lnterrater Reliability of Hand-Held
Dynamometry: Effects of Rater Gender, Body
Weight, and Grip Strength
Carolyn 7. Wadsworth, MS, PT'
David H. Nielsen, PhD, PT2
Diana S. Corcoran, MPT3
Connie E. Phillips, MPT3
Teresa 1. Sannes, MPT3
tate-of-the-art practice,
with increasingly sophisticated equipment, more
exacting health-care practitioners, knowledgeable
consumers, and stringent reimbursement policies, is challenging today's
physical therapists to more accurately and objectively document a
patient's condition. Nowhere are the
demands more urgent than in the
area of strength testing. Hand-held
dynamometry (HH D) has emerged
as a quantitative, efficient method of
obtaining objective clinical measurements of strength (5, 10, 12, 17).
Several studies have addressed both
the intrarater and interrater reliability of HHD (1-4, 6, 7, 8, 20, 22,
23). Although most studies have
found good-to-high reliability, some
have yielded conflicting results.
Byl et al reported intrarater reliability coefficients for dynamometry
measurements ranging from 0.833
to 0.957 (higher values occurred
when the examiners used additional
stabilization) and interrater reliability coefficients from 0.5 18 to 0.840
in normal subjects (7). In a weaker
patient population, Bohannon found
correlations ranging from 0.84 to
0.99 for test-retest scores of 18 mus-
Clinicians and authors of previous publications have not reached agreement on the interrater
reliability of dynametric strength testing. This study investigates the effects of gender, body weight,
and grip strength on the reliability of hand-held dynametric strength measurements. Ten male and
10 female raters tested five muscle groups on the same two subjects (one male and one female) with
a Chatillon Series D hand-held, spring-scale dynamometer. Both the raters and the test subjects
were blinded to the dynametric output readings throughout the testing. lnterrater reliability was
good for all tests except for female raters when testing the male subjects' stronger muscle groups.
Standard deviations were 61% and 50% greater for female vs. male raters for elbow flexors and
knee extensors, respectively. female raters' body weight had a significant correlation with torque
when testing male subjects' wrist extensors, ankle dorsiflexors, and knee extensors (r > 0.64).
likewise, female raters' grip strength significantly correlated with torque when testing males' wrist
extensors and elbow flexors (r z 0.71). The results indicate that gender, body weight, and grip
strength affect a rater's ability to stabilize a hand-held dynamometer and could influence reliability
when testing stronger muscle groups.
Key Words: hand-held dynamometry, strength testing, reliability
' Lecturer, Physical Therapy Graduate Program, College of Medicine, The University of lowa, lowa City, IA
Associate professor, Physical Therapy Graduate Program, College of Medicine, The University of lowa, lowa
City, IA
Mrs. Corcoran, Mrs. Phillips, and Mrs. Sannes were students in the Masters of Physical Therapy Program,
Colleae of Medicine. The Universitv of lowa, lowa Citv, IA, when the studv was conducted.
cle groups tested by an experienced
clinician (4). Bohannon also obtained
good interrater reliability for two
raters on six muscle groups of neurological patients, with correlation
coefficients ranging from 0.84 to
0.94 (2). He stated, however, that
significant differences in the mean
ratings for two of the muscle groups
indicate a need for further evaluation of HHD. Agre et al demonstrated strong interrater reliability
for upper extremity muscle tests,
with correlation coefficients ranging
from 0.85 to 0.99. However, they
found poor reliability for lower extremity muscle tests, with correlation
coefficient values ranging from
-0.20 to 0.96 (1).
A number of investigators recognize that dynamometer stabilization
is of major importance to the reliability of HHD strength testing (1-4,
7, 14, 20, 22). Byl et al state that
Volume 16 Number 2 August 1992 JOSPT
since a hand-held dynamometer is
not a fixed device, reliability will depend on a rater's ability to stabilize
the instrument (7). Agre et al found
HHD to be reliable for upper extremity strength testing, due to the
ease in which the subject's body
parts could be positioned and stabilized, but not reliable for lower extremity strength testing (1). They
concluded that since the lower extremity body parts were larger, it
was more difficult for the examiner
to provide sufficient stabilization in
the standardized testing position (1).
Hosking et al concluded that the limiting factor in measuring more powerful muscle groups, such as hip flexors and knee extensors, with the
Hammersmith Myometer was the examiner's ability to oppose the subjects' contractions. They recommended that muscle strength measurement with a myometer be limited
to monitoring weak muscles in diseased conditions o r smaller muscle
groups in stronger populations (I 2).
Dynamometer stabilization requires that the rater meet the force
produced by the patient o r subject.
Bohannon reported that in studies
with patients rather than healthy individuals as volunteers, the patients
were generally weaker and, therefore, easier to test (2-4). Bohannon
also discussed that stabilization and
meeting force output may be particularly difficult for clinicians who are
not physically strong (4). Wadsworth
et al discussed similar findings, proposing that an examiner's own
strength may affect his or her ability
to use a dynamometer appropriately
with stronger subjects (22). Marino
also suggested that a rater's strength
influences his o r her ability to stabilize a dynamometer ( 1 4).
Since a rater's strength appears
to influence his o r her ability to stabilize a dynamometer, it is relevant
to ask: "What is the minimal strength
necessary to reliably perform HHD?"
and "What other parameters related
to strength affect dynamometer stabilization?" Previous research has
JOSPT Volume 16 Number 2 August 1992
shown that overall body strength
correlates with gender (16, 18). body
weight (13, 2 I), and grip strength
(1 1). But in spite of the association
of these variables with overall
strength, and the assumption that
strength influences dynamometer
stabilization, no publication has addressed the correlation between
these factors and reliability in HHD.
T h e purpose of this study was to
investigate the effects of gender,
body weight, and grip strength on
the rater's ability to determine muscle performance using a hand-held
dynamometer and to examine the
impact of these variables on consistency of measurement between
raters. In this context, stabilization
was defined as the ability of the rater
to keep full dynamometer contact on
an extremity in the appropriate perpendicular position and to maintain
an isometric contraction.
EXPERIMENTAL DESIGN
T h e experimental design was a
time series quasiexperimental model
consisting of two groups of raters
(10 male and 10 female) and repeated measures (three trials) on five
muscle groups, each performed on
one male and one female subject.
T h e five muscle groups tested were
the wrist extensors (WEX), ankle
dorsiflexors (DFL), shoulder abductors (SAB), elbow flexors (EFL),
and knee extensors (KEX). T h e
Muscle Group
- ---- -
RESEARCH S T U D Y
-.--
-
-
.-
-
.
specific test positions were adopted
from Smidt (19) and Daniels and
Worthingham (9). with slight modifications (see Table 1). T h e treatment
assignment (testing order) was determined by the Latin Square Technique.
Equipment
Raters used a calibrated ChatilIon (John Chatillon & Sons, 83-30
Kew Gardens Rd, Kew Gardens,
NY 1 1415) Series D hand-held,
spring-scale dynamometer with a
maximum loading capacity of 150
Ibs (68.2 Kg) and a built-in peak
force indicator, which was zeroed
prior to each test trial. T h e dynamometer attachments included two
handles and a padded concave cuff
(Figure 1). T h e rater's grip strength
was measured with a Jamar handheld, spring-scale dynamometer with
a maximum loading capacity of 90
Kg-
Subjects
T h e authors selected one man
and one woman as test subjects. Both
were 24 years of age, with heights of
I78 cm and 157 cm and weights of
77.3 Kg and 56.5 Kg, respectively.
Criteria for subject eligibility were:
I ) no current or previous history of
neuromuscular deficits and 2) a normal grade on the manual muscle
tests for the five muscle groups
Limb Position
Dynamometer Placement
Wrist extensors
(WEXI
Forearm supported on armrest, elbow
flexed, wrist in 15' extension
Ankle dorsiflexors
(DFL)
Knee at 90". ankle at 0"
Shoulder abductors
90" of shoulder abduction, elbow
fully extended, forearm pronated
Just proximal to metacarpophalangeal joints on extensor surface of hand
just proximal to metatarsophalangeal joints on dorsal surface of foot
Just proximal to ulnar styloid
process on extensor surface
of forearm
Just proximal to wrist joint on
flexor surface of forearm
(SABI
Elbow flexors (EFL)
Knee extensors (KEX)
Arm beside trunk, elbow flexed to
90°, forearm supinated, wrist in
neutral
Knee in 150' of extension
Just proximal to ankle on anterior surface of leg
TABLE 1. Specific testing positions of each muscle group.
75
R-E-S-E-A-R- C
H STUDY
-- .-.
.
-
-
over a 2-week time period. This
physically strong raier, with high
body weight and grip strength, was
chosen to ensure optimal stabilization for all test conditions. This rater
was also a physical therapy student
who had previous experience with
HHD, similar to that of the experimental raters. One-way analysis of
variance of the mean torque values
for each subject revealed no significant betweenday differences. With
subject reliability established, any inconsistencies in the repeated
strenah
., measurements r e ~ o r t e din
future testing sessions could be attributed to characteristics of the
raters and not the subjects.
I
I
FIGURE 1. Chatillon hand-held dynamometer with a
maximum loading capacity of 68.2 Kg (I50 Ibs).
being examined. T h e study was a p
proved by the Human Subjects Review Committee of the College of
Medicine at the University of Iowa.
Informed written consent was obtained from the test subjects and
raters prior to participation.
Raters
T h e raters who performed the
strength tests were 20 healthy volunteers, including 10 men (mean age =
25 f 2.8 years, mean body weight =
85.2 f 27.3 Kg, mean grip strength
= 52.6 f 10.2 Kg) and 10 women
(mean age = 27 f 3.1 years, mean
body weight = 64.5 f 7.4 Kg, mean
grip strength = 28.5 f 7.3 Kg). All
raters were physical therapy students
who had similar previous experience
with HHD muscle testing through
related course work.
lntrasubjed Reliability Test
Fundamental to the study design
was the ability of the two subjects to
generate consistent forces during all
tests. T o verify consistency of testing
prior to the initiation of the primary
study, an additional male rater (age
26, body weight 1 13.6 Kg, and grip
strength of 64 Kg) administered the
test battery according to the study
protocol, testing the two study subjects five times on alternate days
0
Procedure
Orientation sessions were held
for the raters prior to the testing sessions. During this time, the raters
were given a verbal and written explanation of the study. T h e raters
were instructed in the isometric
"make test" protocol (3) and testing
technique to use during the testing
sessions. T h e raters practiced on the
test subjects using the dynamometer
on all five muscle groups with the
designated technique and verbal
commands.
Prior to the actual testing sessions, the testing order of each rater,
the order of the muscle groups
tested, and the order in which subjects were tested was randomly assigned. T h e testing sessions were
conducted 1 day a week for 5 consecutive weeks. On a given day, four
different raters (two male and two
female) performed dynametric testing of the five muscle groups on the
two subjects. Two investigators attended each test session and helped
with the preliminary test procedures
and the recording of the data.
At the beginning of each test session, the rater's body weight was
measured to the nearest quarterpound with a standard clinical balance scale. Grip strength was measured to the nearest 0.5 Kg with a
Jamar dynamometer (Jamar hydraulic hand dynamometer, Therapeutic Equipment Corporation,
Clifton, NJ 070 15). T h e raters performed three trials of grip strength
with their dominant hand while
maintaining 90" of elbow flexion
and a neutral forearm position.
Hand-held dynamometry testing
was conducted with the subjects sitting on a padded seat (backrest at
95" and feet off floor). An armrest
supported the forearm with the elbow at 90" of flexion. No manual o r
external stabilization was used other
than the back and forearm supports.
Raters tested only the right side of
the body. T o ensure test consistency,
one investigator premarked the cuff
position on the subject's arm with a
felt tip pen and used a goniometer to
place the subject's limb in the correct position for each test (Figure 2).
T h e verbal commands for each
isometric test effort were "slowly
push, push, push, push, relax" so
that the contraction lasted approximately 4-5 seconds. T h e raters were
required to grip both handles of the
dynamometer to keep the dynamom-
dynamometer force-indicating dial hidden hom
view. Investigator goniometr~allymonitors the joint
angle.
Volume 16 Number 2 August 1992 JOSPT
-.
eter in place and perpendicular t o
the limb segment and t o maintain a
counter force equivalent t o that exerted by the subject. To be considered a "stabilized" test, the rater was
not t o allow the subject t o move
more than 20" beyond the starting
angle o r allow the cuff t o lose more
than 50% contact with the subject's
skin. T h e raters knew the loss of stabilization criteria. In the event that
the criteria were not met, the test
was still completed and trial readings
recorded. O n e investigator was responsible for determining if stabilization criteria were met and for recording the muscle tests in which
loss of stabilization occurred (Figure
3). After the rater completed the
strength test, another investigator
read the dynamometer and recorded
the reading t o the nearest halfpound. T h e dynamometer force-indicating dial was covered during the
test so that both the subject and
rater were blinded t o the reading.
T h e above procedures were considered t o be one trial. Each muscle
group was tested in sequence according t o the above protocol and prede-
termined testing order. T h e test sequence was repeated two times, and
the means of the three individual
muscle trials were used for statistical
analysis. In order t o allow ample
time for muscle recovery and avoid
subject and rater fatigue, 5-minute
rest periods were provided between
test sequences and between test subjects (1 5).
T o calculate torque, the recorded dynamometer force measurements were converted t o N, then
multiplied by the distance between
the joint axis and midcuff placement
for each muscle group.
Statistical Analysis
Descriptive statistics, means, and
standard deviations were calculated
for each group of raters for both the
male and the female subjects for the
five muscle groups. Pearson product
moment correlation coefficients
were computed for body weight and
grip strength versus torque for individual muscle groups t o show the influence of rater body weight and
grip strength on strength torque
measurements. Analysis of variance
(ANOVA) was used t o test for interaction and main effects of raters (female and male) and muscle groups
independently for each of the two
subjects. Bonferroni adjusted t-tests
were used for follow-up pairwise
contrasts between raters for each
muscle group. A probability of p <
0.05 was considered the criterion for
statistical significance. Intraclass correlation analysis was not performed
since the study was limited to only
two test subjects and the raters were
not randomly selected t o assure
heterogenous representation.
Hypotheses
FIGURE 3. Testing wrist extension, with rater
maintaming perpendicular dynamometer contact
and investigator monitoring for loss of stabilization.
JOSPT Volume 16 Number 2 August 1992
This study evaluated the interrater reliability of dynametric strength
testing by examining the consistency
of the torque measurements among
and between two groups of raters
(one group of 10 female raters and
one group of 10 male raters) for five
-.
DY
- - -- - R E S E A-R-C H S T.U
.
selected muscle groups for each of
two test subjects. Five muscle groups
were selected t o represent a functional range of strength measurements from weak t o strong. Male
and female test subjects and raters
were employed to further ensure
variation in strength as well as body
stature. Table 2, section A, specifies
the null hypotheses that were tested
according to each muscle group.
T h e alternate hypotheses follow in
section B.
RESULTS
Tables 3 and 4 summarize the
group means and standard deviations for the different muscle groups
for the female subject and the male
subject respectively. As expected,
there were appreciable differences in
torque values between different muscle groups. This reflects the basic design of the study in which selected
muscles were selected t o produce a
wide range of strength.
As seen in Table 3 there were
little differences between the female
vs. male raters' mean torque values
in any one muscle group. T h e variability in measurement, as indicated
by the standard deviations, was similar between female and male raters.
T h e lack of between female vs. male
rater differences in mean strength
measurements and the relatively
small magnitude of the standard deviations suggest reasonably good interrater reliability for the female test
subject.
Table 4 indicated similar trends
for the first three muscle groups for
the male test subject. For the last
two muscle groups, EFL and KEX,
however, in which strength measurements were considerably higher,
there were appreciable differences
small magnitude of the standard deviations suggest reasonably good interrater reliability for the female test
subject.
Table 4 indicated similar trends
for the first three muscle groups for
the male test subject. For the last
RESEARCH STUDY
TABLE 2. Hypotheses.
Muscle
Croup
WEX
DFL
SAB
EFL
KEX
Mean Torque (NM)
Female
Male
Raters
Raters
(N = 10)
(N = 10)
5.0
14.2
25.8
29.6
68.8
5.4
14.9
24.9
31.1
72.4
Standard Deviation
Female
Raters
(N = 10)
Male
Raters
(N = 10)
0.4
2.4
2.7
1.5
7.8
0.7
2.1
2.5
1.6
4.7
WEX (wrist extensors)
DFl (ankle dorsiflexors)
SAB (shoulder abductors)
Efl (elbow flexors)
KEX (knee extensors)
TABLE 3. Means and standard deviations for male and female raters according to muscle group for female
subject.
Muscle
Croup
-
WEX
DFL
SAB
EFL
KEX
Mean Toque (NM)
Female
Male
Raters
Raters
(N = 10)
(N = 10)
15.4
20.4
78.4
68.0
97.6
16.3
21.6
79.6
81.1
118.3
Standard Deviation
Female
Male
Raters
Raters
(N = 10)
(N = 10)
2.7
3.3
5.0
10.7
14.5
2.0
2.0
4.8
4.2
7.2
kVtX (wr~stextensors)
DFL (ankle dorsiflexors)
SAB (shoulder abductors)
EFL (elbow flexors)
KEX (knee extensors)
TABLE 4. Means and standard deviations for male and female raters according to muscle group for male
subject.
78
two muscle groups, EFL and KEX,
however, in which strength measurements were considerably higher,
there were appreciable differences
between the female and male raters,
with the female raters' mean torques
being less than the male raters' mean
torque values. For the EFL, the difference was 13.1 Nm and for KEX
the difference was 20.7 Nm. For
EFL the standard deviation was 10.7
Nm for the female rater compared
to 4.2 Nrn for the male rater. Similarly for KEX, the standard deviation for the female rater was 14.5
Nm compared to 7.2 Nm for the
male rater. These results suggest
that for female raters, interrater reliability is compromised during testing
of stronger muscle groups. A contributing factor could be the greater
amount of stabilization required for
stronger muscle groups.
During testing, female raters lost
stabilization more frequently when
testing the EFL and KEX compared
t o the other muscle groups. T h e loss
of stabilization records indicated that
when testing the male subject, female raters lost stabilization 12 times
on the EFL and 15 times for the
KEX. Male raters did not lose stabilization on the EFL and lost it only
once when testing the KEX.
T h e ability t o adequately stabilize the dynamometer may be d u e t o
rater body weight and grip strength,
in addition t o the gender differences
just noted. Pearson product moment
correlation analysis was used t o address this question. Table 5 summarizes the correlation coefficients between body weight and torque and
Table 6 summarizes the correlators
between grip strength and torque.
T h e r e were significant positive correlations between the torque measurements for female raters testing
the male subject for WEX, EFL, and
KEX. Significant positive correlations in male and female raters were
also seen for grip strength and
torque measurements for female raters testing male subjects' WEX and
EFL (Table 6). Male raters, while
testing the female subject, showed
Volume 16 Number 2 August 1992
JOSPT
RESEARCH S T U D Y
icant rater by muscle interaction.
T h e main effect F-test was significant for muscle but not significant
for rater. Since between-musclegroup differences were not a primary concern of this study, no follow-up analysis was performed on
this variable. T h e results of this
ANOVA suggest that female and
male raters were equally consistent
negative correlations for grip
strength vs. EFL and KEX torque
measurements.
Analysis of variance was used to
evaluate the consistency of the
strength measurements between the
female vs. male raters for the different muscle groups. As seen in Table
7, ANOVA revealed that on the female test subject, there was no signif-
Female
Male
Female
Male
Female
Male
-0.21
-0.51
0.78'
-0.60
0.24
-0.46
0.62
-0.16
-0.08
-0.08
0.48
0.44
(
-(
(
-(
' p < .05
W E X (wrist extensors)
DFL (ankle dorsiflexors)
SAB (shoulder abductors)
EFL (elbow flexors)
KEX (knee extensors)
TABLE 5. Pearson product moment correlation coefficients-Body weight vs toque according to rater and
subject gender.
Subject
Raten
Female
Female
Male
Female
Male
Male
Muscle Croup
WM
0.08
-0.46
0.83'
-0.22
0.19
-0.58
0.50
-0.1 1
SAB
E
n
KM
-0.47
-0.21
0.28
0.1 7
0.16
-0.71'
0.76'
-0.33
0.12
-0.78'
0.48
0.03
' p < .O5
W E X (wrist extensors)
DFL (ankle dorsiflexors)
SAB (shoulder abductors)
EFL (elbow flexors)
KEX (knee extensors)
TABLE 6. Pearson product moment correlation coefficients-grip strength w. toque according to rater and
subject gender.
A. Female Subject
Source
DF
SS
MS
F
P
Rater
Muscle
Rater by muscle
Error
1
4
4
72
27.86
50493.44
54.87
628.56
27.86
12623.36
13.72
8.73
1.31
1445.98
1.57
0.2676
0.0001
0.1912
Source
DF
SS
MS
F
Rater
Muscle
Rater by muscle
Error
1
4
4
72
1376.41
126987.87
1650.88
1959.95
1376.41
34246.97
412.72
27.22
10.99
1166.25
15.16
6. Male Subject
0.0038
0.0001
0.0001
TABLE 7. Analysis of variance summary of dynametric strength (torque) measurements for female subject and
male subject.
JOSPT Volume 16 Number 2 August 1992
in performing the strength test on
the female subject. As graphically illustrated in Figure 4, the group
mean values for male and female raters were almost identical, and standard errors were small.
Analysis of variance on the male
subject revealed different findings
(Table 7). T h e F-test for rater by
muscle interaction was highly significant. This result suggested nonparallel strength measurement between
male and female raters for the respective muscle groups. T h e significant double interaction negated
looking at the main effects for muscle and rater. Figure 5 presents the
results of the Bonferroni adjusted
t-tests, which were utilized in the
post hoc analysis. As indicated, the
pairwise between-rater contrasts for
EFL and KEX were significant. In
both cases, the mean torque values
were less for the female raters compared to male raters, suggesting impaired measurement reliability.
Again, no between muscle group follow-up analysis was done for the reason stated previously.
DISCUSSION
T h e present study is based on
the following assumptions: I ) a
rater's ability to stabilize a dynamometer is positively correlated with
his o r her gender, body weight, and
grip strength, and 2) dynamometer
stabilization affects a rater's reliability in obtaining torque measurements.
Bohannon suggests that for
healthy, normal subjects, less physically strong clinicians may have difficulty stabilizing the dynamometer
(4). Morrow and Hosler found that
men are significantly stronger than
women in upper and lower body
strength (1 6). Men generally possess
greater body weight and grip
strength, which correlate with general strength, providing them with a
strength advantage over females.
Given this inherent strength difference, the authors expected male
raters in comparison with female
RESEARCH STUDY
-
-
---
-..-----..
-
.
- .
- - ..
Famale m t e n
nale m t e n
60
L
w r l r t Ent.
Dorslflsn.
Shoulder lbd.
Elbow Flen.
Knsa ~ n t .
Muscle Group
FIGURE 4. Muscle strength for female test subject (means and standard errors). No significant painvise
contrasts.
0
Female n t e n
Male m t e n
s
t E n
Oorslflen.
Shoulder l b d .
Elbow Flen.
Knea Ent.
Muscle Group
FIGURE 5. Muscle strength lor male test subject (means and standard errors). 'Significant painvise contrast.
raters in the study to be able t o stabilize a dynamometer more effectively,
leading t o higher torque measurements and a higher interrater reliability.
T h e study results showed that
both male and female raters had reasonably good interrater reliability
when testing the female subject.
This finding was expected because
the female subject was not capable of
generating large forces. Male raters
testing a male subject also had good
interrater reliability. However, female raters testing a male subject
had poor interrater reliability for
EFL and KEX, the two strongest.
T h e hypotheses for the present
study included not only gender
strength differences but also the
variables of rater body weight and
grip strength. Studies by Jones (1 3)
and Viitasalo (2 1) both found a positive correlation between isometric
strength and body weight. This may
be o n e reason why the male raters,
with a higher mean body weight, obtained greater torque measurements
than the female raters, with a lower
mean body weight, for the stronger
muscle groups. Another reason may
be due t o the male raters' ability t o
utilize their body weight as leverage
during the testing technique. T h e female raters' lower torque measurements and increased variability may
be attributed t o more frequent loss
of stabilization on EFL and KEX.
Along with body weight, the authors chose to look a t grip strength
as another factor of dynamometer
stability because grip strength is r e p
resentative of overall body strength,
as Everett and Sills report (1 1). Grip
strength is also important because
the raters were required t o stabilize
the dynamometer by having a firm
grip on the dynamometer before
trying t o resist another force. T h e
present investigators found that
greater grip strengths did indeed
correlate with greater torque values
in stronger o r harder to stabilize
muscle groups.
This study showed that when
forces were low, as in the female
subject, raters had n o difficulty stabilizing the dynamometer, and there
were n o significant correlations between body weight and grip strength
for either group of raters testing the
female subjects (Tables 5 and 6).
However, when forces were high, as
in the stronger male subject, female
raters' body weight and grip
strength both became significant factors when testing the WEX and EFL.
Body weight appears t o be more important than grip strength in testing
KEX. Body weight and grip strength
were not significant for DFL o r
SAB.
Due t o the testing position for
WEX and EFL, it was difficult for
the raters to use their body weight as
leverage. Therefore, the authors
speculate that the positive correlations of body weight and grip
strength t o torque measurements
were probably due t o the relationship of body weight and grip
strength t o overall body strength.
T h e significant correlation between
torque measurements of KEX and
body weight was expected d u e t o the
raters' ability t o use their body
Volume 16 Number 2 August 1992 JOSPT
weight as leverage while testing this
muscle group.
Unexpectedly, significant negative correlations were found f o r grip
strength vs. torque f o r male raters
while testing the female subject f o r
E F L and K E X (Table 6). Although
n o significant negative correlations
were found between the weight o f
the male raters and the measurement o f torque, n o physiological rationale can be given for this later
finding. A possible explanation could
be that stronger, heavier males,
while testing weak females, may be
afraid o f overpowering their subjects, resulting in lower strength
measurements. Or, perhaps, the
male raters are too aggressive in
their resistance, and they overcome
their subjects. An experimental limitation o f the present study i s that
only two subjects, one female and
one male subject, were employed.
Inclusion o f additional test subjects
would have enabled a more thorough investigation o f the negative
correlations found with the male
raters. Also, additional test subjects
would make the results more generalilable.
Follow-up studies are warranted
t o determine the limits o f body
weight and grip strength that are
needed t o stabilize a dynamometer
in order t o obtain reliable and accurate strength measurements for
given muscles. This type o f information could possibly be used as a
guideline f o r individual raters t o objectively assess their reliability using
a hand-held dynamometer. In addition, a study investigating the independent effects o f gender vs. body
weight and grip strength would further delineate rater characteristics
that affect interrater reliability in
HHD.
CONCLUSIONS
T h i s study confirms that gender,
body weight, and grip strength, as
indicators o f a rater's strength, d o
affect a rater's ability t o stabilize a
hand-held dynamometer and influence the consistency o f measuring
JOSPT Volume 16 Number 2
August 1992
torque with a hand-held dynamometer. Gender, body weight, and grip
strength o f the rater d o not appear
t o influence interrater reliability o f
H H D strength testing f o r weaker
subjects o r weaker muscle groups of
stronger subjects, b u t gender, body
weight, and grip strength d o appear
t o reduce the measurement accuracy
and, subsequently, the reliability o f
H H D strength testing o f stronger
muscle groups o f stronger subjects.
T h e relationships described in
this study demand that body weight
and grip strength be considered as
factors in future reliability research.
These findings have clinical relevance as well. An individual must determine if he/she is strong enough t o
stabilize a hand-held dynamometer
for each individual patient. Ifthere
is a question, then another fixed
strength assessment tool should be
used.
JOSPT
REFERENCES
I.Agre /C, Magness IL, Hull SZ, Wright
KC, Baxter TL, Patterson R, Stradel L:
Strength testing with a portable dynamometer: Reliability for upper and
lower extremities. Arch Phys Med Rehabil68:454-457, 1987
2. Bohannon RW, Andrews AW: Interrater reliability of hand-held dynamometry. Phys Ther 67:93 1-933, 1987
3. Bohannon RW: Make tests and break
tests of elbow flexor muscle strength.
Phys Ther 68: 193- 194, 1988
4. Bohannon RW: Test-retest reliability of
hand-held dynamometry during a single session of strength assessment.
Phys Ther 66:206-209, 1986
5. Bohannon RW: Upper extremity
strength and strength relationships
among young women. IOrthop Sports
Phys Ther 8: 128- 133, 1986
6. Burdett RC, Whitney SL: Reliability of
hand-helddynamometry in measuring
muscle strength. Phys Ther 67:748,
1987
7. By1NN, Richard S, Asturias I:lntrarater
and interrater reliability of the biceps
and deltoid using a hand-held dynamometer. IOrthop Sports Phys Ther
9:399-405, 1988
8. Clarke HH: Comparison of instruments
for recording muscle strength. Res Q
25:398-4 11, 1954
9. Daniels L, Worthingham C: Muscle
Testing: Techniques of Manual Examination (6th Ed), p 2. Philadelphia, PA:
W.B. Saunders Company, 1986
10. Edwards RTH, McDonnel M: Handheld dynamometer for evaluating voluntary muscle function. Lancet 2:757758, 1974
1I.Everett PW, Sills FD: The relationship
of grip strength to stature, somatotype
components, and anthropometric
measurements of the hand. Res Q
23:161-166, 1952
12. Hosking CP, Dubowitz V, Edwards
RHT: Measurements of muscle
strength and performance in children
with normal and diseased muscle.
Arch Dis Child 5 1:957-963, 1976
13. )ones HE: The relationship of strength
to physique. Am 1 Phys Anthropol
5:29-39, 1947
14. Marino M, Nicholas IA, Cleim CW,
Rosenthal P, Nicholas S): The efficacy
of manual assessment of muscle
strength using a new device. Am I
Sports Med 10:360-364, 1982
15. Milner-Brown HS, Mellenthin M,
Miller RC: Quantifying human muscle
strength, endurance, and fatigue. Arch
Phys Med Rehabil67:530-535, 1986
16. MorrowlR, Hosler WW: Strength comparisons in untrained men and trained
women athletes. Med Sci Sports Exerc
13:194-198, 1981
17. Saraniti Al, Cleim CW, Melvin M,
Nicholas IA: The relationship between
subjective and objective measurements of strength. I Crthop Sports
Phys Ther 2: 15- 19, 1980
18. Schmidt RT, Toews /V: Crip strength
as measured by the lamar dynamometer. Arch Phys Med Rehabil 5 1:32 1327, 1970
19. Smidt CL: Muscle Strength Testing: A
System Based on Mechanics, pp 19128. Iowa City, IA: Spark Instruments
and Academics, 1984
20. Sullivan Sj, Chesley A, Hebert C,
McFaull 5, Scullion D: The validity and
reliability of hand-held dynamometry
in assessing isometric external rotator
performance. I Orthop Sports Phys
Ther 1O(6):2 13-2 17, 1988
21. Viitasalo IT, Era P, Leskinen AL, Heikkinen E: Muscular strength profiles and
anthropometry in random samples of
mean 3 1-35, 5 1-55, and 71-75. Ergonomics 28: 1563- 1574, 1985
22. Wadsworth CT, Krishnan R, Sear M,
Harrold 1, Nielsen DH: lntrarater reliability of manual muscle testing and
hand-held dynametric muscle testing.
Phys Ther 67:1342-1347, 1987
23. Wiles CM, Karni Y: The measurement
of muscle strength in patients with peripheral neuromuscular disorders. I
Neurol Neurosurg Psychiatry46: 10061013, 1983