CONCURRENT VALIDITY OF FOUR CLINICAL TESTS

CONCURRENT VALIDITY OF FOUR CLINICAL TESTS

USED TO MEASURE HAMSTRING FLEXIBILITY

D. SCOTT DAVIS, RICH O. QUINN, CHRIS T. WHITEMAN, JASON D. WILLIAMS, AND COREY R. YOUNG

Division of Physical Therapy, Department of Human Performance and Exercise Science, West Virginia University, Morgantown,West Virginia

ABSTRACT

The purpose of this study was to examine the concurrent validity of

4 clinical tests used to measure hamstring muscle length. A pilot

study (N = 10) was conducted to determine the intratester

reliability of 4 hamstring length measures: knee extension angle

(KEA), sacral angle (SA), straight leg raise (SLR), and sit and reach

(SR). The pilot investigation revealed good to excellent intratester

reliability (intraclass correlation coefficient = 0.92–0.95) for each

of the 4 tests. Eighty-one subjects (42 men and 39 women)

participated in the main investigation. Subjects were randomly

tested for each of 4 assessments of hamstring length. Concurrent

validity was determined using linear regression, correlation, and k

statistics. Correlation coefficients corresponding to the concurrent

validity of the six combinations of the 4 clinical tests revealed poor

to fair correlation (r = 0.45–0.65). The correlation coefficients for

each pair from greatest to least were SR-SA= 0.65, SLR-SR =

0.65, KEA-SLR = 0.63, KEA-SR = 0.57, SLR-SA = 0.50, and

KEA-SA = 0.45. Despite the common clinical use of these

measures to assess hamstring length, these tests do not have

sufficient concurrent validity to be used interchangeably or to

assume that they each measure the same construct (hamstring

length). Based on the results of this investigation and a review of

the literature, the authors recommend that researchers, clinicians,

and strength and conditioning specialists adopt the KEA test as

the gold standard measure for hamstring muscle length.

KEY WORDS assessment, examination, muscle length

INTRODUCTION

The literature describes 4 common methods ofassessing hamstring muscle length (2–25,28,29).These include the knee extension angle (KEA) test,the sacral angle (SA) test, the straight leg raise

(SLR) test, and the sit-and-reach (SR) test. The choice ofwhich test to use is often based on examiner preference, ease ofuse, professional discipline, or tradition rather than scientific

evidence. In order for a test to be clinically meaningful, the testmust possess a high degree of reliability and validity (26). Thepurpose of this investigation was to examine the concurrentvalidity of these 4 measures of hamstring muscle length.

The KEA test is performed with the patient lying supine withboth lower extremities extended. The tester flexes theipsilateral hip to 90� and maintains this angle while theipsilateral knee is passively extended. The contralateral lowerextremity is stabilized on the examination plinth. Clinically, theendpoint is reached when either the tester feels slight resistanceor the subject reports a strong but tolerable stretching sensationin the hamstring musculature. Using either a universal goni-ometer or two gravity inclinometers, the tester can measure theKEA, which is the degree of knee flexion from terminal kneeextension. Alternatively, the tester can measure the obtuseadjacent angle measured between the femur and the tibia. Thelater angle is called the popliteal angle (PA). The sum of theKEA and the PA is 180�. The KEA test and the associated PAtest have been used extensively in the literature as a measure ofhamstring muscle length (2–4,7,12). The intratester reliabilityof the KEA test has been reported to be 0.99 (11,29). A KEA of20� has been defined as a cutoff score indicating hamstringmuscle tightness (7). Therefore, a KEA .20� indicateshamstring muscle tightness.

The SLR test is performed with the patient lying supinewith both lower extremities extended. While maintaining thecontralateral lower extremity on the examination plinth, theexaminer passively lifts the ipsilateral lower extremity byflexing the hip joint. The ipsilateral knee is maintained in fullextension throughout the test. The ankle is maintained inslight plantarflexion throughout the test to avoid adverseneural tension (10). At the end of the available range ofmotion, the hip joint angle is measured with either a universalgoniometer or a gravity inclinometer. Previous research usinga gravity inclinometer reported the intratester reliability ofthe SLR test to be .0.97 (14,22). A SLR of 80� has beensuggested as a cutoff score to indicate hamstring muscletightness (19). Therefore, a SLR measurement ,80�indicates hamstring muscle tightness.

The SLR test is an attempt to indirectly measure thefunctional excursion of the hamstring muscle group; however,several confounding variables have been identified in theliterature. These variables include the possibility of neural

Address correspondence to D. Scott Davis, [email protected].

22(2)/583–588

Journal of Strength and Conditioning Research� 2008 National Strength and Conditioning Association

VOLUME 22 | NUMBER 2 | MARCH 2008 | 583

stretch, stretching of the hip joint capsule, inconsistency inpelvic position, contralateral hip flexor tightness, andlimitations of motion due to fascial connections betweenthe posterior cruel fascia and the fascia of the posterior thigh(5,6,10,12).

The SR test has many variations, but they are all derivedfrom the classic test (13,15–25). The classic SR test requiresa measuring box with a mounted ruler. The subject sits on thefloor or examination plinth with both knees extended and thefeet flat against the device. The subject is then asked to reachforward over the measuring device and hold this position for2 seconds. Referencing the plantar surface of the feet (zero),the median distance for young adults 20 to 29 years of age is5 and 8 cm for men and women, respectively (1). Severalvariations of the classic SR have been proposed and exam-ined in the literature. These include the modified SR test, theV SR test, and the modified back saver SR test (13,15–25).The intratester reliability of the classic SR test was reportedto be 0.98 (27).

Kendall and colleagues (19) advocate a modified version ofthe SR test, called the SA test, which is intended to eliminatemany of the confounding variables associated with the SR test.Kendall et al. (19) propose that the angle of the sacrum relativeto the examination plinth should be used instead of measuringthe distance that the subject is able to reach relative to his orher feet. According to Kendall et al. (19), normal hamstringlength is achieved when a sacral angle of $80� relative to thesupporting surface is achieved. They suggest that this testeliminates the influence of lumbar spine range of motion andanthropometric issues related to the length of the arms, trunk,and lower extremities. Kendall and colleagues (19) suggest thatthis test is consistent with the results of the SLR test; however,no evidence is offered to support this claim. A review of theliterature failed to identify any previously established intra-tester reliability data for the SA test.

Hamstring muscle length is a fundamental measureperformed by many different disciplines across a wide varietyof settings. Although several investigations have examinedthe individual validity of these tests, the authors were unableto identify a study that examined the concurrent validity of all4 of these methods used to measure hamstring muscle length.Therefore, the purpose of this investigation was to determinethe concurrent validity of 4 tests (KEA, SA, SLR, SR) used tomeasure hamstring muscle length.

METHODS

Experimental Approach to the Problem

In the pilot study, a test-retest research design was used toexamine the intratester reliability of 4 common measures ofhamstring muscle length (KEA, SA, SLR, and SR). The maininvestigation examined the concurrent validity of the 4 ham-string length measures using a quasi-experimental correla-tional design. The primary focus of this investigation was todetermine a gold standard clinical measure of hamstringmuscle length. Based on previously reported anatomical and

biomechanical differences among the 4 tests, it was hypoth-esized that there would be sufficient variability among thetests to result in weak concurrent validity. Before recruitingthe subjects, the investigation was approved by the Uni-versity’s Institutional Review Board for the Protection ofHuman Subjects.

Subjects

The subjects (N = 81) were 42 college-age men (age, 23.6 6

4.1 years) and 39 college-age women (age, 24.1 6 4.3 years).Initially, 117 subjects were recruited to participate in thisinvestigation; however, 36 subjects were excluded based onprevious medical conditions identified in the exclusionquestionnaire. Subjects were also excluded if they werefound to have a KEA ,10� in order to exclude those subjectswith hyperflexibility.

All participants were asked to complete a medical historyquestionnaire. Exclusion criteria consisted of previous orcurrent complaints of numbness, tingling, or burning in theirarms or legs; neck or back injury/pain within the past 2 years;circulatory or neurological complaints; or lower extremityfracture. Before testing commenced, the procedures wereexplained to each participant by one of the investigators andeach subject was required to sign a consent form.

Procedures

Pilot study. A pilot investigation was conducted before themain investigation to establish the intratester reliability of the4 hamstring length measures (KEA, SA, SLR, and SR). TheKEA test was performed by having the subject lie on his or herback with the hips and knees fully extended. Gravityinclinometers (Macklanburg-Duncan, Oklahoma City, OK)were placed at two points on the tested lower extremity. Oneinclinometer was placed on the distal thigh immediatelysuperior to the patella, and the second inclinometer wasplaced on the distal anterior tibia. The distal edge of theinclinometer was aligned with the superior aspect of themedial malleolus. The tested lower extremity was thenpassively raised by the examiner to 90� of hip flexion asrecorded by the inclinometer placed on the distal thigh. Thesubject’s knee was then passively straightened to a pointwhere the subject reported a strong but tolerable stretch intheir posterior thigh. The contralateral lower extremity wasfixed to the table in full knee extension using a nylon strapover the distal thigh. The angle of the knee (KEA) was thenmeasured using the inclinometer placed on the lower leg. Theadjacent PA (180 2 KEA) was used for all statistical analyses.

The SLR test was performed by having the subject liesupine with both hips and knees fully extended. Thecontralateral lower extremity was secured to the examina-tion plinth using a nylon strap placed over the distal thigh.A gravity inclinometer was placed on the distal leg at the levelof the medial malleolus. With the ipsilateral knee fullyextended, the tested extremity was passively raised to a pointwhere the subject experienced a strong but tolerable stretch inthe posterior thigh. The angle of the subject’s lower extremity

584 Journal of Strength and Conditioning Researchthe TM

Concurrent Validity of Four Clinical Tests

from the horizontal surface of the examination plinth wasrecorded from the inclinometer strapped to the distal tibia.

The SA test was performed by having the subject sit on theexamination plinth with the knees fully extended and theirhips in neutral rotation and fully adducted. The subject wasinstructed to reach forward with his or her arms extended,toward the toes, until a strong but tolerable stretch was felt inthe posterior thigh. The angle formed between the sacrumand the horizontal plane was measured by placing a gravityinclinometer flat against the subject’s sacrum. If the subject’ssacrum was vertical, a measurement of zero was recorded.Positive scores were recorded if the sacrum was unable toobtain a vertical position as the subject reached forward.Negative scores were recorded if the sacrum was able to flexbeyond a vertical position.

The SR test was performed by having the subject sit on theexamination plinth with the knees fully extended. The hipswere placed in neutral rotation and fully adducted. Thesubject’s feet were placed flat against a standard SR measuringdevice. The subject was instructed to keep his or her kneesstraight while reaching forward until a strong but tolerablestretch was felt in the posterior thigh. The distance that thesubject was able to reach forward was recorded in centimetersfrom a ruler located on the SR measuring device.

Four investigators each performed one of the hamstringlength measures. Each investigator practiced their respectiveexamination technique for a total of 4 hours before the start of

the study. Ten college-age participants (5 men and 5 women)were recruited to participate in the pilot study. The examinerswere blinded to the subject’s identity by a drape placed at thesubject’s waist. Each examiner randomly tested each subject3 times. Each subject was asked to avoid communication duringthe testing procedure to ensure the blinding of the subject fromthe tester. Additionally, each examiner was assigned a researchassistant who recorded the values as they were taken. Thisprevented the examiner from having contact with the databetween trials. Data from the pilot study were analyzed usinga repeated-measures analysis of variance (ANOVA) and anintraclass correlation coefficient (ICC 3,1) according to theformula described by Shrout and Fleiss (27). The pilot studyrevealed good to excellent intratester reliability (ICC = 0.92–0.95) for each of the examiners using their assigned technique.

Testing procedure. The main study was conducted ina laboratory with 4 examination stations (KEA, SA, SLR,and SA) on two separate occasions within a 1-week period. Toprevent measurement error related to repeat testing (sarco-mere give), subjects randomly rotated among the 4 stations.The right hamstring was measured for the SLR and KEAtests, while the SR and SA tests required a simultaneous testof both lower extremities. The tests were performed in thesame manner as described in the pilot investigation.

Statistical Analyses

Data from the main investigation were analyzed using JMP5.1 statistical software (SAS Institute, Cary, NC). Descriptivestatistics (mean and SD) were calculated for age, gender,and each of the 4 measures of hamstring muscle length. Asdescribed by Portney and Watkins (26), simple linear regres-sion analysis and Pearson correlation coefficients were cal-culated for each combination of the 4 testing procedures todetermine concurrent validity. Beta coefficients, probabilityvalues, and SEs were calculated for each regression analysis.Additionally, the test results were dichotomized based onpublished cutoff scores for each of the 4 tests, and k statisticswere calculated for each pair of the 4 testing procedures.Gender differences were examined using an independentt-test with a predetermined a level of 0.05.

TABLE 1. Intratester reliability.

Test ICC (3,1)

KEA 0.94SA 0.95SLR 0.92SR 0.94

KEA= knee extension angle; SA= sacral angle; SLR=straight leg raise; SR = sit and reach.

TABLE 2. Mean (SD) for age and hamstring length measures.

Variables Men (n = 42) mean (SD) Women (n = 39) mean (SD) All subjects (N = 81) mean (SD)

Age 23.6 (4.1) 24.2 (4.3) 23.7 (3.9)PA (180 2 KEA) 71.6 (9.6) 77.7 (9.5) 74.6 (10.0)SA 15.7 (5.4) 18.6 (6.6) 17.1 (6.1)SLR 75.5 (10.3) 82.8 (10.8) 79.0 (11.1)SR 23.5 (6.7) 26.8 (7.6) 25.1 (7.3)

PA = popliteal angle; KEA = knee extension angle; SA = sacral angle; SLR = straight leg raise; SR = sit and reach.


Journal of Strength and Conditioning Researchthe TM

| www.nsca-jscr.org

RESULTS

The pilot study yielded excellent intratester reliability withICC (3,1) values ranging from 0.92 to 0.95 for all 4 techniques(Table 1). Table 2 contains the mean and SD for age and eachof the hamstring length measures. Independent t-tests (Table 3)revealed a significant difference in hamstring length betweenmen and women for each of the 4 hamstring length measures.Women were found to have 8.5, 9.7, 14.0, and 18.5% moreflexibility than their male counterparts for KEA, SLR, SR,and SA tests, respectively.

Although each of the 4 tests has been reported to bea measure hamstring muscle length, the measurements areunique and agreement among the raw scores is impossible.In order to determine the concurrent validity of the 4 ham-string length measures, simple linear regression and Pearson’scorrelation coefficient (Table 4) were calculated for each com-bination of the 4 tests. This analysis identified poor to faircorrelation among all pairs of the 4 tests.

In addition to examining the reliability of the 4 tests usingraw values, the data were dichotomized to indicate a positiveor negative test for each subject based on normative cutoffvalues reported for each testing procedure. After reducing thedata to a nominal variable (positive = tight hamstrings;

negative = normal hamstring length), k correlations (Table 5)were calculated to determine the agreement among the 4tests. Kappa values were unattainable for each of the pairingsof SA because all the subjects were found to be negative usingthe SA test. Kappa agreement among the 4 tests was found tobe fair.

DISCUSSION

Previous investigations have identified good to excellentintratester reliability of these commonly used measures ofhamstring muscle length; however, to date, no investigationhas attempted to identify the concurrent validity of all 4 tests.The results of this investigation support the excellentintratester reliability of these tests, but the results fail tosupport adequate concurrent validity among these clinicalmeasures using either the raw scores or the dichotomizedscores based on previously published cutoff values. Thecoefficient of determination (R2), which is simply the squareof the Pearson product moment correlation coefficient (r),represents the amount of variability in y that is explained by x.The largest correlation coefficient in this investigation was(r = 0.65) for the SR-SA pairing, which represents a R2 of0.43. Therefore, only 43% of the variability in the SA test wasexplained by the SR test.

The lack of concurrent validity among these 4 tests raisesquestions related the appropriateness of some of these tests asa valid measure of hamstring muscle length. This investiga-tion also supports previous concerns related to the anatomicalstructures, which may confound some of these tests. Variablesthat may have contributed to the poor correlation amongthese tests include adverse neural tension, differences in thetesting positions (supine vs. seated), bilateral vs. unilaterallower extremity testing, differences associated with pelvicposition and stability, variability in anthropometric variablessuch as body segment length, and the contribution of othernoncontractile tissues such as the deep and superficial fascia.

Neural tension associated with some of these tests maypose a threat to obtaining accurate range of motionmeasurements because muscle length testing often requiresa subjective response of a strong but tolerable stretch from thesubject. Neural tension has been reported to be a confounding

TABLE 3. Comparison of hamstring length amongmen and women.

Variable Prob . | t |

SLR 0.0027*KEA 0.0056*SR 0.0421*SA 0.0281*

*P , 0.05.SLR = straight leg raise; KEA = knee extension angle;

SR = sit and reach; SA = sacral angle.

TABLE 4. Regression analysis among the 4 tests.

Tests b r R2 SE Prob . | t |

SLR and KEA 0.57 0.63 0.40 0.08 ,0.0001SLR and SR 0.42 0.65 0.41 0.06 ,0.0001SLR and SA 0.28 0.50 0.25 0.05 ,0.0001KEA and SR 0.41 0.57 0.33 0.07 ,0.0001KEA and SA 0.27 0.45 0.20 0.06 ,0.0001SR and SA 0.55 0.65 0.43 0.07 ,0.0001

SLR = straight leg raise; KEA = knee extension angle;SR = sit and reach; SA = sacral angle.

TABLE 5. Kappa correlations.

KEA SR SA

SLR 0.36 0.39 Unable to evaluateKEA 0.42 Unable to evaluateSR Unable to evaluate

KEA = knee extension angle; SR = sit and reach; SA =sacral angle; SLR = straight leg raise.



factor that is exacerbated by ankle dorsiflexion (10,21).Liemohn et al. (21) found that ankle dorsiflexion comparedto ankle plantarflexion significantly limited the availablerange of motion when performing a 1-leg SR test. Theyreported a 3-cm improvement when the ankle was allowed toplantarflex. They speculated that fascial attachments be-tween the gastrocnemius and the hamstrings and neuraltension contributed to the difference associated with ankleposition. Three of the 4 tests (KEA, SLR, and SA) in thisinvestigation were performed with the ankle(s) comfortablyplantarflexed, while the SR test was the only test in which theankles were dorsiflexed.

Previous research has shown that there is significantly lesspelvic rotation when performing the KEA test than whenperforming the SLR test (5). Bohannon et al. (5) reporteda mean pelvic rotation of 32.1� associated with the SLR test.Additionally, they reported that for every 1.7� degrees of hipmotion, there is 1� of pelvic motion when performing theSLR test. Fredriksen et al. (9) examined the pelvic con-tribution associated with a modified KEA test with the hipinitially in 120� of hip flexion. They reported the medianpelvic contribution to be only 4.1�.

Concern has also been raised about the influence of con-tralateral hip flexor length when performing the SLR test(5,9). Gadjosik et al. (10) found similar results when test-ing the SLR with the contralateral leg stabilized on theexamination plinth and when the contralateral hip wasslightly flexed to allow the lumbar spine to rest flat on theexamination plinth. While Cameron et al. (6) found thatplacing the contralateral knee in 90� flexion significantlyincreased the SLR test compared to when the contralateralleg is maintained on the examination plinth. This demon-strates the potential affects associated with contralateral hipand pelvis position.

The authors anticipated poor concurrent validity betweenthe supine and seated tests since the seated SA and SR testsmeasure bilateral hamstring length rather than unilateralhamstring length measured by the KEA and SLR tests.Another potential confounding variable is related to anthro-pometric variability among the subjects. As discussed byKendall et al. (19), trunk and extremity length variability hasthe potential to significantly affect the SR test.

Finally, it is important to note that the SA test was unable toclassify any subject in this sample of young healthy adults astight, using the guidelines offered by Kendall et al. Given thatthe other three tests (KEA, SLR, and SR) all identifiedsubjects in this sample as positive (tight hamstring) raisesconcerns about the cutoff score offered by Kendall et al. (19)for the SA test in young healthy adults.

Despite the fact that the authors were unable to find a studythat examined all 4 tests concurrently in the same sample,several investigations have examined the concurrent validityof individual pairs. Liemohn and colleagues (22) examinedthe concurrent validity of the SR test compared to thecriterion SLR test. Using a relatively small sample (N = 40),

Liemohn et al. reported a fair correlation (r = 0.71) betweenthe SR test and the SLR test. The correlation between the SRtest and SLR test in this investigation was found to be 0.65,which is only slightly lower than the results reported byLiemohn et al. (22).

Gajdosik et al. (12) examined the concurrent validity of theSLR test and KEA test and reported a correlation of 20.66,which is very consistent with the 0.63 correlation found inthis investigation. The sign difference is related to the factthat this investigation used PA = 180 2 KEA for all statisticalanalysis. Thus, as SLR increased, the popliteal angle alsotended to increase.

Minkler and Patterson (24) reported the concurrent validityof the SR test with the SLR test to be 0.75. Additionally,Simoneau (28) examined the concurrent validity of the SR testand SLR test (N = 34) and reported a correlation of 0.78.Although Simoneau (28) acknowledged the potential contri-bution of the spine to the SR test, he indicated that hamstringflexibility was the largest single contributor of motion in theSR test. Despite the unquestionable influence of the ham-strings on the SR test, Simoneau suggested muscle-specificlength tests (KEA and SLR) were more appropriate thana single flexibility test that measures the contribution ofmultiple body segments. The SR to SLR correlation in thisinvestigation was 0.65, slightly less than that reported by bothMinkler and Patterson (24) and by Simoneau (28).

This investigation has several potential limitations. Theforemost limitation was the use of multiple testers. In order tomaximize the sample size and to improve the logistics of theinvestigation, 4 testers were used. To minimize potentialdifferences based on tester, each tester practiced and gainedproficiency in 1 of the 4 tests. While the intratester reliabilityof the testers was established, it is unclear whether the sametester performing each test would demonstrate a strongercorrelation among the 4 tests. To promote consistency amongthe 4 tests, the degree of stretch was determined by the subjectrather than the testers. Future investigations should con-sider using a potentiometer normalized to body weight as anattempt to consistently apply the same stretching force.Although the authors attempted to control for sarcomere giveby randomly assigning the testing order of the 4 tests, futureinvestigation should consider a short bout of stretching priorto the testing session to help eliminate the short-term gainsassociated with repeated testing. This investigation useda relatively large sample size (N = 81); however, the subjectswere homogeneous based on age, which limits the externalvalidity of the results.

The 4 tests were not compared to a specific criterion test;instead, the concurrent validity of each pair was examined.Despite not identifying a specific criterion test for thisinvestigation, the investigators suggest that the KEA test withthe ankle plantarflexed be adopted as the gold standard basedon the body of knowledge currently available. Further study iswarranted to determine the influence of the fascia and othernoncontractile tissues on these tests. Based on the results of


Journal of Strength and Conditioning Researchthe TM

| www.nsca-jscr.org

this investigation, there is poor to fair concurrent validityamong these 4 common measures of hamstring muscle lengthand ,50% of the variability of any test is explained by anyother test.

PRACTICAL APPLICATIONS

Several reliable tests are available for clinicians to use to assesshamstring muscle length; however, the validity of these testsas a measure of hamstring muscle length has been questioned.As such, a gold standard test has not been clearly establishedand adopted by clinicians, coaches, and strength andconditioning specialists. This investigation found poor to fairconcurrent validity among the 4 most common techniques(KEA, SA, SLR, and SR) used to measure hamstring musclelength. The results of this investigation also found pooragreement among these tests when dichotomized to positiveand negative tests based on previously published cutoff values.

This investigation supports previously published concernsregarding the potential for multiple confounding variableswhen using the SA, SLR, and SA tests. Based on the results ofthis investigation, these tests should not be used interchange-ably when assessing hamstring muscle length. Based on thebody of knowledge currently available, the authors suggestthat the KEA test be adopted by all strength and conditioningspecialists as the gold standard hamstring length measureuntil further research can identify a more valid measure ofhamstring length.

REFERENCES

1. American College of Sports Medicine. ACSM’s Guidelines for ExerciseTesting and Prescription (6th ed.). Baltimore: Lippincott Williams &Wilkins, 2000.

2. Bandy, WD and Irion, JM. The effect of time on static stretch on theflexibility of the hamstring muscles. Phys Ther 74: 845–850, 1994.

3. Bandy, WD, Irion, JM, and Briggler, M. The effect of time andfrequency of static stretching on flexibility of the hamstring muscles.Phys Ther 77: 1090–1096, 1997.

4. Bandy, WD, Irion, JM, and Briggler, M. The effect of static stretchand dynamic range of motion training on the flexibility of thehamstring muscles. J Orthop Sports Phys Ther 27: 295–300, 1998.

5. Bohannon, R, Gajdosik, R, and Leveau, BF. Contribution of pelvicand lower limb motion to increases in the angle of passive straightleg raising. Phys Ther 64: 474–476, 1985.

6. Cameron, DM, Bohannon, RW, and Owen, SV. Influence of hipposition on measurements of the straight leg raise test. J OrthopSports Phys Ther 19: 168–172, 1994.

7. Davis, DS, Ashby, P, McCale, K, McQuain, J, and Wine, J. Theeffectiveness of three stretching techniques on hamstring flexibility.J Strength Cond Res 19: 27–32, 2005.

8. Dixon, JK and Keating, JL. Variability in straight leg raisemeasurements. Physiotherapy 86: 361–370, 2000.

9. Fredriksen, H, Dagfinrud, H, Jacobsen, V, and Maehlum, S.Passive knee extension test to measure hamstring muscle tightness.Scand J Med Sci Sports 7: 279–282, 1997.

10. Gajdosik, RL, Leveau, BF, and Bohannon, RW. Effects of ankledorsiflexion on active and passive unilateral straight leg raising. PhysTher 65: 1478–1482, 1985.

11. Gajdosik, RL and Lusin, GF. Hamstring muscle tightness: Reliabilityof an active knee-extension test. Phys Ther 63: 1085–1088. 1983.

12. Gajdosik, RL, Rieck, MA, Sullivan, DK, and Wightman, SE.Comparison of four clinical tests for assessing hamstring musclelength. J Orthop Sports Phys Ther 18: 614–618, 1993.

13. Hopkins, DR and Hoeger, WWK. A comparison of the sit-and reachtest and the modified sit-and-reach test in the measurement offlexibility for males. J Appl Sport Sci Res 6: 7–10, 1992.

14. Hsieh, CY, Walker, JM, and Gillis, K. Straight-leg-raising test:comparison of three instruments. Phys Ther 63:1429–1433, 1983.

15. Hui, SS and Yuen, PY. Validity of the modified back-saver sit-and-reach test: a comparison with other protocols. Med Sci Sports Exerc32: 1655–1659, 2000.

16. Hui, SC, Yuen, PY, Morrow, JR, and Jackson, AW. Comparison ofthe criterion-related validity of sit-and-reach tests with and withoutlimb length adjustment in Asian adults. Res Q Exerc Sport 70: 401–406, 1999.

17. Jackson, AW, Morrow, JR, Brill, PA, Kohl, HW, Gordon, NF, andBlair, SN. Relations of sit-up and sit-and-reach tests to low back painin adults. J Orthop Sports Phys Ther 27: 22–26, 1998.

18. Jones, CJ, Rikli, RE, Max, J, and Noffal, G. The reliability and validityof a chair sit-and-reach test as a measure of hamstring flexibility inolder adults. Res Q Exerc Sport 69: 338–343, 1998.

19. Kendall, FP, McCreary, EK, and Provance, PG. Muscle Testing andFunction (4th ed.). Philadelphia: Lippincott Williams & Wilkins,1993.

20. Lemmink, KA, Kemper, HC, Degreef, MH, Rispens, P, andStevens, M. The validity of the sit-and-reach test and the modifiedsit-and-reach test in middle-aged to older men and women. Res QExerc Sport 74: 31–336, 2003.

21. Liemohn, W, Martin, SB, and Pariser, GL. The effect of ankleposture on sit-and-reach test performance. J Strength Cond Res 11:239–241, 1997.

22. Liemohn, W, Sharpe, GL, and Wasserman, JF. Criterion relatedvalidity of the sit-and-reach test. J Strength Cond Res 8: 91–94, 1994.

23. Martin, SB, Jackson, AW, Morrow, JR, and Wiemohn, WP. Therationale for the sit and reach revisited. Meas Phys Ed Exerc Sci 2:85–92, 1998.

24. Minkler, S and Patterson, P. The validity of the modifiedsit-and-reach test in college-age students. Res Q Exerc Sport 65:189–192, 1994.

25. Patterson, P, Wiksten, DL, Ray, L, Flanders, C, and Sanphy, D.The validity and reliability of the back saver sit-and-reach test inmiddle school girls and boys. Res Q Exerc Sport 67: 448–451, 1996.

26. Portney, LG and Watkins, MP. Foundations of Clinical Research:Applications to Practice (2nd ed.) Upper Saddle River, NJ: PrenticeHall, 2000.

27. Shrout, PE and Fleiss, JL. Intraclass correlation: uses in assessingrater reliability. Psychol Bull 86: 420–428, 1979.

28. Simoneau, GG. The impact of various anthropometric andflexibility measurements on the sit-and-reach test. J Strength CondRes 12: 232–237, 1998.

29. Sullivan, MK, Dejulia, JJ, and Worrell TW. Effect of pelvicposition and stretching method on hamstring muscle flexibility.Med Sci Sports Exerc 24: 1383–1389, 1992.



Documents

CONCURRENT VALIDITY OF FOUR CLINICAL TESTS