated. We have greatly appreciated the
guidance and extensive help with the revision of
this paper generously provided by Professor J.
Goodnow and Dr. J. Ungerer of the School of
Behavioural Sciences, Macquarie University, Syd-
ney. Miss Kate Juniper contributed extensively to
the revision of the paper and prepared the
Exact instructions to the children phases 1-IV
All instructions for phases I and II began with
the "most painful" position to start the child's
thinking on the concept of pain.
Phase I. The two types of instructions were
(a) 1. Inside this shape, draw a face that shows a
real lot of pain, as much as a person can
2. Then, on this page, inside the shape, draw a
face that shows a bit less pain.
3. And on this, a bit less again.
4. And, then, less again - just a little bit of
5. On this last one, draw a face that shows no
pain at all. The pain is all gone.
(b) 1. Inside this shape, draw a face that shows a
real lot of pain, as much as a person can
2. Then, on this page, inside the shape, draw a
face that shows a bit less pain, by changing
the look of the eyes and the mouth.
3. Then, still changing the eyes and the mouth,
draw a face with a bit less pain again.
4. Keep on changing the eyes and the mouth
in this next one, to show a bit less pain
again - only a little bit of p:an.
5. On this last page, change the eyes and the
mouth to show no pain at all. The pain is
Phase II. Two sets of instructions; 1 for each
(a) Ranking of the 7 faces.
1. I have a pack of 7 cards here; each has a
drawing of a face, showing different
amounts of pain. Some show a lot of pain,
others only some pain and even no pain at
2. I want you to look at these cards very
carefully and then put them in a line that
starts with a face that you think shows the
most pain; then one that shows less pain
and keep on going, choosing faces that
show less and less pain. End the line with
the face that you think shows no pain at all.
3. Take your time and look at each face care-
4. You can change their position on the line,
if you change your mind, by looking care-
fully at how much pain each face shows.
(b) Paired comparisons.
1. I am going to show you some pictures of
faces, 2 at a time.
2. Look at these 2 faces and choose the one
that ~/ou think shows more pain.
3. Now look at these 2 and again choose the
one that you think shows more pain.
The last instruction was repeated for each of
the remainder of 21 pairs.
Phase III. The instructions were given using a
hand to indicate areas talked about.
(a) Placement of faces singly.
1. Look at this red line that becomes bigger
and bigger in width until it is a wide wedge
at this end.
2. Imagine that red means pain and this draw-
ing shows how pain can get bigger and
bigger like the amount of red.
3. See this face is placed just before any red
begins and it shows no pain.
4. Now at this end is another face that shows
the most pain a person can have and it is
placed where the whole paper is red.
5. I will now give you some more pictures of
faces that show pain - one at a time - and
I want you to look at it carefully, then
choose the AMOUNT OF RED that you
think that each face shows. Remember red
6. Now here is another picture of a face. Look
at it carefully and then put it on the amount
of red you think it shows. Remember red
This last instruction is repeated for the next 3
(b) After this task was completed, the investigator
replaced all faces together at points chosen by
the child previously, with the following in-
1. I have now put all the faces you saw, back
in places you chose on the amounts of red
- which is the amount of pain - they
2. Now that you see them all in position to-
gether, are you happy with the amount of
red - which is the amount of pain - under
3. If you want to move any face to show a
different amount of pain you can do so.
Phase IV. General introduction to the whole
group, prior to subdividing for experimental con-
1. I am going to ask you to do some remem-
2. It might sound like a test to you, but it is
not like the tests you do for your teachers.
3. This is to help me in my studies, because I
am trying to find out how children of your
age remember different expressions on faces
that are drawn on paper.
4. So you will not be getting marks or any-
thing for it but you will be helping me find
out things about memory.
(a) Initially several children were seen at the same
time f~r looking to remember, with the follow-
ing instruction: (for both conditions)
1. I have a line of faces here that I am going
to show you for a few minutes. You will
each get a sheet of paper with a line of
2. Please look at the expressions on each face
and note carefully which face comes after
which on the line and try to remember the
3. When I see you again, a little while iater
today (tomorrow for delayed), I win want
you to remember and show me, the way the
faces come after each other in the line-up.
(b) For the memory session, children were tested
singly, with the following instruction:
1. Remember, just a while back (yesterday for
delayed condition) I showed you some faces
arranged in a line and asked you to remem-
2. Now I have exactly the same faces, but
each on a separate card.
3. I want you to look at them carefully and
arrange all the faces in exactly the same
line-up as you saw before.
4. Just try to remember as well as you can,
but do not worry if you are unsure. It is not
like a school test. I just want to know what
you can remember.
At completion of the task, children were asked:
5. Can you now tell me what you think these
1 Abu-Saad, H., Assessing children's responses to pain, Pain,
19 (1984) 163-171.
2 Abu-Saad, H. and Holzemer, W.L., Measuring children's
self-assessment of pain, Iss. Compr. Pediat. Nurs., 5 (1981)
3 Aradine, C.R., Beyer, J.E. and Tompkins, J.M., Children's
pain perception before and after analgesia: a study of
instrument construct validity and related issues, J. Pediat.
Nurs., 3 (1988) 11-23.
4 Asch, S.E. and Nedove, H., The development of double
function terms in childr, a: an exploratory investigation. In:
B. Kaplan and S. Wapner (Eds.), Perspectives in Psycho-
logical Theory, International Universities Press, New York,
5 Beyer, J.E. and Aradine, C.R., Content validity of an
instrument to measure young children's perceptions of the
intensity of their pain, J. Pediat. Nurs., 1 (1986) 386-395.
6 Beyer, J.E. and Aradine, C.R., Patterns of pediatric pain
intensity: a mett~odological investigation of a self-report
scale, Clin. J. Pain, 3 (1987) 130-141.
7 Bieri, D., White, L., O'Gorman Hughes, D., Vowels, M.,
Newsum, M., Warren, J. and Miller, C., Communication of
self perceived pain in children, Med. Pediat. Oncol., 13
8 Charlesworth, W.R. and Kreutzer, M.A., Facial expressions
of infants and children. In: P. Ekman (Ed.), Darwin and
Facial Expression, Academic Press, New York, 1973, pp.
9 Craig, K.D., McMahon, R.J., Morison, J.D. and Zaskow,
C., Developmental changes in infant pain expression during
immunization injections, Soc. Sci. Med., 19 (1984) 1331-
10 Craig, K.D. and Patrick, C.J., Facial expression during
induced pain, J. Pers. Soc. Psychol., 48 (1985) 1080-1091.
11 Ekman, P. and Friesen, W., Constants across cultures in the
face and emotion, J. Pers. Soc. Psychol., 17 (1971) 124-129.
12 Ekman, P. and Oster, H., Facial expressions of emotion,
Annu. Rev. Psychol., 30 (1979) 527-554.
13 Eland, J.M., Minimizing pain associated with prekinder-
garten intramuscular injections, Iss. Compr. Pediat. Nurs.,
5 (1981) 361-372.
14 Eland, J.M. and Anderson, J.E., The experience of pain in
children. In: A. Jacox (Ed.), Pain: a Source Book for
Nurses and Other Professionals, Little, Brown, Boston,
MA, 1977, pp. 453-473.
15 Felleman, E.S., Barden, R.C., Carlson, C.R., Rosenberg, L.
and Masters, J.C., Children's and adults' recognition of
spontaneous and posed emotional expressions in young
children, Dev. Psychol., 19 (1983) 405-413.
16 Grunau, R.V.E. and Craig, K.D., Pain expressions in
neonates: facial action and cry, Pain, 28 (1987) 395-410.
17 Hester, N.K.O., The preoperational child's reaction to im-
munization, Nurs. Res., 28 (1979) 250-255.
18 Ivhelder, B. and Piaget, J., The Growth of Logical Thinking
from Childhood to Adolescence, Basic Books, New York,
19 lzard, C.E., The Face of Emotion, Appleton-Century-
Crofts, New York, 1971.
20 Izard, C.E., Hembree, E.A., Dougherty, L.M. and Spizzirri,
C.C., Changes in facial expressions of 2-19-month-old
infants following acute pain, Dev. Psychol., 19 (1983) 418-
21 Izard, C.E., Huebner, R.R., Risser, D., McGinnes, G.C.
and Dougherty, L.M., The younger infant's ability to pro-
duce discrete emotion expressions, Dev. Psychol., 16 (1980)
22 .lay, 8., Ozolins, M., Elliott, C.H. and Caldw~ll, S., Assess-
ment of children's distress during painful medical proce-
dures, Hlth. Psychol., 2 (1983) 133-147.
23 Johnston, C.C. and Strada, M.E., Acute pain response in
infants: a multidimensional description, Pain, 24 (1986)
24 Le Baron, 8. and Zeltzer, L., Assessment of acute pain and
anxiety in children and adolescents by self-reports, observer
reports, and a behaviour checklist, J. Cons. Clin. Psychol.,
52 (1984) 729-738.
25 Le Resche, L., Facial expression in pain: a study of candid
photographs, J. Nonverb. Behav., 7 (1982) 46-56.
26 Lollar, D.J., Smith, S.J. and Patterson, D.L., Assessment of
pediatric pain: an empirical perspective, J. Pedlar. Psychol.,
7 (1982) 267-277.
27 Maunuksela, E.L., Olkkala, K.T. and Korpela, R., Mea-
surement of pain in children with self-reporting and be-
havioural assessment, Clin. Pharmacol. Ther., 42 (1987)
28 McGrath, P.A., An assessment of children's pain: a review
of behavioural, physiological and direct scaling techniques,
Pain, 31 (1987) 147-176.
29 McGrath, P.A., DeVeber, L.L. and Hearn, M.T., Multidi-
mensioiml pain assessment in children, Adv. Pain Res.
Ther., 9 (1985) 387-393.
3~ ~4cGrath, PJ. and Craig, K.D., Developmental and psy-
chological factors in children's pain, Pediat. Clin. N.Am.,
34 (1989) 823-836.
31 M¢~rath, P.J. and Unruh, A.M., The measurement and
assessment of pain. In: P.J. McGrath and A.M. Unruh
(Eds.), Pain in Children and Adolescents, Eisevier,
Amsterdam, 1987, pp. 73-104.
32 Piaget, J. and Inhelder, B., F~ychology of the Child, Basic
Books, New York, 1969.
33 Popper, K., The Open Society and its Enemies, Vol. 1, 5th
edn., Routledge and Kegan Paul, London, 1966, pp. 284-
34 Ross, D.M. and Ross, S.A., Assessment of pediatric pain:
an overview, Iss. Compr. Pediat. Nuts., 11 (1988) 73-91.
35 Ross, D.M. and Ross, S.A., Assessment and measuremein.
In: Childhood Pain. Current Issues, Research, and Man-
agement, Urban and Schwarzenberg, Baltimore, MD, 1988,
36 Szyfelbein, S.K., Osgood, P.F. and Can', D.B., The assess-
ment of pain and plasma beta-endorphin immunoactivity
in burned children, Pain, 22 (1985) 173-182.
37 Thompson, K.L. and Varni, J.W., A developmental cogni-
tive-biobehavioural approach to pediatric pain assessment,
Pain, 25 (1986) 283-296.
38 Unruh, A., McGrath, P., Cunningham, S.J. and Humphreys,
P., Children's drawings of their pain, Pain, 17 (1983) 385-
39 Walden, T.D. and Field, T.M., Discrimination of facial
expressions by preschool children, Child Dev., 53 (1982)
40 Whaley, L. and Wong, D.L., Nursing Care of Infants and
Children, Mosby, St. Louis, MO, 1987, pp. 1068-1070.
41 White, J.B. and Stow, P., Rationale and experience with
visual analogue toys, Anaesthesia, 40 (1985) 601-603.