A Method for Predicting Maximal Strength in

SAVE OFFLINE

Journal of Strength and Conditioning Research, 2003, 17(2), 324–328 q 2003 National Strength & Conditioning Association

A Method for Predicting Maximal Strength in Collegiate Women Athletes MICHAEL HORVAT, VINCENT RAMSEY,1 CHRISTINE FRANKLIN,1 CHRISTINE GAVIN,1 TOM PALUMBO,1 AND L. ANTHONY GLASS2 Movement Studies Laboratory, University of Georgia, Athens, Georgia 30602; 2Ohio State University, Columbus, Ohio 43210.

1

ABSTRACT The purpose of this study was to develop a regression equation capable of accurately predicting a 1 repetition maximum bench press in collegiate women athletes. The findings of this study could benefit future women athletes by providing coaches and trainers with an easy method of determining maximum upper body strength in women athletes. Sixty-five University of Georgia NCAA Division 1 women athletes from 9 different sports were measured prior to the start of their season utilizing 2 repetition tests to fatigue (25 kg: REPS55; 31.8 kg: REPS70) and a 1 repetition maximum (1RM) bench press test in random order. Other independent variables that were used with a submaximal weight to predict 1RM were total body weight, lean body mass (LBM), height, and percent body fat. The variables of REPS70 and LBM were the best predictors of 1RM utilizing Pearson product correlations (r 5 0.909, p 5 0.000; r 5 0.445, p 5 0.000) and multiple regression results (R2 5 0.834, p 5 0.000) for this population. The results from this study indicate muscular endurance repetitions using an absolute weight of 31.8 kg in conjunction with LBM can be used to accurately predict 1RM bench press strength in collegiate women athletes.

Key Words: muscular endurance, strength measurement, performance prediction Reference Data: Horvat, M., V. Ramsey, C. Franklin, C. Gavin, T. Palumbo, and L.A. Glass. A method for predicting maximal strength in collegiate women athletes. J. Strength Cond. Res. 17(2):324–328. 2003.

Introduction

W

ithin the last 2 decades, the participation of women in collegiate athletics has increased dramatically. Concurrently, the need for strength and conditioning programs has increased as the level of participation has expanded for women. To prepare athletes for competition and to compare performance, an accurate prediction of muscular strength and endurance specific to women is needed. Performance standards for men cannot be applied to women because of size, 324

maturation, and hormonal variations that are gender specific. For male football athletes, the 225-lb repetitions-to-fatigue test is a submaximal estimate of a 1 repetition maximum (1RM) bench press performance that is commonly used to measure absolute muscular strength (6, 18, 25). The 1RM techniques have been used to measure muscular strength (the ability of a muscle to exert a force against a resistance in 1 maximal effort) (19) and appear to be accurate measures of upper body strength (17, 20, 25). Absolute muscular strength represents a maximal level of strength that is not adjusted for body size or fitness differences among the individuals and generally is measured in units of work or power. Relative muscular endurance load represents an amount of weight typically based on some proportion of the 1RM, total body weight, or lean body mass. The correlation between 1RM strength and the number of repetitions completed with a relative load vary according to the number of repetitions performed and training status (trained vs. untrained) (3, 4, 9, 15, 17). The correlation between 1RM strength and the number of repetitions completed with absolute loads is high (r 5 0.95), indicating that greater values of absolute muscular endurance translate to higher levels of muscular strength (13). Based on previous research, a method for estimating 1RM from at least 10 repetitions includes addition of weight continuously with a rest interval of 2–4 minutes between lifts. The approach used by the National Football League and collegiate football programs, which includes use of an absolute muscular endurance load to estimate strength, requires a player to complete as many bench press repetitions as possible with a prescribed weight of 225 lb. Although this test is not suitable for women athletes, it has been used to estimate absolute strength with reasonable accuracy (6, 18). Although a number of studies have been undertaken to develop predictive equations for men to determine absolute muscular strength (13, 17, 19, 20, 25),

Predicting Maximal Strength 325

Table 1. Personal characteristics and physical attributes. Variable*

Mean

SE

Age (y) TBW (kg) HGT (cm) PBF (%) LBM (kg) REPS55 REPS70 1RM (lb)

20.02 59.21 166.75 18.64 48.03 26.78 15.31 98.43

1.98 0.94 1.12 0.53 0.68 1.57 1.25 2.20

* TBW 5 total body weight; HGT 5 height; PBF 5 % body fat; LBM 5 lean body mass; REPS55 5 25-kg repetitions-tofatigue test; REPS70 5 31.8-kg repetitions-to-fatigue test; 1RM 5 1 repetition maximum.

little work has been done to develop similar techniques for women athletes. Several researchers have attempted to develop methods for predicting absolute muscular strength for untrained women using field tests or other upper body lifting techniques (5, 7, 12, 22), but no study has focused on the development of a predictive 1RM equation, from multiple repetitions, for NCAA collegiate women athletes. The purpose of this study was to develop and validate a method of predicting an absolute 1RM for collegiate women athletes using absolute weights of 25 kg (55 lb) and 31.8 kg (70 lb).

Methods Subjects Sixty-five women athletes 17–23 years of age from a Division 1 NCAA institution volunteered to participate in this study. Athletes were assessed prior to their competitive season as a component of their general strength and performance evaluation. Table 1 includes the demographics of height, weight, and age of the subjects from 9 sports teams at the University of Georgia. All athletes were tested prior to the in-season cycle of preseason training. Athletes whose sports were in season, who were undergoing rehabilitation, or who were adverse to testing were not included in the study. All subjects participated in an orientation about the expectations prior to consent. Upon signing the Institutional Review Board consent form, a personnel data sheet including age, height, weight, percent body fat, history of training, sport, and position was completed. Testing Procedure Three bench press tests were administered to all 65 athletes in a counterbalanced order. A warm-up prior to testing was conducted by the university’s strength and conditioning coach and included sit-ups, pushups, quadriceps stretches, hamstring stretches, and chest, back, and arm stretches with a pole.

The 1RM bench press strength was measured using a free-weight Olympic bar (20.4 kg) and plates. The subject grasped the bar with a comfortable grip with hands slightly more than shoulder width apart. A spotter assisted in lifting the bar from the rack. The standardized touch-and-go method (2) was followed, where the bar is slowly lowered to the chest and then is pressed to full extension. Bouncing the bar off the chest was not allowed, and the testers ensured that the back and buttocks remained on the bench throughout the lift. All lifts were supervised by a certified strength and conditioning specialist. Prior to the 1RM test, a warm-up set of 10 repetitions with an unweighted bar was performed and completed by all participants. The University of Georgia Strength and Conditioning Program procedure for the 1RM bench press test was utilized because the protocol was familiar to the athletes and was consistent with the protocol suggested by Bachle et al. (2). A recovery time of 3 minutes between sets was allowed. Weight was added to each consecutive lift until the subject could not perform the lift successfully. The highest weight lifted successfully was recorded as the 1RM. Testing was completed within a 10- to 14-day period in conjunction with other performance measures conducted by the strength and conditioning coach but not on consecutive days. Each subject was required to perform a repetition test to failure using a 25-kg (55 lb) or 31.8-kg (70 lb) barbell. Following the warm-up procedure, the subject grasped the bar using the same format as the 1RM test. The bar was lowered to the chest without bouncing, and the arms were fully extended on each repetition. If there was a significant hesitation ($2 seconds) or a repetition was not completed with proper form, the test was terminated. The highest number of properly executed repetitions was recorded as the repetition maximum. The 25 kg (55 lb) used for one of the tests in this project is slightly more than the 15.9 kg (35 lbs) used in the YMCA test for untrained women (10). The weight selected was also based on the background and training of the participants and pilot testing in which 5 athletes performed a minimum of 20 repetitions with the 25-kg weight. An additional test weight of 31.8 kg was utilized based on prior strength data that revealed the average 1RM for University of Georgia female athletes to be 47.2 kg (104 lb) and represented approximately 70% of the maximum lifts. All strength tests were randomized, and none were conducted on consecutive days Statistical Analyses All data were entered into the SPSS 10 statistical analysis software program (24). Means and SDs were calculated for all variables. Pearson product correlation coefficients were used to determine relationships be-

326 Horvat, Ramsey, Franklin, Gavin, Palumbo, and Glass Table 2. Correlation of selected variables included in regression analysis. Variable

LBM

1RM

REPS70

1RM REPS70 REPS55

0.445** 0.374** 0.282*

0.909** 0.866**

0.938**

* p 5 0.05. ** p 5 0.01.

tween the variables. A multiple linear regression analysis was used to determine the equation that best predicted the 1RM bench press. A random split case analysis was conducted to determine the validity of the prediction equation (24). All analyses were performed at the a 5 0.05 level, and the power analysis was 0.96. Variables that were included in the analysis were 1RM weight (dependant variable), 25-kg repetitions (REPS55), 31.8-kg repetitions (REPS70), height (HGT), total body weight (TBW), lean body mass (LBM), and percent body fat (PBF) (all possible explanatory variables).

Results Means and SEs for the dependant (1RM) and possible explanatory variables considered are given in Table 1. Height (0.44) and percent body fat (0.51) had the smallest SEs of all variables. The remaining variables also exhibited small SEs, indicating the homogeneous composition of this subject sample. A Levene test was utilized to compare the homogeneity of variance between groups (sport) by variable. Because the p values are .0.05 for all comparisons, there is no indication that the groups differed significantly (24). At 25 kg, the minimum number of repetitions obtained from testing with was 3 (a member of the golf team), and the maximum number was 47 (a gymnastics athlete). At 31.8 kg, the minimum number of repetitions was 0 (a member of the golf team), and the maximum number was 34 (a gymnastics athlete). A stepwise regression was used to analyze which of the 6 possible explanatory variables were useful in predicting 1RM. The 3 explanatory variables, REPS70, REPS55, and LBM had p values of ,0.05, indicating significant unique contributions to the prediction of

1RM. The correlation matrix for the 3 explanatory variables and the dependant variable, 1RM, that were included in the regression analysis are presented in Table 2. The highest Pearson correlations were found between repetitions at 25 and 31.8 kg (r 5 0.938, p , 0.001), repetitions at 31.8 kg and 1RM (r 5 0.909, p , 0.001), and repetitions at 25 kg and 1RM (r 5 0.866, p , 0.001). Additional significant correlations were TBW and LBM (r 5 0.916, p , 0.001) and HGT and LBM (r 5 0.750, p , 0.001). In addition to the Pearson product correlation (24), a colinear multivariate regression was employed to determine a valid equation for predicting a 1RM bench press; one analysis was based on 25 kg and one was based on 31.8 kg. A split-case, cross-validation design was also utilized to validate the predictive equation (21). The sample population was randomly split into 2 groups. A stepwise regression analysis was performed on one sample and then validated on the remaining sample of subjects. Results of this analysis technique revealed similar adjusted R2 values and predictive equations (Table 3). These results were further validated by a t-test. Along with the 2 split-case models (1 and 2), 2 total models (3 and 4) were developed that resulted in similar adjusted R2 values and similar coefficients (Table 3). These models were also validated by a t-test. All 4 models were significant at the a 5 0.01 level. The equations encompassing the total sample for REPS70 and REPS55 yielded adjusted R2 values of 0.832 and 0.787, respectively, indicating that the explanatory variables REPS70 and LBM accounted for more than 83% of the explained variability in 1RM and the variables REPS55 and LBM accounted for approximately 79% of the explained variability in 1RM. Both of these results may be applicable to other similar populations for predicting absolute maximum upper body strength for collegiate women athletes. The results of this study indicte that LBM and the estimated 1RM derived from the regression equation using multiple repetitions-to-failure (REPS70) with a submaximal weight (31.8 kg) are highly correlated (Figure 1) with the actual maximum bench press performance (r 5 0.916, p , 0.001). Figure 2 represents the correlation of the REPS55 and LBM regression prediction and the actual 1RM values (r 5 0.891, p , 0.001). A final assessment uti-

Table 3. Split-case and overall regression models. Model 1 2 3 4

Equation 1RM 1RM 1RM 1RM

5 5 5 5

53.165 57.037 54.592 31.857

1 1 1 1

(1.61)REPS70 (1.72)REPS70 (1.68)REPS70 (1.29)REPS55

1 1 1 1

(0.197)LBM (0.125)LBM (0.167)LBM (0.298)LBM

N

Adjusted R2

32 33 65 65

0.781 0.870 0.834 0.787

Predicting Maximal Strength 327

Figure 1. Correlation between actual and predicted 1RM values using the REPS70 and LBM regression equation (r 5 0.916).

Figure 3. Correlation between actual and predicted 1RM values using the REPS70 regression equation (r 5 0.909).

ing the explanatory variables of REPS70 and LBM would more accurately predict a 1RM for this group of trained athletes based on their higher levels of strength and conditioning and lower overall percentage of body fat compared with their untrained peers.

Discussion

Figure 2. Correlation between actual and predicted 1RM values using the REPS55 and LBM regression equation. (r 5 0.891).

lizing a Pearson product also revealed high correlations between predicted 1RM values generated by the REPS70/LBM and REPS55/LBM regression equations (0.951, p , 0.001). The results of this correlation of the predicted values from each regression model corroborate the strong relationship between the individual variables of REPS70 and REPS55. An initial stepwise analysis of the data revealed significant regression models of 1 (REPS70) and 2 (REPS70, LBM) variables. The 1-variable model using only REPS70 to predict the 1RM yielded an adjusted R2 value of 0.815 and a strong correlation (r 5 0.909, Figure 3), whereas the 2-variable model of REPS70 and LBM produced an adjusted R2 of 0.832. These results combined with the high correlations between the variables REPS70 and REPS55 (r 5 0.938) could indicate that the 1-variable model would be adequate in predicting a 1RM bench press for this population. However, we concluded that the prediction equation utiliz-

Previous studies have indicated that submaximal bench press repetition tests of 80–90% of the actual 1RM are more accurate than tests requiring repetitions of 10 or more (1, 4, 16, 20), whereas other research has shown that repetition tests using 55–95% of the 1RM are highly correlated (r 5 0.94) with other maximum lifts such as the leg press (3). Mayhew et al. (18) revealed strong correlations between predicted and actual 1RMs (r 5 0.96) with 10 or less repetitions in a sample of collegiate football players. Our study population are considered ‘‘power’’ athletes and therefore require a much higher degree of explosive upper body strength. These findings also are consistent with those of Morales and Sobonya (20), who reported a similar variance and 1RM predictability in class athletes. The athletes tested in the current study have a much different body composition and rely much more on explosive power and muscular endurance than on absolute upper body strength, except perhaps for the gymnasts. Based on the data analysis, we concluded that muscular endurance repetitions with an absolute load of 25 or 31.8 kg combined with other components such as LBM can be used to accurately estimate a 1RM bench press for collegiate women athletes. These results are similar to those reported by Harman et al. (11) for women athletes but are slightly higher than estimates for untrained women in an earlier study (23). In addition, our values for the ratios of 1RM weight to LBM and TBW are higher than those reported in previous studies (8).

328 Horvat, Ramsey, Franklin, Gavin, Palumbo, and Glass

Practical Applications These data indicate that LBM and repetition tests with 25 and 31.8 kg are highly correlated with the 1RM bench press test. The 31.8-kg load was selected as the best predictor from the analyses. This finding was consistent with previous findings that indicated loads between 70% and 95% of the 1RM produce the most accurate indicators of maximal strength (6, 7, 14, 18, 20). In general, the subjects in the present study had experience in resistance training. Whether the weights achieved would generalize to untrained or younger women is unclear and is a fertile area for future research. For a population that is emerging in overall performance and resistance training, this project was a first step in accurately gauging strength levels in women athletes. Submaximal repetition tests for these athletes could be useful for individuals who do not often test for levels of maximum absolute strength and for those who are unfamiliar with proper lifting techniques. This type of testing would also provide coaches and trainers with a simple and effective method of testing trained and untrained women athletes that could be easily incorporated into existing lifting regimes.

References 1.

2.

3.

4.

5.

6.

7.

8.

ARNOLD, M.D., J.L. MAYHEW, D. LESEUR, AND M. MCCORMICK. Accuracy of predicting bench press and squat performance from repetitions at low and high intensity [Abstract]. J. Strength Cond. Res. 9:205–206. 1995. BACHLE, T.R., R.W. EARLE, AND D. WORTHEN. Resistance training. In: Essentials of Strength Training and Conditioning. T.R. Baechle and R.W. Earle, eds. Champaign, IL: Human Kinetics, 2000. pp. 395–425. BALL, T.E., C. VAN FLEET, R.J. LAHEY, AND A.L. GLASS. The relationship of relative muscular endurance to maximal lifting capacity in college men and women [Abstract]. J. Strength Cond. Res. 9:277. 1995. BATES, E.L., C.T. CLEMENS, J.S. BUSBY, J.L. MAYHEW, J.E. GLOVER, AND M.G. BEMBEN. Measurement specificity in the determination of 1-RM bench press strength changes following training [Abstract]. J. Strength Cond. Res. 9:277. 1995. CHANDLER, T.G., S.C. WEST, R.B. LARKIN, B.J. CRADY, AND J.L. MAYHEW. Relationship of pullup and lat-pull performances to 1-RM lat-pull strength [Abstract]. J. Strength Cond. Res. 9:205. 1995. CHAPMAN, P.P., J.R. WHITEHEAD, AND R.H. BRINKERT. The 225 lb reps-to-fatigue test as a submaximal estimate of 1-RM bench press performance in college football players. J. Strength Cond. Res. 12:258–261. 1998. CUMMINGS, B., AND K.J. FINN. Estimation of a one repetition maximum bench press for untrained women. J. Strength Cond. Res. 12:262–265. 1998. DALTON, N.J., AND J.E. WALLACE. Strength testing protocols for college-age women. Natl. Strength Cond. Assoc. J. 16:7–10. 1996.

9.

10. 11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

EVERETT, H.P., A.B. JOUNER, AND J.L. MCMILLAN. The relationship among absolute endurance, relative endurance, and 1 RM tests for upper body strength and endurance in college males and females [Abstract]. J. Strength Cond. Res. 9:279. 1995. GOLDING, L.A. (ed.). YMCA Fitness Testing and Assessment Manual (4th ed.). Champaign, IL: Human Kinetics, 2000. HARMAN, E., J. GARHAMMER, AND C. PANDORF. Administration, scoring, and interpretation of selected tests. In: Essentials of Strength Training and Conditioning. T.R. Baechle and R.W. Earle, eds. Champaign, IL: Human Kinetics, 2000. p. 309. KURAMOTO, A.K., AND V.G. PAYNE. Predicting muscular strength in women: A preliminary study. Res. Q. Exerc. Sport. 66:168–172. 1995. LESUER, D.A., J.H. MCCORMICK, J.L. MAYHEW, R.L. WASSERTEIN, AND M.D. ARNOLD. The accuracy of prediction equations for estimating 1-RM performance in the bench press, squat, and deadlift. J. Strength Cond. Res. 11:211–213. 1997. MAYHEW, J., T.E. BALL, M. ARNOLD, AND J. BROWN. Relative muscular endurance performance as a predictor of bench press strength in college men and women. J. Appl. Sport Sci. Res. 6: 200–206. 1992. MAYHEW, J.L., T.E. BALL, AND J.C. BOWEN. Prediction of bench press lifting ability from submaximal repetitions before and after training. Sports Med. Train. Rehabil. 3:195–201. 1992. MAYHEW, J.L., J.C. CLEMENS, K.L. BUSBY, J.S. CANNON, J.S. WARE, AND J.C. BOWEN. Cross-validation of equations to predict 1-RM bench press from repetitions-to-failure [Abstract]. Med. Sci. Sport Exerc. 27:S209. 1995. MAYHEW, J.L., J.L. PRINSTER, J.S. WARE, D.L. ZIMMER, J.R. ARBAS, AND M.G. BEMBEN. Muscular endurance repetitions to predict bench press strength in men of different training levels. Sports Med. Phys. Fitness 35:108–113. 1995. MAYHEW, J.L., J.S. WARE, M.G. BEMBEN, B. WILT, T.E. WARD, B. FARRIS, J. JURASZEK, AND J.P. SLOVAK. The NFL-225 test as a measure of bench press strength in college football players. J. Strength Cond. Res. 13:130–134. 1999. MAYHEW, J.L., J.S. WARE, AND J.L. PRINSTER. Using lift repetitions to predict muscular strength in adolescent males. Natl. Strength Cond. Assoc. J. 15:35–38. 1993. MORALES, J., AND S. SOBONYA. Use of submaximal repetition tests for predicting 1-RM strength in class athletes. J. Strength Cond. Res. 10:186–189. 1996. PEDHUZUR, E.J. Multiple Regression in Behavioral Research. Explanation and Prediction (3rd ed.). Fort Worth, TX: Harcourt Brace College Publishers, 1997. ROSE, K., AND T.E. BALL. A field test for predicting maximum bench press lift of college women. J. Appl. Sport Sci. Res. 6:103– 106. 1992. SIMPSON, S.R., R. ROZNEK, J. GARHAMMER, M. LACOURSE, AND T. STORER. Comparison of one repetition maximums between free weights and universal machine exercises. J. Strength Cond. Res. 11:103–106. 1997. SPSS. Base 10.0 for Windows User’s Guide. Upper Saddle River, NJ: SPSS Inc., 1999. WARE, J.S., C.T. CLEMENS, J.L. MAYHEW, AND T.J. JOHNSTON. Muscular endurance repetitions to predict bench press and squat strength in college football players. J. Strength Cond. Res. 9:99–103. 1995.

Address correspondence to: Michael Horvat, mhorvat@ coe.uga.edu.