MEASURING FITNESS IN FEMALE GYMNASTS: THE GYMNASTICS FUNCTIONAL MEASUREMENT TOOL

Abstract

Purpose/Background:

A reliable and valid method of measuring and monitoring a gymnast's total physical fitness level is needed to assist female gymnasts in achieving healthy, injury-free participation in the sport. The Gymnastics Functional Measurement Tool (GFMT) was previously designed as a field-test to assess physical fitness in female competitive gymnasts. The purpose of this study was to further develop the GFMT by establishing a scoring system for individual test items and to initiate the process of establishing the test-retest reliability and construct validity of the GFMT.

Methods:

A total of 105 competitive female gymnasts ages 6-18 underwent testing using the GFMT. Fifty of these subjects underwent re-testing one week later in order to assess test-retest reliability. Construct validity was assessed using a simple regression analysis between total GFMT scores and the gymnasts' competition level to calculate the coefficient of determination (r²). Test-retest reliability was analyzed using Model 1 Intraclass correlation coefficients (ICC). Statistical significance was set at the p<0.05 level.

Results:

The relationship between total GFMT scores and subjects' current USAG competitive level was found to be good (r² = 0.60). Reliability testing of the GFMT total score showed good test-retest reliability over a one week period (ICC=0.97). Test-retest reliability of the individual component items was good (ICC = 0.80-0.92).

Conclusions:

The results of this study provide initial support for the construct validity and test-retest reliability of the GFMT.

INTRODUCTION

Women's competitive gymnastics is a multifaceted sport that requires a high level of physical fitness and skill to succeed. Speed,^1–4 strength,^2,3,5,6 endurance,⁵ agility,⁷ flexibility,^3,8–12 balance,^2,13 and power^8,14–16 are all physical abilities that play a role in the success of a competitive gymnast. A gymnast's physical abilities may also be related to the ability to sustain injury free participation in the sport.^7,17–19 As such, it is imperative that the coaches, trainers, and therapists involved in the sport be able to monitor an individual gymnast's physical abilities and overall fitness level as a means of promoting healthy, injury-free participation in the sport.

Traditionally, field-testing has been done in a variety of sports in an effort to measure sport-specific physical abilities.^20–28 For example, speed, power and agility are physical abilities needed for success in the sport of soccer. Field-tests have been developed in an attempt to quantify each of those physical abilities.^26,29,30 Some field-tests, such as the hop test³¹ or the agility T-test,³² focus on a specific aspect of sport function. Other tests, such as the Functional Movement Screen™ (FMS™),^33,34 include a battery of individual items designed to assess an athlete's abilities across multiple aspects of function.

Within the United States Association of Gymnastics (USAG), a system of competitive levels ranging from a low of 4 to a high of 10 is used to rank the skills and abilities of individual gymnasts. To move from one competitive level to the next, a gymnast must achieve a specific all-around score and be able to perform specific skills that increase in difficulty as the competitive level increases. Individual tests for flexibility, strength, endurance, and power have been suggested as useful tools to gauge gymnastic potential.^35–38 These physical abilities are included in the USAG Talent Opportunity Programs (TOPs) Test, a multi-test battery designed to measure a gymnast's basic skill in addition to the physical abilities of strength, endurance, power, and flexibility.³⁹ Although the TOPs protocol has changed a number of times since its development by William Sands,³⁷ it is used primarily with young club gymnasts ages 7-10 years of age to identify competitive potential and aid in the development of the United States competitive gymnastics program. The TOPs was thus not designed to address the needs of gymnasts of all ages or those who compete through high school or collegiate programs. While specialized training is needed to administer the TOPs and the number of people deemed qualified to administer the test is limited, the reliability and validity of the TOPs test have not been reported.

Currently there is not a reliable and valid measurement tool to evaluate the specific physical fitness abilities needed for successful competition in either men's or women's gymnastics. Previous studies have examined possible correlations between a gymnast's level of competition or intensity of training and various singular physical fitness traits.^3,12,40 Nelson and co-workers³ investigated the relationship between gymnasts' flexibility and strength and varying training intensity levels. The gymnasts at the highest level of training were reported to be the most flexible, had a slender body type, weighed less, and demonstrated higher amounts of both functional and absolute strength especially in the upper body. In 1989, Faria et al⁴¹ examined the relationship between anthropometric and physical characteristics of male gymnasts and overall competitive performance success. These researchers concluded that the top gymnasts were stronger in both absolute upper body strength and upper body strength relative to bodyweight, possessed greater overall flexibility through the hip region, shoulder girdle, and back, and possessed the least percentage of body fat.⁴¹ Neither of these studies used a standardized measurement tool to determine an overall fitness score or explore the relationship between age or body weight and physical abilities.

Without a reliable and valid field-test for measuring gymnasts' physical abilities, fitness evaluation and training are often left to the tradition-driven ways of individual coaches. As stated by Sands,¹⁹ “…. Gymnasts often simply ‘trick’ themselves into shape meaning they perform skills over and over until they acquire the fitness and skill to perform the movement”.^(p.367) This may lead to an athlete who is simply fit to do certain skills but who does not have the overall fitness level necessary for prolonged participation in the sport. With the consistent increases in the complexity and difficulty of the gymnastics elements being performed during competition,⁷ a reliable and valid method of measuring and monitoring gymnast's total physical fitness levels is needed to collectively measure the physical abilities of gymnasts and monitor their physical state.

Establishing the reliability and validity of a measurement tool is a multi-step and complex process that must be investigated within the context of the tool's intended use. Various types of validity must be considered when evaluating a new measurement tool. Construct validity, or the ability of a tool to measure the abstract concept it is intended to evaluate, is one type of validity that must be assessed. Methods of construct validation include convergence and discrimination, factor analysis, the known groups method, criterion validation, and hypothesis testing.⁴² Methods related to hypothesis testing are based on the ability of a measurement tool to reflect specific assumptions that form the framework underlying the theoretical basis of the construct. Given that a single study cannot definitively verify a theory, construct validation is considered to be an on-going process.

Various forms of reliability such as intra-rater reliability, inter-reliability and test-test reliability must also be considered when evaluating a measurement tool. Test-retest reliability is used to establish that a tool will obtain the same results across repeated administrations of the same test. Intervals between test administrations must be long enough to avoid the impact of factors such as subject fatigue and learning effects but close enough to avoid true changes in the measured variable.

Overview of the GFMT

The Gymnastics Functional Measurement Tool (GFMT) was developed to assess a gymnast's overall fitness level while minimizing the impact of gymnastic skill on testing scores.^43,44 Identifying fitness deficits to be targeted for improvement as part of a gymnast's individual training regime may prove useful in injury prevention. As a field-test for female competitive gymnasts of all ages, the GFMT was designed to be carried out by coaches, trainers and therapists using equipment commonly found in any gymnastics gym (club, high school, collegiate, etc.).

Given that successful participation in women's competitive gymnastics requires a combination of abilities related to flexibility, speed, power, strength, muscular endurance, and balance,^1–16 the individual items of the GFMT were developed based on knowledge of these requirements, a review of the literature, and consultation with experts in the field of women's gymnastics.^43,44 The 10 items comprised in the GFMT are summarized in Table 1 and detailed in Appendix I.

Table 1.

Individual Items Comprising the GFMT.

Item	Targeted Area(s) of Fitness Assessment	Units of Measure for Raw Score
The Rope Climb Test	Strength and endurance, as well as trunk control	Seconds*
The Jump Test	Lower extremity power	cm
The Hanging Pikes Test	Abdominal strength, hip flexor strength, and flexibility as well as grip strength	Number of reps
The Shoulder Flexibility Test	Shoulder complex flexion flexibility	cm/arm length
The Agility Test	Speed, endurance, and agility	Seconds
The Over-grip Pull-up Test	Upper extremity strength and muscular endurance	Number of reps
The Splits Test	Pelvis and lower extremity flexibility	Sum of cm split clearance/leg length
The Push-up Test	Shoulder and upper extremity strength	Number of reps
The 20-Yard Sprint Test	Speed and power	Seconds
The Handstand Test	Upper extremity strength and endurance as well as balance in a head-down position	Seconds

^*Signifies a form component within the final item score

The purpose of this study was to continue developing the GFMT by establishing a scoring system for individual test items and initiating the process of establishing test-retest reliability and construct validity. Given the authors' belief that a gymnast's total GFMT score would vary based on her current USAG competitive level, construct validity was assessed using the known-groups method of construct validation.

METHODS

Approval for the study was obtained from the Office for the Protection of Research Subjects at Northwestern University. Healthy competitive female gymnasts were recruited from gymnastics clubs throughout the Midwestern and Mid-Atlantic United States. Inclusion criteria required the subjects to be female, between 6 and 18 years of age, and competitive in gymnastics at USAG levels 4 to 10. Exclusion criteria included musculoskeletal pathology currently limiting the gymnast's ability to train or compete; a history of, or current systemic illnesses including cardiovascular or pulmonary disease; musculoskeletal disease or rheumatoid arthritis; and a lack of informed assent given by the participant or consent given by the parent/legal guardian. A total of 105 subjects participated in the study. Refer to Figure 1 for a flowchart reflecting subjects' participation in the study.

Figure 1.

Flowchart reflecting subject's participation in study.

All testing was performed in the subjects' home gyms or in a gym familiar to the subject. Subjects did not have prior knowledge or exposure to the specific items composing the GFMT. Each subject provided her own USAG competition level which was recorded by the testers. Prior to testing, subjects completed their regular, coach-directed warm-up routines without regard to the requirements of the GFMT. Given that field-tests composed of multiple items are often administered in stations each consisting of an individual item,^43,44 subjects were placed into groups of 10 to 12 and moved through each of the 10 stations to complete the GFMT. Data was collected by second year Doctor of Physical Therapy students from Northwestern University and by gymnastics coaches with a minimum of 5 years of coaching experience. In an effort to simulate actual practice patterns,^45–47 all data collectors were provided with a detailed set of instructions for administering each item on the GFMT but did not undergo any specialized or extensive training. Raw data for each item of the GFMT was recorded in units of measurement that were appropriate for the item tested. Units of measurement for the raw data of each item are listed in Table 1. Subjects were not intentionally masked as to their item scores. Individual GFMT items were completed in the following order to help reduce the effects of regional fatigue: Rope Climb Test, Jump Test, Hanging Pikes Test, Shoulder Flexibility Test, Agility Test, Over-grip Pull-up Test, Splits Test, Push-up Test, 20-yard Sprint Test and Handstand Hold Test. Subjects were given a minimum of 5 to 10 minutes rest between administrations of each item of the GFMT.

From the 105 total subjects, a convenience sample of 50 subjects was chosen to participate in test-retest reliability testing. These 50 subjects were retested with the GFMT one week after initial testing. Test conditions and administration were consistent between the 2 administrations of the GFMT including warm-up and item order. To help ensure that test-retest reliability rather than intra-rater reliability was assessed, testers administered different items from the GFMT on each of the 2 administration dates.

STATISTICAL METHODS

Development of the Scoring System

To develop the scoring system for the GFMT, raw scores in the appropriate units of measurement were recorded for each of the individual items on the GFMT. The raw scores for each item were used to calculate the range, mean, and standard deviation for each individual item of the GFMT (n=105). Data was then transformed to an ordinal scale using the following procedure. In an attempt to reduce the possibility of ceiling and floor effects, 5 percent of the total range of the raw scores was added to the high score of each item and 5 percent was subtracted from the low score of each item. The resulting range of scores for each individual item was then divided by 11 to create a 0 to 10 ordinal scale for each individual item on the GFMT.^48–50 The ordinal scale for each item was used to create a total GFMT score out of a possible 100 points (10 points for each item). Based on these findings, the scoring for each individual item and for the total GFMT score were finalized and are provided in the GFMT Score Sheet found in Appendix II.

Test-retest and Construct Validity:

Test-retest reliability was analyzed using Model 1 Intraclass correlation coefficients (ICC).³⁴ Although a process of systematic randomization was not employed in the study, a Model 1 ICC was used to reflect the concept that individual items on the GFMT were administered by different testers on each of the 2 test dates.⁴² The variance assessed was thus restricted to differences in the subjects' scores in the test-retest design and necessitated the use of a Model 1 ICC.⁴²

Given that previous studies had reported a positive relationship between various singular fitness traits and a gymnast's level of competition,^2,51,52 it was theorized that the total scores on the GFMT would vary with a gymnast's current competitive level. This was based upon the concept that at each increasing competitive level, a gymnast is required to perform increasingly difficult skills that require a related increase in the gymnast's physical abilities. Construct validity was thus evaluated based on the authors' belief that there would be a direct linear relationship between a gymnast's physical abilities as measured by the GFMT and the gymnast's current level of competition as reflected by the gymnast's USAG level. A simple regression analysis was performed using USAG competitive level to predict total GFMT score.⁴² The coefficient of determination (r²) was used to explore this relationship.⁴² Statistical significance was set at the p<0.05 level.

RESULTS

Of the 148 subjects assessed for eligibility in this study, 105 subjects participated. Forty-three of the recruited subjects were excluded from the study due to recent injury (n=38) or the lack of a signed informed consent or assent (n=5). The mean age of participating subjects was 12.64 years with these subjects reporting participation in competitive gymnastics for a mean of 5.4 years. Mean height and weight of the subjects were 42.76 kg and 149 cm respectively. Subject demographics, categorized by USAG competition level, are summarized in Table 2. Mean GFMT component test raw scores and standard deviations are presented in Table 3.

Table 2.

Subject Demographics by Competitive Level.

Competition Level	4	5	6	7	8	9	10
Number of Subjects	12	9	16	21	11	19	17
Mean Age in Years (sd)	10.4 (2.3)	9.9 (1.5)	10.7 (1.1)	12.0 (1.6)	13.7 (1.8)	14.6 (1.3)	15.2 (1.8)
Mean Years Competing	2.6	1.8	4.3	4.1	6.4	7.6	8.2
Mean Height in cm (sd)	140 (11.0)	137 (8.4)	139 (6.9)	146 (10.9)	154 (3.4)	157 (5.7)	154 (6.2)
Mean Weight in kg (sd)	33.3 (9.1)	32.3 (4.7)	33.8 (3.9)	39.8 (8.2)	46.2 (9.8)	49.0 (5.8)	50.7 (10.7)

sd= Standard Deviation, cm = Centimeters, kg = Kilograms

Table 3.

Mean and Standard Deviation of GFMT Individual Item Scores and GFMT Total Scores (n = 105).

	Rope Climb Time (sees)	Vertical Jump Height (cm)	Hanging Pikes (reps)	Shoulder Flexibility (cm/arm length)	Agility (sees)	Pull-ups (reps)	Splits (combined [cm/leg length])	Push-ups (reps)	20 Yard Sprint (sees)	Handstand (sees)	Total GFMT Score (/100)
	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)
Total sample (Levels 4-10)	12.40	41.70	17.62	0.773	19.12	8.08	−0.101	24.52	3.29	23.48	50.38
	(6.9)	(7.2)	(10.6)	(0.208)	(1.3)	(4.0)	(0.382)	(9.2)	(0.3)	(29.5)	(15.9)
Level 4 scores	16.64	33.33	11.08	0.67	20.73	5.25	−0.38	18.42	3.74	3.67	30.83
	(6.8)	(4.9)	(6.9)	(0.15)	(0.9)	(3.4)	(0.27)	(5.6)	(0.3)	(5.6)	(8.5)
Level 5 scores	16.91	31.50	7.67	0.63	20.89	5.11	−0.15	13.22	3.55	3.69	29.78
	(3.8)	(5.7)	(9.5)	(0.13)	(1.6)	(4.9)	(0.48)	(7.0)	(0.3)	(2.6)	(12.0)
Level 6 scores	16.35	40.69	17.69	0.80	19.29	8.63	−0.10	26.38	3.37	11.72	48.75
	(5.3)	(6.6)	(8.3)	(0.18)	(0.7)	(3.7)	(0.35)	(7.6)	(0.2)	(14.6)	(8.3)
Level 7 scores	15.35	42.74	13.14	0.85	19.23	7.52	−0.15	22.95	3.28	11.31	46.71
	(10.0)	(5.2)	(9.4)	(0.23)	(0.9)	(4.1)	(0.45)	(9.4)	(0.2)	(19.2)	(10.2)
Level 8 scores	12.15	44.45	15.73	0.80	18.66	7.00	−0.05	22.45	3.21	15.38	50.91
	(3.2)	(5.1)	(8.6)	(0.28)	(1.2)	(2.8)	(0.27)	(8.0)	(0.2)	(17.8)	(10.5)
Level 9 scores	11.97	44.42	24.11	0.70	18.24	8.26	−0.02	27.42	3.09	42.38	59.00
	(3.0)	(3.2)	(8.0)	(0.12)	(0.8)	(2.4)	(0.38)	(6.9)	(0.2)	(31.7)	(7.9)
Level 10 scores	11.10	47.82	26.94	0.87	18.04	12.29	0.06	33.12	3.01	58.18	71.18
Level 10 scores	(1.7)	(6.2)	(10.0)	(0.23)	(0.7)	(2.9)	(0.33)	(6.5)	(0.2)	(32.3)	(9.4)

Sd= Standard Deviation, cm = Centimeters, kg = Kilograms, sees = Seconds, deg = Degrees, reps = Repetitions.

Raw scores for all items on the GFMT demonstrated a normal distribution with the exception of the Handstand Test, which presented with a right skew. This skew possibly reflects the complexity of this particular activity. The relationships between the subjects' current USAG competitive level and individual component raw scores are presented in Table 4. As indicated in Table 4, several of these relationships were statistically significant, however, r² values demonstrated moderate to poor relationships between USAG competitive level and individual component raw scores (r² = 0.05-0.47). The relationship between total GFMT scores (out of a possible score of 100) and the subjects' current USAG competitive level was found to be good (r² = 0.62). Figure 2 demonstrates the relationship between USAG competitive level and total GFMT Scores. To rule out alternative explanations for the relationship between USAG competitive level and total GFMT scores, the relationships between total GFMT scores and age and total GFMT scores and bodyweight were also explored. Statistically significant relationships were identified between total GFMT score and age and between total GFMT score and bodyweight (r² = 0.13). However, r² values demonstrated a poor relationship between total GFMT score and age (r² = 0.29) and between total GFMT score and bodyweight (r² = 0.13).

Table 4.

Relationship between GFMT Individual Test Raw Score and Composite Score and the Subjects' Current Competitive Level, Body Weight and Age (n = 105).

	Rope Climb Component Score	Vertical Jump Height (cm)	Hanging Pikes (reps)	Shoulder Flexibility (cm/arm length)	Agility (sees)	Pull-ups (reps)	Splits (combined [cm/leg length])	Push-ups (reps)	20 Yard Sprint (sees)	Handstand (sees)	Total Score (/100)
Test score vs. Competitive Level	0.27	*0.42	*0.27	*0.05	*0.47	*0.20	*0.09	*0.25	*0.47	*0.38	*0.62
Test score vs. Age	*0.14	*0.28	*0.06	0.00	*0.38	*0.04	0.01	*0.10	*0.42	*0.18	*0.29
Test score vs. Body weight (kg)	*0.10	*0.26	0.01	0.00	*0.31	0.00	0.00	0.03	*0.32	0.02	*0.13

Values expressed in terms of Coefficient of Determination (r²)

^*Denotes statistical Significance (p<.05)

Figure 2.

Relationship between USAG score and GFMT score.

Raw item scores were used to examine the test-retest reliability for each item on the GFMT. Test-retest reliability of total GFMT scores was also determined. Reliability testing of the GFMT total score showed good test-retest reliability over a one week period (ICC=0.97). Test-retest reliability of the individual component tests was good to excellent (ICC = 0.80-0.92).⁴² Reliability coefficients are shown in Table 5. A statistically significant difference (p<0.05) between the first and second test scores was identified for the GFMT Total score and for the following test items: the Hanging Pikes Test, the Vertical Jump Test, and the Splits Test.

Table 5.

Score Means and Standard Deviations for Both Test Days and Intraclass Correlation Coefficients for Test Retest Reliability (n=50).

	Rope Climb Time (sees)	Vertical Jump Height (cm)	Hanging Pikes (reps)	Shoulder Flexibility (cm/arm length)	Agility (sees)	Pull-ups (reps)	Splits (combined [cm/leg length])	Push-ups (reps)	20 Yard Sprint (sees)	Handstand (sees)	Total GFMT Score
	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)	Mean (sd)
Test Day 1 scores	5.86	40.47	16.84	0.77	19.54	8.32	−0.008	25.2	3.81	21.46	48.54
	(1.9)	(6.8)	(8.3)	(0.2)	(1.3)	(4.2)	(0.40)	(9.9)	(0.2)	(29.3)	(15.0)
Test Day 2 scores	6.61	41.93*	19.84*	0.79	19.51	8.58	−0.092*	26.42	3.82	24.70	51.26*
	(2.0)	(7.1)	(9.9)	(0.2)	(1.2)	(3.4)	(0.36)	(10.3)	(0.3)	(32.5)	(15.5)
Test re-test reliability ICC	0.80	0.83	0.88	0.92	0.86	0.89	0.91	0.84	0.85	0.92	0.97

^*Signifies statistically significant difference from Test Day 1 scores p<0.05

DISCUSSION

The GFMT provides the coaches, trainers, and therapists who work with female gymnasts of any age or competitive level with a functional tool designed to assess the unique aspects of fitness that are necessary for safe and effective participation in the sport. Given that the GFMT was developed as a field-test that can be administered without extensive training using equipment readily available in a gymnastics gym, the authors believe that the GFMT can be easily incorporated into any gymnastics program. Identifying fitness deficits that can be targeted as part of a gymnast's individual training regime may prove useful in injury prevention.

Raw data collected in this study was used to develop the scoring system for the GFMT. Transformation of the raw data for each individual item to an ordinal scale allowed for a total GFMT score out of a possible 100 points (10 points for each item) and permitted raw data based on a variety of units of measurement to be considered within a total score. As reflected in Appendix I, the raw score for the Rope Climb item reflects both the amount of time needed to complete the climb and the qualitative analysis of the climbing technique used by the gymnast during the climb. Scoring for this item thus reflects a 0 to 5 ordinal scale for time developed using the procedures outlined above as well as a 0 to 5 score for climbing technique as outlined in Appendix I.

The results of this study provide initial support for the construct validity and test-retest reliability of the GFMT. Although construct validity is only one of the many forms of validity to be considered when evaluating a measurement tool,^42,46 the relationship between a gymnast's total GFMT score and current USAG competitive level provides support for the concept that GFMT scores will vary based on a gymnast's current competitive level. Examining data from the individual items comprising the GFMT reveals that certain items such as the Jump Test, the Agility Test, and the 20-Yard Sprint Test relate more strongly to a gymnast's current competitive level than items such as the Shoulder Flexibility Test and the Splits Test. Despite the variations in the strength of the relationship between individual items and competitive level, the authors believe that all items on the GFMT must be administered to fully assess a gymnast's fitness across multiple domain areas (strength, flexibility, power, etc.). Maintaining a complete representation of fitness within the GFMT is necessary in order to adequately identify a gymnast's fitness deficits and aid in the development of a fitness program tailored to address individual fitness needs.

The procedures and methods used in this study allowed the researchers to evaluate the GFMT within the context of its intended use as a field-test to assess a gymnast's overall fitness level while minimizing the impact of gymnastic skill on testing scores.^43,44 As such, testing was conducted in a manner consistent with the sport in an environment familiar to the individual athletes. Each item on the GFMT was administered at a separate station by different testers to reflect the common practices of field-test administration. Testers were intentionally provided with detailed instructions for administering each item but did not undergo extensive or additional training. Results are therefore felt to reflect the application of the GFMT within the setting for which it was intended to be used.

The intended purpose and use of a measurement tool dictate the relative importance of various forms of reliability. Given that the GFMT was designed as a physical fitness field-test, assessment of test-retest reliability was felt to be essential. The one week interval between test administrations attempted to control for factors such as fatigue or learning effects that may have impacted a gymnast's performance while trying to avoid enough passage of time to permit a true change in a gymnast's overall fitness.

This study was limited by several factors. The total number of participants at any given USAG level ranged from 9 to 21. Increasing these numbers to ≥30 participants at each USAG level may have yielded different results. Although methods such as using physical therapy students and coaches to collect the data may have helped to reflect the use of the GFMT within the context of its intended use, greater methodological control and therefore different results may have been obtained through the use of more stringent techniques such as employing highly trained, researching physical therapists to collect the data. While attempts were made in the test-retest procedures to decrease the possibility of a practice or learning effect, the authors' recognize that such factors may have impacted score differences between the first and second administrations of the GFMT.

Further research is needed to continue the process of establishing the various types of reliability and validity of the GFMT. The possibility of correlations between total GFMT score and such factors as body composition/percentage of body fat and body mass index must be explored. Future studies should also explore the ability of the GFMT total score and individual item scores to identify a gymnast's risk for specific injuries and whether the GFMT could be used to help determine if and when an injured gymnast can safely resume high-level training and competition. Finally, since the GFMT was developed exclusively for female gymnasts, a different tool that reflects the demands and specifications of men's competitive gymnastics should also be developed.

CONCLUSION

Although the process of establishing the reliability and validity of any measurement tool is a complex and lengthy procedure, the results of this study provide initial support for the construct validity and test-retest reliability of the GFMT.

APPENDIX 1: INSTRUCTIONS FOR ADMINISTRATION OF THE GFMT.

APPENDIX 2