Concurrent Validity and Intertester Reliability of
Universal and Fluid-based Goniometers for Active
Elbow Range of Motion
MICHELE PETHERICK,
WENDY RHEAULT,
SHARON KIMBLE,
CONNIE LECHNER,
and VIRGINIA SENEAR
The purposes of this study were 1) to establish the concurrent validity of the
universal goniometer and the fluid-based goniometer and 2) to determine the
intertester reliability of these two instruments. A correlational study was performed in which two testers used the universal goniometer and the fluid-based
goniometer in measuring elbow range of motion in 30 healthy subjects. The fluidbased goniometer had high intertester reliability (R = .92), and the standard
goniometer had poor reliability (R = .53). The Pearson product-moment correlation
between the two instruments was .83. A significant difference was shown between the standard goniometer and the fluid-based goniometer by the t test (t =
4.4, df = 28, p < .05). The results support the use of the fluid-based goniometer
between testers on elbow range of motion; however, the two instruments cannot
be used interchangeably.
Key Words: Elbow joint; Physical therapy; Tests and measurements, range of motion;
Upper extremity, elbow.
The goniometer long has been recognized as an essential
tool for clinical physical therapists. Goniometers are used by
physical therapists to assess treatment effectiveness and dis­
ease progression and to document this information to thirdparty payers.1"7 Unfortunately, the homogeneity of this im­
portant assessment tool and the technique associated with its
use have evolved slowly. This problem was addressed by
Moore, who attributed errors to a lack of standardization in
nomenclature, numerical expression, patient position, and
instrumentation.6,8 Hellebrandt et al suggested that rigid con­
trol of these variables would result in greater reliability for
measurement of joint range of motion.3
In 1949, Moore advocated the use of an instrument that is
currently considered the universal (standard) goniometer
(UG).8 This goniometer is a protractor with an extended
M. Petherick, BS, is a physical therapist affiliated with the Manhattan
Psychiatric Center-Ward's Island, 600 E 125th St, New York, NY 10035.
W. Rheault, MA, is Chairman and Associate Professor, Department of
Physical Therapy, University of Health Sciences-The Chicago Medical School,
3333 Green Bay Rd, Bldg 51, North Chicago, IL 60064 (USA).
S. Kimble, BS, is Physical Therapist, Shalom Home Inc, 1554 Midway Pkwy,
St. Paul, MN 55108.
C. Lechner, BS, is Staff Physical Therapist, Paulsen Rehabilitation Center,
Willowbrook, IL 60439.
V. Senear, BS, is Staff Physical Therapist, Virginia Mason Hospital, Depart­
ment of Physical Medicine and Rehabilitation, 925 Seneca St, Seattle, WA
98101.
Ms. Petherick, Ms. Kimble, Ms. Lechner, and Ms. Senear were senior
students in the physical therapy program, University of Health Sciences-The
Chicago Medical School, at the time this study was conducted.
Address all correspondence to Mrs. Rheault.
This article was submitted February 13, 1987; was with the authors for
revision 17 weeks; and was accepted September 8, 1987. Potential Conflict of
Interest: 4.
966
stationary arm and a fulcrum-mounted moveable arm. Tape
measures and trigonometry, in addition to pendulum, bubble,
and "over-the-joint" goniometers, have also been used to
measure joint ROM.9-11 These other measuring tools, how­
ever, have not gained wide clinical acceptance. A fluid-based
type of goniometer (FG), which works on the principle of a
carpenter's level, was recently introduced. The fluid in this
type of instrument is contained in a circular chamber and is
displayed by straight-plane movement in gravity. Although
manufacturers of the FG have provided no literature on the
instrument, Clarke and associates have described the use of a
type of FG for the glenohumeral joint,12,13 and other authors
have used it to measure cervical ROM.10
For a measurement instrument to be useful, its reliability
and validity must be established. Goniometric measurement
has been found to have greater intratester reliability than
estimation by observation.4 High intratester reliability has
been found by several authors in both clinical and research
settings when using the UG.1,7 Boone et al have found that
for a given motion no significant intratester variation exists
in three repetitions of the same motion.1 Rothstein and asso­
ciates found that intratester reliability for elbow and knee
joint ROM measurements was very high (r = .91-.99).7 The
results of that study support the finding of Boone et al1 that
multiple measurements do not increase intratester reliability.
These findings, however, disagree with those of Low, who
contends that the average of multiple measurements leads to
greater accuracy and reliability.4
In general, past research has found intertester reliability to
be less than intratester reliability.1,9,11 Hamilton and Lachenbruch measured finger joint ROM and found significant
PHYSICAL THERAPY
RESEARCH
variance in mean measurements taken by several therapists
over a period of time.11 Rothstein et al studied intertester
reliability of elbow and knee joint ROM measurements and
found Pearson product-moment correlations of .88 to .97 for
knee and elbow flexion and elbow extension but correlations
of only .63 to .70 for knee extension.7 Fitzgerald et al found
that intertester reliability was .88, .76, and .91 for spinal
extension, right lateral spinal flexion, and left lateral spinal
flexion, respectively.14 Boone et al found less intertester vari­
ation in three upper extremity motions (r = .86) than for
lower extremity motions (r = .58).1 They found that the
average standard deviation for upper extremity motions was
2 degrees, but for elbow flexion was 3.7 degrees. They con­
cluded that a change of greater than 5 degrees in the upper
extremities and greater than 6 degrees in the lower extremities
is necessary to determine improvement in ROM when more
than one tester measures the same motion.1
High validity of the measuring instrument is required for
the tool to provide meaningful results. Burdett and associates,
in a study measuring lumbar spine movements in healthy
subjects, found low validity correlations between four meas­
urement instruments and roentgenographic measurements.15
Gogia et al, however, found high Pearson product-moment
correlation coefficients (r = .97-.98) and intraclass correla­
tion coefficients (ICCs) (ICC = .98-.99) between goniometric
measurements and roentgenograms taken of the knee.16 They
concluded that the goniometer was a valid instrument for
measuring ROM of the knee joint. Enwemeka also reported
no significant difference between goniometric and roentgen­
ographic measurements of the knee, except in the first 15
degrees of ROM.17
Because little literature is available on the FG, we felt that
a study exploring the validity and reliability of this instrument
was necessary. The purposes of this study were 1) to establish
the concurrent validity of the UG and the FG and 2) to
determine intertester reliability of these two instruments in
measuring elbow joint ROM. We hypothesized 1) that no
significant difference would exist between the UG and the FG
in measuring elbow flexion ROM and 2) that the intertester
reliability of these two instruments would be high.
METHOD
Subjects
Subjects were 10 male and 20 female volunteers with no
previous history of musculoskeletal or neurological problems
of the right upper extremity (Tab. 1). The mean age of the
subjects was 24 years (s = 4.2 years). All subjects were
informed of the testing procedure, and a signed statement of
consent was obtained from each subject.
Age
Height (m)
Weight (kg)
Right arm girth (cm)a
Right arm length (cm)b
s
24.00
1.68
63.62
26.40
74.26
Two goniometers were used for comparison of their validity
and intertester reliability, a UG and a FG.* The UG was a
full-circle plastic goniometer with 25.4-cm moveable arms,
marked in 1-degree increments.
The FG had afluid-filledchamber that responded to posi­
tion changes in space. It had a 360-degree scale divided into
1-degree increments and was read from the bottom of the
fluid meniscus. The FG weighed about 11 g, had a 0.7-cm
base, and was 10 cm in height. Each goniometer was calibrated
before each testing session to known angles of 0, 90, and 180
degrees as determined from a standard protractor.
Procedure
Subjects were placed supine in the anatomical position on
a treatment table. The subject's right arm was positioned with
a 5.08-cm thick towel under the distal humerus and with the
forearm off the edge of the plinth. This arm position allowed
full elbow extension ROM. The right shoulder was stabilized
with a 2.3-kg weight over the deltopectoral groove to prevent
shoulder protraction. The subject's right arm was exposed for
easy identification of bony landmarks.
The UG was aligned according to the procedure outlined
by Norkin and White.18 The stationary arm of the goniometer
paralleled the lateral midline of the humerus, pointing at the
acromion process. The goniometer's moveable arm was
aligned with the lateral midline of the radius and the styloid
process. The lateral epicondyle was used to center the fulcrum
of the goniometer.
The FG was placed on the dorsal aspect of the subject's
forearm, 6 cm distal to the olecranon process. The scale of
the liquid level was set at 0 degrees when the limb was at the
starting position of full extension. The scale was read when
the subject's elbow was in full flexion.
Two testers (S.K. and C.L.) each took three measurements
on every subject with both instruments. This testing procedure
was followed because the literature disagrees on whether
multiple measurements lead to greater intratester reliability.
The order of measurement was randomized to eliminate
sampling bias. The researchers practiced the measurement
procedure before the experiment to ensure testing reliability.
Subjects activelyflexedand extended their right elbow five
times before thefirstmeasurement to ensure constant ROM
throughout the testing session. The goniometer was aligned
with the arm in full extension. The subjects then actively
flexed their right elbow through the full ROM, avoiding
forceful contraction. The full arc of motion was recorded by
separate researchers (M.P. and V.S.) to blind the testers to
previous measurements and thus prevent tester bias. All data
were collected in the same laboratory at the university.
Data Analysis
TABLE 1
Physical Characteristics of Subjects (N = 30)
Characteristic
Instrumentation
4.20
.08
11.25
3.60
4.45
Two statistical methods were used to assess the concurrent
validity of the goniometers. We used a Pearson productmoment correlation to determine the covariance of measure­
ments taken with both goniometers. Because the correlation
coefficient can be high even if large differences exist between
covarying paired measurements, the paired-data t test was
also used.19 This method measured the agreement between
a
Girth measured 10 cm proximal to olecranon process.
Arm length measured from acromion process to distal end of
third digit.
b
Volume 68 / Number 6, June 1988
Chattanooga Corp, 101 Memorial Dr, PO Box 4287, Chattanooga, TN
37405.
967
testers to establish the concurrent validity of the goniometers.
To compute these statistics, we used the mean of all six
measurements per subject taken with each instrument.
Separating the data by instrument and tester, we calculated
the mean of each subject's measurements. These values were
used to compute the ICC (formula 2,3), which was used to
determine the intertester reliability of each instrument.519 The
subjects acted as their own control, thus increasing the sensitivity of the experiment.19 The use of ICCs allowed the determination of the source of variability between and within
testers. All statistical tests were two-tailed at an alpha level
of .05.
RESULTS
The Pearson product-moment correlation between the two
goniometers equaled .83, which is significant at an alpha level
of .05 (Tab. 2). The paired-data t test revealed a significant
difference between the two instruments (t = 4.4, p < .05)
(Tab. 2). Intertester reliability was determined by the ICC as
.53 for the UG and .92 for the FG (Tab. 3).
DISCUSSION
The results of this study, although conducted on a small
group of subjects, showed high correlation between the FG
and the UG. Because the instruments were measuring the
same joint angle, a high correlation would be expected. The
instruments were not shown to be concurrently valid by the t
test.
The intertester reliability was found to be higher with the
FG than with the UG. Our results do not support those of
Rothstein and associates, who found high intertester reliability
(r = .95 for elbow flexion and extension) with the UG.7
Twelve testers took measurements in their study. Because
standard deviations and degrees of freedom were not published with their study results, we could not determine how
their statistics were calculated. The subjects in their study
included patients with elbow disorders, which produced wider
variability in the data. Statistically, this variability, if large,
will produce higher correlation values. In our study, subjects
with normal elbow mobility were measured, producing a
a
b
s
SEM
Universal
145.8
6.3
1.2
Fluid-based
149.4
7.9
1.5
ra
t
.83
4.4b
CONCLUSION
This study compared the UG and the FG and found that
concurrent validity did not exist between these two instruments. It also addressed intertester reliability for the two
instruments and found that the FG had greater intertester
reliability than the UG. Although the UG appears to be a
mechanically sound measurement instrument, even the use
of a standardized testing procedure allowed excessive variability between therapists. Although more research is needed,
we foresee potential usefulness of the FG for improving
goniometric measurement in physical therapy.
REFERENCES
1. Boone DC, Azen SP, Lin C-M, et al: Reliability of goniometric measurements. Phys Ther 58:1355-1360, 1978
2. Loessin Grohmann JE: Comparison of two methods of goniometry. Phys
Ther 63:922-925, 1983
3. Hellebrandt FA, Duvall EN, Moore ML: The measurement of joint motion:
Part 3. Reliability of goniometry. Phys Ther Rev 29:302-307, 1949
4. Low JL: The reliability of joint measurement. Physiotherapy 62:277-229,
1976
5. Miller PJ: Assessment of joint motion. In Rothstein JM (ed): Clinics in
Physical Therapy: Measurement in Physical Therapy. New York, NY,
Churchill Livingstone Inc. 1985, vol 7, pp 103-136
6. Moore ML: The measurement of joint motion: Part I. Introductory review
of the literature. Phys Ther Rev 29:195-205, 1949
7. Rothstein JM, Miller PJ, Roettger RF: Goniometric reliability in a clinical
setting: Elbow and knee measurements. Phys Ther 63:1611-1615, 1983
TABLE 2
Elbow Joint Range-of-Motion Measurements Taken with
Universal and Fluid-based Goniometers
Goniometer
restricted statistical range, which leads to attenuated correlation values.
We found the FG to be more reliable than the UG, perhaps
because of its simple technique of use. The UG, in contrast
to the FG, requires careful and sustained alignment to bony
landmarks during movement. Other advantages of the FG
include its light weight, small size, and speed of application.
Its ease of application allows the therapist to obtain ROM
measurements with one hand, leaving the other hand free to
passively move the patient's extremity.
The FG, however, does have several limitations. It must be
used perpendicular to the gravitational field, which necessitates frequent patient position changes. Its size precludes use
on smaller joints such as the interphalangeal joints. Another
disadvantage is that the FG measures only total excursion as
opposed to movement from neutral. This factor may account
for the significant difference we found between the two measurement instruments with the t test. Because the FG measures
only total movement, it cannot specify the part of the ROM
where movement occurs. Lastly, the cost of the FG is about
three times that of the UG.
Future research involving the FG is necessary to fully
understand its clinical applicability. This research could be in
the areas of 1) protocol establishment for standardization, 2)
determining reliability for other joints, and 3) clinical trials.
Further study of both the FG and the UG is required to
determine their validity.
Pearson product-moment correlation.
p < .05 using the paired-data f test.
TABLE 3
Intertester Reliability of Universal and Fluid-based Goniometers
Universal Goniometer
Fluid-based Goniometer
Source
df
Between subjects
29
SS
2292.9
MS
R
79.0
df
SS
MS
29
2651.7
125.9
30
155.0
5.2
.92
.53
Within subjects
968
30
711.1
23.7
R
PHYSICAL THERAPY
RESEARCH
8. Moore ML: The measurement of joint motion: Part 2. The technic of
goniometry. Phys Ther Rev 29:256-264, 1949
9. Hsieh C-Y, Walker JM, Gillis K: Straight-leg-raising test: Comparison of
three instruments. Phys Ther 63:1429-1433, 1983
10. Buck CA, Dameron FB, Dow JM, et al: Study of normal range of motion in
the neck utilizing a bubble goniometer. Arch Phys Med Rehabil 40:390392, 1959
11. Hamilton GF, Lachenbruch PA: Reliability of goniometers in assessing
finger joint angle. Phys Ther 49:465-469, 1969
12. Clarke GR, Willis LA, Fish WW, et al: Preliminary studies in measuring
range of motion in normal and painful stiff shoulders. Rheumatol Rehabil
14:39-46,1975
13. Clarke GR: Measurement in shoulder problems. Rheumatol Rehabil
15:191-193,1976
Volume 68 /Number 6, June 1988
14. Fitzgerald GK, Wynveen KJ, Rheault W, et al: Objective assessment with
establishment of normal values for lumbar spinal range of motion. Phys
Ther 63:1776-1781, 1983
15. Burdett RG, Brown KE, Fall MP: Reliability and validity of four instruments
for measuring lumbar spine and pelvic positions. Phys Ther 66:677-684,
1986
16. Gogia PP, Braatz JH, Rose SJ, et al: Reliability and validity of goniometric
measurements at the knee. Phys Ther 67:192-195, 1987
17. Enwemeka CS: Radiographic verification of knee goniometry. Scand J
Rehabil Med 18:47-50, 1986
18. Norkin CC, White DJ: Measurement of Joint Motion: A Guide to Goniometry. Philadelphia, PA, F A Davis Co, 1985, p 38
19. Currier DP: Elements of Research in Physical Therapy, ed 2. Baltimore,
MD, Williams & Wilkins, 1984, pp 163-164,171, 230
969