- Email: [email protected]
Assessment of grip force and subjective hand force exertion under handedness and postural conditions
Applied Ergonomics 42 (2011) 929e933
Contents lists available at ScienceDirect
Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo
Assessment of grip force and subjective hand force exertion under handedness and postural conditions Kai Way Li a, *, Ruifeng Yu b a b
Department of Industrial Management, Chung Hua University, 707, Sec 2, WuFu Road, Hsin-Chu, Taiwan Department of Industrial Engineering, Tsinghua University, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 July 2010 Accepted 1 March 2011
The Borg CR-10 scale has been used to quantify the perception of physical exertion. An experiment was conducted to test the grip force of males on four levels of the CR-10 scale under experimental conditions. It was found that the subjects applied higher grip forces than they perceived at levels 2, 5, and 7 on the scale. The grip forces between dominant and non-dominant hands at low CR-10 levels were negligible. The grip forces were significantly different between the two hands at level 10. Similar results were found for the postural conditions. A follow-up experiment was conducted to estimate the subjective rating when applying a pre-determined grip force under the same conditions. Regression models were established to link the relationship between the subjective rating and hand force. The estimated ratings were lower than those actual values under all the tested conditions, even though the models have high R2 values. Ó 2011 Elsevier Ltd and The Ergonomics Society. All rights reserved.
Keywords: Borg CR-10 Grip force Hand exertion Subjective rating
1. Introduction Forceful exertion has been considered one of the major contributing factors to the occurrence of upper extremity workrelated musculoskeletal disorders (Silverstein et al., 1986; Moore and Garg, 1994). Electromyography (EMG) has been adopted in measuring muscular exertions when specific muscle groups are identified in performing various tasks. Load cells and force sensors, mounted on handles to capture the force signals, may also be used to measure muscular effort especially in push and pull tasks. Both the EMG and load cells are, however, seldom used by practitioners in field measurements because preparation and post data processing are time consuming. Forceful hand exertions may also be measured indirectly using either a dynamometer or a strain gauge. The force-matching method has been proposed to measure hand forces (Wiktorin et al., 1996; Casey et al., 2002; Bao and Silverstein, 2005). This method required a subject to grip a dynamometer immediately after a task been performed to mimic the amount of force exerted. The reading on the dynamometer was recorded as the grip force applied during the task. Large between-subject variation, however, has been reported using this method (Koppelaar and Wells, 2005; Bao et al., 2009).
* Corresponding author. Tel.: þ886 3 518 6583; fax: þ886 3 518 6575. E-mail address: [email protected] (K.W. Li).
Subjective rating is one of the alternatives to measure forceful hand exertion. This approach is probably the most cost-effective and easy way to obtain force exertion data especially in large population studies (Hjelm et al., 1995; Bao et al., 2009). Borg’s concept of perceived exertion (Borg, 1970) has been adopted in quantifying subjective exertion. His ratings of perceived exertion (RPE) scale (ranged from 6 to 20) have been widely used in analyzing various manual tasks and physical works as a supplementary measure in addition to physiological measures (Wang et al., 2000; Chung and Wang, 2001; Li et al., 2007, 2009). The RPE scale has also been used in determining the perceived weight handled (Deeb, 1999). In addition to the RPE scale, Borg (1982) has constructed a category ratio scale, called CR-10 scale. This scale, ranged from 0 (nothing at all) to 10 (extremely strong or almost maximal), provides direct level estimates for determinations of ratio relationships between perceptual responses (Borg, 1990). Based on this scale, Buchholz et al. (2008) examined the subjective ratings of upper extremity exposures, wrist and shoulder postures, and grip force for automobile workers. They found significant correlation between subjective rating of manual effort and grip force. They concluded that worker’s self-report is an alternative to more resource-intensive exposure assessment methods. Park and Yun (1996) conducted a study to estimate hand force exertion using the CR-10 rating. The correlation coefficients between the measured force and perceived exertion for each of their trial ranged from 0.8 to 0.99.
0003-6870/$ e see front matter Ó 2011 Elsevier Ltd and The Ergonomics Society. All rights reserved. doi:10.1016/j.apergo.2011.03.001
930
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
Spielholz (2006) conducted a study to evaluate the efficacy of calibrating subjective worker ratings of hand exertions. The subjects applied power grip forces corresponding to their perceptions of different Borg CR-10 scale levels using the so-called “gripto-scale” procedure. The data were used to define relationships between subjective ratings and actual force application. Gripping tasks were performed and corresponding subjective hand force ratings were calibrated using the grip-to-scale calibration data. The estimation error for power grip task was found significantly reduced using this calibration procedure. Spielholz’s (2006) study indicated that calibration of hand force ratings using the grip-toscale procedure improved the accuracy of hand exertion measurements using the CR-10 scale. The Borg’s CR-10 rating has been discussed in the literature. But whether usage of this rating scale in quantifying power grip force is affected by dominant or non-dominant hand has not been discussed. Neither have the effects of arm posture been studied. The objectives of this study were to i) test the hypothesis that the exertions of power handgrip at the same perceived exertion levels between dominant hand and non-dominant hand are different; ii) test the hypothesis that hand posture affects the exertion of power grip measured at the same perceived exertion levels; and iii) compare the subjective ratings with the estimations based on regression models established using the actual CR-10 rating and power grip force data. 2. Methods This study was conducted in the laboratory.
right-handed. They were requested to refrain from physical activities at least 1 h before joined the experiment. 2.2. Apparatus A TAKEIÒ 5001 hand dynamometer was used to measure the grip force of the hand. The dynamometer was calibrated by the supplier before the experiment. A PEAKLIEÒ FT515 handgrip to build up hand/arm muscles was also adopted. A Borg CR-10 scale (Borg, 1982) was used for subjective rating. 2.3. Hand/posture conditions The hand exertions were measured under two hand conditions and two posture conditions. The measurements were conducted using either dominant hand or non-dominant hand. The postures included 90 and 180 at the elbow (see Fig. 1). 2.4. Grip force measurements Before the study, the subjects were verbally instructed the nature and use of the CR-10 scales. The grip-to-scale calibrating procedure proposed by Spielholz (2006) was adopted. In this procedure, each subject gripped the dynamometer to one of the four CR-10 levels: 2 (weak), 5 (strong), 7 (very strong) and 10 (extremely strong). The grip span of the dynamometer was 5 cm. The power grip sustained for 4 s and the peak value of the grip force was recorded. Each subject took a break for 5 min or more after he finished one grip measurement so as to avoid the effects of fatigue. Each subject performed no more than four grips at the same day.
2.1. Subjects
2.5. Subjective ratings for a pre-determined hand force
Twenty male college students were recruited as human subjects. Based on self-reports, all of them were free from musculoskeletal injuries. Their age, stature, and body weight were 22.1 (2.5) yrs, 172.3 (5.3) cm, and 69.8 (12.5) kg, respectively. All the subjects received payment and had signed informed consent for their participation in the study. All the subjects, except one, were
The handgrip was adopted as a basis to examine the CR-10 rating for the subjects under different experimental conditions. The subjects were requested to apply a 10 kgf force on the handgrip without knowing the force level (see Fig. 2). The grip span was approximately 4 cm. Each subject gripped and sustained for 4 s. He, then, reported a CR-10 rating of the hand exertion.
Fig. 1. Hand posture at (a) 90 and (b) 180 .
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
Fig. 2. Hand grip applying a 10 kgf force.
2.6. Experiment design & data analysis The grip force measurement experiment was conducted using a three-factor completely randomized design. The factors were the level of exertion, hand used, and posture. Four samples were collected for each treatment. For each subject, the order of the experiment was randomized among all the conditions. The subject completed four trials for the same treatment in the same visit. Experimental errors were assumed to be evenly distributed among all treatments under this randomization arrangement. A total of 1280 (4 CR-10 levels 2 hands 2 postures 4 replicates 20 subjects) measures were collected. Both the descriptive statistical analysis and analysis of variance (ANOVA) were performed. Duncan’s multiple range test was conducted if a factor was found statistically significant at a ¼ 0.05 level. In addition, linear regression equations without intercept were built to describe the relationship between the CR-10 rating and grip force under each hand/posture condition for each subject:
CR 10 rating ¼ b grip force þ 3i
(1)
where b is the regression coefficient and 3i is a random error of the ith trial which is normally distributed with a mean of 0 and a variance of s2. In the handgrip test, the CR-10 ratings after applying a 10 kgf power grip force were collected under a two-factor completely randomized design with one trial in each treatment. A total of 80 measures were collected (2 hands 2 postures 20 subjects). These CR-10 ratings were compared with the estimated rating calculated using Eq. (1). Analyses of variance were performed for the actual value, estimated value, and estimation error of the CR-10 rating. The statistical analyses were performed using the SPSSÒ 12.0 computer software.
931
multiple range test results comparing the two postures showed that the grip force at the 180 posture (28.3 kgf) was significantly (p < 0.05) higher than that at the 90 posture (26.5 kgf). Dominant hand (28.0 kgf) showed a significantly (p < 0.05) higher grip force than that of the non-dominant hand (26.7 kgf). The interaction effects of the CR-10 level and the hand used were significant (p < 0.0001). The interaction effects of the CR-10 level and the posture were also significant (p < 0.0001). The interactions of these effects are shown in Figs. 3 and 4, respectively. The overall Pearson’s correlation coefficient between the CR-10 rating and the grip force was 0.92 (p < 0.0001). A total of 80 linear regression models were established for each hand-posture condition for each subject. All the regression models were statistically significant at p < 0.0001 with coefficients of determination (R2) of 0.96 or higher. All the hypotheses testing whether the regression coefficient was zero were rejected (p < 0.0001). Table 1 shows the means and standard deviations of the estimated regression coefficients. In the handgrip test, the ANOVA results showed that the effects of both the hand and posture on the reported subjective rating were not statistically significant. For the estimated rating, however, the effects of both the hand and posture were significant. Nondominant hand (2.36) showed significant (p < 0.05) higher values than those of dominant hand (2.22). The estimated subjective ratings at 90 posture (2.37) were significantly (p < 0.01) higher than those at the 180 posture (2.20). The effects of hand and posture on the estimation error were not statistically significant. Table 2 shows the means and standard deviations of actual values, estimated values, and estimation errors of the subjective rating. 4. Discussions The 2, 5, 7, and 10 on the CR-10 scale were assumed to represent 20%, 50%, 70%, and 100% of perceived maximum voluntary contraction (MVC), respectively. The results of the study indicated a strong linear correlation between the CR-10 rating and grip force. This was consistent with the findings in the literature (Borg, 1990; Park and Yun, 1996). The interaction effects between exertion level and hand used in Fig. 3 showed that at low CR-10 ratings such as 2 and 5, the grip forces between dominant and non-dominant hands were negligible. At level 10, however, the grip force of dominant hand (43.7 kgf) was higher than that of the non-dominant hand (40.1 kgf). This was consistent with the 10% rule that dominant hand possesses a 10% greater grip strength than non-dominant hand (Petersen et al., 1989). It should be noted that, however, the grip forces at level 10 for the two hands were lower than the grip strengths of a similar group reported by Ho et al. (2000). This might due to the fact that level 10 on the CR-10 scale represents an “almost maximal” force exertion but not the maximal. When divided the grip forces at levels of 2, 5, and 7 by the force at level 10, the grip forces were converted into %MVC. The %MVC at levels of 2, 5, and 7 for dominant hand were 24.7%, 56.2%, and
3. Results The ANOVA results showed that the main effects of the CR-10 level, hand used, and posture on power grip force were all significant (p < 0.0001). The Duncan’s multiple range test results showed that the mean grip force at the level 10 of the CR-10 (41.9 kgf) was significantly (p < 0.05) higher than those at levels 7 (32.6 kgf), 5 (24.4 kgf), and 2 (10.5 kgf). The grip force at level 7 was significantly (p < 0.05) higher than those at levels 5 and 2. The grip force at level 5 was significantly higher than that at the level 2. Duncan’s
Fig. 3. Grip force at different CR-10 ratings for dominant and non-dominant hands.
932
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933 Table 2 Means and standard deviations of actual values, estimated values, and estimation errors of the CR-10 rating. Actual CR-10 rating
Fig. 4. Grip force at different CR-10 ratings under two postures.
75.6%, respectively. The %MVC values at the three levels for nondominant hand were 25.6%, 60.5%, and 80.0%, respectively. All the % MVC values were higher than their corresponding CR-10 levels times 10. This implies that the subjects applied a force exertion higher than they perceived. Comparing the %MVC between the two hands at the 2, 5, and 7 levels, non-dominant hands showed higher values than those of the dominant hand at all three levels. A possible explanation of this phenomenon was, as mentioned in the previous paragraph, that the level 10 did not correspond to the maximal but the sub-maximal of the force exertion. The interaction effects between the exertion level and posture in Fig. 4 showed similar trends as in Fig. 3. The grip forces between the two postures at low exertion levels such as 2 and 5 were negligible. At level 10, the grip forces at 180 posture were higher than those at the 90 . The %MVC at levels of 2, 5, and 7 for the 90 posture were 26.1%, 59.2%, and 79.0%, respectively. The %MVC values at the three levels for the 180 posture were 24.3%, 57.4%, and 76.6%, respectively. All the %MVC values were higher than their corresponding CR-10 levels times 10. The %MVC values for the 90 posture were higher than those of the 180 posture for all the 2, 5, and 7 levels. This might be attributed to the fact that, as compared to the 180 posture, the 90 posture was associated with a flexion of the arm which required higher muscular exertion. Buchholz et al. (2008) examined the agreement of subjective ratings of upper extremity exposures with corresponding direct measurements obtained from workers. Psychophysical ratings of exposure based on the CR-10 were obtained for the period of time in which direct measurements were acquired. Ratings of manual effort were significantly correlated with directly measured grip force. Grant et al. (1994) reported that the perceived exertion, based on the CR-10, showed strong and positive correlation with grip force. These were consistent with the findings our study. It might be noteworthy that the correlation coefficient between the subjective rating and the grip force in the current study (0.92) was higher than those (0.33e0.72) reported by Grant et al. (1994). This difference might be attributed to the force ranges in the two studies. The force range in Grant et al. (1994) was more restricted (1%e46% MVC) than the one (0%e100% MVC) adopted in the current study. The regression coefficients in Table 1 indicate the increase of the CR-10 rating due to an increase of grip force depend on the hand used and posture. With each unit (kgf) increase in grip force, the CR-10 ratings at the 90 posture increased 0.24 and 0.23 for non-dominant and dominant hands, respectively. At the 180 posture, the increases were 0.23 and 0.21 for non-dominant and dominant hands, respectively. A pair-wised t-test results comparing the regression coefficients showed that the difference between dominant and nonTable 1 Means and standard deviations of estimated regression coefficients. Posture
Non-dominant hand
Dominant hand
90 180
0.24 (0.03) 0.23 (0.03)
0.23 (0.03) 0.21 (0.02)
Estimated CR-10 rating
Estimation error
Non-dominant hand 3.93 (1.78) 90 180 3.83 (1.77)
2.43 (0.03) 2.29 (0.28)
1.50 (1.73)** 1.54 (1.71)**
Dominant hand 90 3.50 (1.57) 180 3.20 (1.61)
2.32 (0.30) 2.12 (0.23)
1.18 (1.51)* 1.08 (1.54)*
Note: Estimation error ¼ actual CR-10 rating estimated CR-10 rating; *p < 0.01, **p < 0.001.
dominant hands was statistically significant (p < 0.001). In Table 1, lower increases in the CR-10 rating associated with an increase in grip force were also observed for the 180 posture than those of the 90 posture. A pair-wised t-test results comparing the regression coefficients also showed that the difference between the two postures was statistically significant (p < 0.001). Table 2 shows the actual ratings, estimated ratings, and the differences between the two for applying a grip force of 10 kgf. The subjects reported slightly higher CR-10 ratings when using their non-dominant hands than those of dominant hands even though the difference was not statistically significant. This is reasonable as the non-dominant hands have weaker muscles and need to apply a higher exertion so as to generate the same amount of force as compared to the dominant hands. The subjects also reported slightly higher CR-10 ratings when applying a force at the 90 posture than at the 180 posture even without statistical significance. The R2 represents the percentage of variation of the dependent variable which could be explained by the independent variable. An R2 of 0.96 or higher indicates that almost all the variation of the subjective rating may be explained by the grip force. It was shown in Table 2 that the estimated CR-10 ratings were significantly (p < 0.01) lower than those of the actual values under all the hand/ posture conditions. For non-dominant hand, the estimation errors for the 90 and 180 postures were 1.50 (38.1%) and 1.54 (40.2%), respectively. For dominant hand, the estimation errors for the two postures were 1.18 (33.7%) and 1.08 (33.7%), respectively. The subjects reported subjective scores higher than those estimations based on their own perception and grip force data even though the corresponding regression models had very high R2 values. The results of the current study indicated that the CR-10 rating scale has high correlation with the power grip force. This was consistent with those in the literature (Grant et al., 1994; Park and Yun, 1996; Buchholz et al., 2008). However, perceived exertions of power grip might be affected by both the hand used and arm posture. In addition, the subjects had a tendency to report the subjective ratings of power grip exertion higher than the actual % MVC values. 5. Conclusion An experiment was conducted to test the grip force of male subjects at four perceived exertion levels on the CR-10 scale under handedness and postural conditions. Generally speaking, the subjects tended to apply a higher power grip force (%MVC) than they perceived at levels 2, 5, and 7 on the CR-10 scale. It was found that the grip forces between dominant and non-dominant hands at low CR-10 levels were negligible. The grip forces were significantly different between the two hands at level 10. Similar results were found for the arm posture conditions. The grip forces between the two postures at low CR-10 levels were negligible. At level 10, the grip forces at 180 posture were significantly higher than those at
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
the 90 . The overall correlation coefficient between the CR-10 rating and the grip force was significant (r ¼ 0.92; p < 0.0001). Eighty regression models have been established between the CR-10 rating and the grip force. The subjects reported higher subjective ratings than the estimations which were based on their own perception and grip force data, even though the regression models had very high R2 values. Acknowledgments This research was financially support by the National Science Council of the ROC under grant NSC 98-2221-E-216-013. The author thanks Shang Wei Tseng and Tsung-Yu Pei, graduate students with the Department of Industrial Management at Chung-Hua University, for their assistance in the study. References Bao, S., Silverstein, B., 2005. Estimation of hand force in ergonomic job evaluations. Ergonomics 48, 288e301. Bao, S., Spielholz, P., Howard, N., Silverstein, B., 2009. Force measurement in field ergonomic research and application. Int. J. Ind. Ergon. 39, 333e340. Borg, G., 1970. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med. 2 (3), 92e98. Borg, G., 1982. Psychophysical bases of perceived exertion. Med. Sci. Sports. Exerc. 14 (5), 377e381. Borg, G., 1990. Psychophysical scaling with applications in physical work and the perception of exertion. Scand. J. Work. Environ. Health. 16 (10), 55e58. Buchholz, B., Park, J.-S., Gold, J.E., Punnett, L., 2008. Subjective ratings of upper extremity exposures: inter-method agreement with direct measurement of exposures. Ergonomics 51 (7), 1064e1077.
933
Casey, J.S., McGorry, R.W., Dempsey, P.G., 2002. Getting a grip on grip force estimates: a valuable tool for ergonomic evaluations. Prof. Saf., 18e24. Chung, H.C., Wang, M.J., 2001. The effects of container design and stair climbing on maximal acceptable lift weight, wrist posture, psychophysical, and physiological responses in wafer-handling tasks. Appl. Ergon. 32, 593e598. Deeb, J.M., 1999. Muscular fatigue and its effects on weight perception. Int. J. Ind. Ergon. 24, 223e233. Grant, K.A., Habes, D.J., Putz-Anderson, V., 1994. Psychophysical and EMG correlates of force exertion in manual work. Int. J. Ind. Ergon. 13, 31e39. Hjelm, E.W., Winkel, J., Nygard, C.H., Wiktorin, C., Karlqvist, L., 1995. Can cardiovascular load in ergonomic epidemiology be estimated by self-report? Stockholm MUSIC 1 study group. J. Occup. Environ. Med. 37, 1210e1217. Ho, R.W.D., Chang, S.Y., Wang, C.W., Hwang, M.H., 2000. Grip and key pinch strength: norms for 15- to 22-year-old Chinese students. Chin. Med. J. 63 (1), 21e27. Koppelaar, E., Wells, R., 2005. Comparison of measurement methods for quantifying hand force. Ergonomics 48, 983e1007. Li, K.W., Yu, R-f., Han, X.L., 2007. Physiological and psychophysical responses in handling maximum acceptable weights under different footweardfloor friction conditions. Appl. Ergon. 38, 259e265. Li, K.W., Yu, R-f., Gao, Y., Maikala, R.V., Tsai, H.-H., 2009. Chinese workers performing combined manual materials handling tasks. Int. J. Ind. Ergon. 39, 422e427. Moore, J.S., Garg, A., 1994. Upper extremity disorders in a pork plant: relationship between task factors and morbidity. Am. Ind. Hyg. Assoc. J. 55, 703e715. Park, D., Yun, M.H., 1996. An application of psychophysical ratings to external force estimation. Comput. Ind. Engng. 31, 675e680. Petersen, P., Petrick, M., Connor, H., Conklin, D., 1989. Grip strength and hand dominance: challenging the 10% rule. Am. J. Occup. Ther. 43 (7), 444e447. Silverstein, B.A., Fine, L.J., Armstrong, T.J., 1986. Hand wrist cumulative trauma disorders in industry. Br. J. Ind. Med. 43, 779e784. Spielholz, P., 2006. Calibrating Borg scale ratings of hand force exertion. Appl. Ergon. 37, 615e618. Wang, M.J., Chung, H.C., Chen, H.C., 2000. The effect of handle angle on MAWL, wrist, posture, RPE, and heart rate. Hum. Factors. 42 (4), 553e565. Wiktorin, C., Selin, K., Ekenvall, L., Kilbom, A., Alfredsson, L., 1996. Evaluation of perceived and self-reported manual forces exerted in occupational materials handling. Appl. Ergon. 27, 231e239.
Contents lists available at ScienceDirect
Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo
Assessment of grip force and subjective hand force exertion under handedness and postural conditions Kai Way Li a, *, Ruifeng Yu b a b
Department of Industrial Management, Chung Hua University, 707, Sec 2, WuFu Road, Hsin-Chu, Taiwan Department of Industrial Engineering, Tsinghua University, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 July 2010 Accepted 1 March 2011
The Borg CR-10 scale has been used to quantify the perception of physical exertion. An experiment was conducted to test the grip force of males on four levels of the CR-10 scale under experimental conditions. It was found that the subjects applied higher grip forces than they perceived at levels 2, 5, and 7 on the scale. The grip forces between dominant and non-dominant hands at low CR-10 levels were negligible. The grip forces were significantly different between the two hands at level 10. Similar results were found for the postural conditions. A follow-up experiment was conducted to estimate the subjective rating when applying a pre-determined grip force under the same conditions. Regression models were established to link the relationship between the subjective rating and hand force. The estimated ratings were lower than those actual values under all the tested conditions, even though the models have high R2 values. Ó 2011 Elsevier Ltd and The Ergonomics Society. All rights reserved.
Keywords: Borg CR-10 Grip force Hand exertion Subjective rating
1. Introduction Forceful exertion has been considered one of the major contributing factors to the occurrence of upper extremity workrelated musculoskeletal disorders (Silverstein et al., 1986; Moore and Garg, 1994). Electromyography (EMG) has been adopted in measuring muscular exertions when specific muscle groups are identified in performing various tasks. Load cells and force sensors, mounted on handles to capture the force signals, may also be used to measure muscular effort especially in push and pull tasks. Both the EMG and load cells are, however, seldom used by practitioners in field measurements because preparation and post data processing are time consuming. Forceful hand exertions may also be measured indirectly using either a dynamometer or a strain gauge. The force-matching method has been proposed to measure hand forces (Wiktorin et al., 1996; Casey et al., 2002; Bao and Silverstein, 2005). This method required a subject to grip a dynamometer immediately after a task been performed to mimic the amount of force exerted. The reading on the dynamometer was recorded as the grip force applied during the task. Large between-subject variation, however, has been reported using this method (Koppelaar and Wells, 2005; Bao et al., 2009).
* Corresponding author. Tel.: þ886 3 518 6583; fax: þ886 3 518 6575. E-mail address: [email protected] (K.W. Li).
Subjective rating is one of the alternatives to measure forceful hand exertion. This approach is probably the most cost-effective and easy way to obtain force exertion data especially in large population studies (Hjelm et al., 1995; Bao et al., 2009). Borg’s concept of perceived exertion (Borg, 1970) has been adopted in quantifying subjective exertion. His ratings of perceived exertion (RPE) scale (ranged from 6 to 20) have been widely used in analyzing various manual tasks and physical works as a supplementary measure in addition to physiological measures (Wang et al., 2000; Chung and Wang, 2001; Li et al., 2007, 2009). The RPE scale has also been used in determining the perceived weight handled (Deeb, 1999). In addition to the RPE scale, Borg (1982) has constructed a category ratio scale, called CR-10 scale. This scale, ranged from 0 (nothing at all) to 10 (extremely strong or almost maximal), provides direct level estimates for determinations of ratio relationships between perceptual responses (Borg, 1990). Based on this scale, Buchholz et al. (2008) examined the subjective ratings of upper extremity exposures, wrist and shoulder postures, and grip force for automobile workers. They found significant correlation between subjective rating of manual effort and grip force. They concluded that worker’s self-report is an alternative to more resource-intensive exposure assessment methods. Park and Yun (1996) conducted a study to estimate hand force exertion using the CR-10 rating. The correlation coefficients between the measured force and perceived exertion for each of their trial ranged from 0.8 to 0.99.
0003-6870/$ e see front matter Ó 2011 Elsevier Ltd and The Ergonomics Society. All rights reserved. doi:10.1016/j.apergo.2011.03.001
930
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
Spielholz (2006) conducted a study to evaluate the efficacy of calibrating subjective worker ratings of hand exertions. The subjects applied power grip forces corresponding to their perceptions of different Borg CR-10 scale levels using the so-called “gripto-scale” procedure. The data were used to define relationships between subjective ratings and actual force application. Gripping tasks were performed and corresponding subjective hand force ratings were calibrated using the grip-to-scale calibration data. The estimation error for power grip task was found significantly reduced using this calibration procedure. Spielholz’s (2006) study indicated that calibration of hand force ratings using the grip-toscale procedure improved the accuracy of hand exertion measurements using the CR-10 scale. The Borg’s CR-10 rating has been discussed in the literature. But whether usage of this rating scale in quantifying power grip force is affected by dominant or non-dominant hand has not been discussed. Neither have the effects of arm posture been studied. The objectives of this study were to i) test the hypothesis that the exertions of power handgrip at the same perceived exertion levels between dominant hand and non-dominant hand are different; ii) test the hypothesis that hand posture affects the exertion of power grip measured at the same perceived exertion levels; and iii) compare the subjective ratings with the estimations based on regression models established using the actual CR-10 rating and power grip force data. 2. Methods This study was conducted in the laboratory.
right-handed. They were requested to refrain from physical activities at least 1 h before joined the experiment. 2.2. Apparatus A TAKEIÒ 5001 hand dynamometer was used to measure the grip force of the hand. The dynamometer was calibrated by the supplier before the experiment. A PEAKLIEÒ FT515 handgrip to build up hand/arm muscles was also adopted. A Borg CR-10 scale (Borg, 1982) was used for subjective rating. 2.3. Hand/posture conditions The hand exertions were measured under two hand conditions and two posture conditions. The measurements were conducted using either dominant hand or non-dominant hand. The postures included 90 and 180 at the elbow (see Fig. 1). 2.4. Grip force measurements Before the study, the subjects were verbally instructed the nature and use of the CR-10 scales. The grip-to-scale calibrating procedure proposed by Spielholz (2006) was adopted. In this procedure, each subject gripped the dynamometer to one of the four CR-10 levels: 2 (weak), 5 (strong), 7 (very strong) and 10 (extremely strong). The grip span of the dynamometer was 5 cm. The power grip sustained for 4 s and the peak value of the grip force was recorded. Each subject took a break for 5 min or more after he finished one grip measurement so as to avoid the effects of fatigue. Each subject performed no more than four grips at the same day.
2.1. Subjects
2.5. Subjective ratings for a pre-determined hand force
Twenty male college students were recruited as human subjects. Based on self-reports, all of them were free from musculoskeletal injuries. Their age, stature, and body weight were 22.1 (2.5) yrs, 172.3 (5.3) cm, and 69.8 (12.5) kg, respectively. All the subjects received payment and had signed informed consent for their participation in the study. All the subjects, except one, were
The handgrip was adopted as a basis to examine the CR-10 rating for the subjects under different experimental conditions. The subjects were requested to apply a 10 kgf force on the handgrip without knowing the force level (see Fig. 2). The grip span was approximately 4 cm. Each subject gripped and sustained for 4 s. He, then, reported a CR-10 rating of the hand exertion.
Fig. 1. Hand posture at (a) 90 and (b) 180 .
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
Fig. 2. Hand grip applying a 10 kgf force.
2.6. Experiment design & data analysis The grip force measurement experiment was conducted using a three-factor completely randomized design. The factors were the level of exertion, hand used, and posture. Four samples were collected for each treatment. For each subject, the order of the experiment was randomized among all the conditions. The subject completed four trials for the same treatment in the same visit. Experimental errors were assumed to be evenly distributed among all treatments under this randomization arrangement. A total of 1280 (4 CR-10 levels 2 hands 2 postures 4 replicates 20 subjects) measures were collected. Both the descriptive statistical analysis and analysis of variance (ANOVA) were performed. Duncan’s multiple range test was conducted if a factor was found statistically significant at a ¼ 0.05 level. In addition, linear regression equations without intercept were built to describe the relationship between the CR-10 rating and grip force under each hand/posture condition for each subject:
CR 10 rating ¼ b grip force þ 3i
(1)
where b is the regression coefficient and 3i is a random error of the ith trial which is normally distributed with a mean of 0 and a variance of s2. In the handgrip test, the CR-10 ratings after applying a 10 kgf power grip force were collected under a two-factor completely randomized design with one trial in each treatment. A total of 80 measures were collected (2 hands 2 postures 20 subjects). These CR-10 ratings were compared with the estimated rating calculated using Eq. (1). Analyses of variance were performed for the actual value, estimated value, and estimation error of the CR-10 rating. The statistical analyses were performed using the SPSSÒ 12.0 computer software.
931
multiple range test results comparing the two postures showed that the grip force at the 180 posture (28.3 kgf) was significantly (p < 0.05) higher than that at the 90 posture (26.5 kgf). Dominant hand (28.0 kgf) showed a significantly (p < 0.05) higher grip force than that of the non-dominant hand (26.7 kgf). The interaction effects of the CR-10 level and the hand used were significant (p < 0.0001). The interaction effects of the CR-10 level and the posture were also significant (p < 0.0001). The interactions of these effects are shown in Figs. 3 and 4, respectively. The overall Pearson’s correlation coefficient between the CR-10 rating and the grip force was 0.92 (p < 0.0001). A total of 80 linear regression models were established for each hand-posture condition for each subject. All the regression models were statistically significant at p < 0.0001 with coefficients of determination (R2) of 0.96 or higher. All the hypotheses testing whether the regression coefficient was zero were rejected (p < 0.0001). Table 1 shows the means and standard deviations of the estimated regression coefficients. In the handgrip test, the ANOVA results showed that the effects of both the hand and posture on the reported subjective rating were not statistically significant. For the estimated rating, however, the effects of both the hand and posture were significant. Nondominant hand (2.36) showed significant (p < 0.05) higher values than those of dominant hand (2.22). The estimated subjective ratings at 90 posture (2.37) were significantly (p < 0.01) higher than those at the 180 posture (2.20). The effects of hand and posture on the estimation error were not statistically significant. Table 2 shows the means and standard deviations of actual values, estimated values, and estimation errors of the subjective rating. 4. Discussions The 2, 5, 7, and 10 on the CR-10 scale were assumed to represent 20%, 50%, 70%, and 100% of perceived maximum voluntary contraction (MVC), respectively. The results of the study indicated a strong linear correlation between the CR-10 rating and grip force. This was consistent with the findings in the literature (Borg, 1990; Park and Yun, 1996). The interaction effects between exertion level and hand used in Fig. 3 showed that at low CR-10 ratings such as 2 and 5, the grip forces between dominant and non-dominant hands were negligible. At level 10, however, the grip force of dominant hand (43.7 kgf) was higher than that of the non-dominant hand (40.1 kgf). This was consistent with the 10% rule that dominant hand possesses a 10% greater grip strength than non-dominant hand (Petersen et al., 1989). It should be noted that, however, the grip forces at level 10 for the two hands were lower than the grip strengths of a similar group reported by Ho et al. (2000). This might due to the fact that level 10 on the CR-10 scale represents an “almost maximal” force exertion but not the maximal. When divided the grip forces at levels of 2, 5, and 7 by the force at level 10, the grip forces were converted into %MVC. The %MVC at levels of 2, 5, and 7 for dominant hand were 24.7%, 56.2%, and
3. Results The ANOVA results showed that the main effects of the CR-10 level, hand used, and posture on power grip force were all significant (p < 0.0001). The Duncan’s multiple range test results showed that the mean grip force at the level 10 of the CR-10 (41.9 kgf) was significantly (p < 0.05) higher than those at levels 7 (32.6 kgf), 5 (24.4 kgf), and 2 (10.5 kgf). The grip force at level 7 was significantly (p < 0.05) higher than those at levels 5 and 2. The grip force at level 5 was significantly higher than that at the level 2. Duncan’s
Fig. 3. Grip force at different CR-10 ratings for dominant and non-dominant hands.
932
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933 Table 2 Means and standard deviations of actual values, estimated values, and estimation errors of the CR-10 rating. Actual CR-10 rating
Fig. 4. Grip force at different CR-10 ratings under two postures.
75.6%, respectively. The %MVC values at the three levels for nondominant hand were 25.6%, 60.5%, and 80.0%, respectively. All the % MVC values were higher than their corresponding CR-10 levels times 10. This implies that the subjects applied a force exertion higher than they perceived. Comparing the %MVC between the two hands at the 2, 5, and 7 levels, non-dominant hands showed higher values than those of the dominant hand at all three levels. A possible explanation of this phenomenon was, as mentioned in the previous paragraph, that the level 10 did not correspond to the maximal but the sub-maximal of the force exertion. The interaction effects between the exertion level and posture in Fig. 4 showed similar trends as in Fig. 3. The grip forces between the two postures at low exertion levels such as 2 and 5 were negligible. At level 10, the grip forces at 180 posture were higher than those at the 90 . The %MVC at levels of 2, 5, and 7 for the 90 posture were 26.1%, 59.2%, and 79.0%, respectively. The %MVC values at the three levels for the 180 posture were 24.3%, 57.4%, and 76.6%, respectively. All the %MVC values were higher than their corresponding CR-10 levels times 10. The %MVC values for the 90 posture were higher than those of the 180 posture for all the 2, 5, and 7 levels. This might be attributed to the fact that, as compared to the 180 posture, the 90 posture was associated with a flexion of the arm which required higher muscular exertion. Buchholz et al. (2008) examined the agreement of subjective ratings of upper extremity exposures with corresponding direct measurements obtained from workers. Psychophysical ratings of exposure based on the CR-10 were obtained for the period of time in which direct measurements were acquired. Ratings of manual effort were significantly correlated with directly measured grip force. Grant et al. (1994) reported that the perceived exertion, based on the CR-10, showed strong and positive correlation with grip force. These were consistent with the findings our study. It might be noteworthy that the correlation coefficient between the subjective rating and the grip force in the current study (0.92) was higher than those (0.33e0.72) reported by Grant et al. (1994). This difference might be attributed to the force ranges in the two studies. The force range in Grant et al. (1994) was more restricted (1%e46% MVC) than the one (0%e100% MVC) adopted in the current study. The regression coefficients in Table 1 indicate the increase of the CR-10 rating due to an increase of grip force depend on the hand used and posture. With each unit (kgf) increase in grip force, the CR-10 ratings at the 90 posture increased 0.24 and 0.23 for non-dominant and dominant hands, respectively. At the 180 posture, the increases were 0.23 and 0.21 for non-dominant and dominant hands, respectively. A pair-wised t-test results comparing the regression coefficients showed that the difference between dominant and nonTable 1 Means and standard deviations of estimated regression coefficients. Posture
Non-dominant hand
Dominant hand
90 180
0.24 (0.03) 0.23 (0.03)
0.23 (0.03) 0.21 (0.02)
Estimated CR-10 rating
Estimation error
Non-dominant hand 3.93 (1.78) 90 180 3.83 (1.77)
2.43 (0.03) 2.29 (0.28)
1.50 (1.73)** 1.54 (1.71)**
Dominant hand 90 3.50 (1.57) 180 3.20 (1.61)
2.32 (0.30) 2.12 (0.23)
1.18 (1.51)* 1.08 (1.54)*
Note: Estimation error ¼ actual CR-10 rating estimated CR-10 rating; *p < 0.01, **p < 0.001.
dominant hands was statistically significant (p < 0.001). In Table 1, lower increases in the CR-10 rating associated with an increase in grip force were also observed for the 180 posture than those of the 90 posture. A pair-wised t-test results comparing the regression coefficients also showed that the difference between the two postures was statistically significant (p < 0.001). Table 2 shows the actual ratings, estimated ratings, and the differences between the two for applying a grip force of 10 kgf. The subjects reported slightly higher CR-10 ratings when using their non-dominant hands than those of dominant hands even though the difference was not statistically significant. This is reasonable as the non-dominant hands have weaker muscles and need to apply a higher exertion so as to generate the same amount of force as compared to the dominant hands. The subjects also reported slightly higher CR-10 ratings when applying a force at the 90 posture than at the 180 posture even without statistical significance. The R2 represents the percentage of variation of the dependent variable which could be explained by the independent variable. An R2 of 0.96 or higher indicates that almost all the variation of the subjective rating may be explained by the grip force. It was shown in Table 2 that the estimated CR-10 ratings were significantly (p < 0.01) lower than those of the actual values under all the hand/ posture conditions. For non-dominant hand, the estimation errors for the 90 and 180 postures were 1.50 (38.1%) and 1.54 (40.2%), respectively. For dominant hand, the estimation errors for the two postures were 1.18 (33.7%) and 1.08 (33.7%), respectively. The subjects reported subjective scores higher than those estimations based on their own perception and grip force data even though the corresponding regression models had very high R2 values. The results of the current study indicated that the CR-10 rating scale has high correlation with the power grip force. This was consistent with those in the literature (Grant et al., 1994; Park and Yun, 1996; Buchholz et al., 2008). However, perceived exertions of power grip might be affected by both the hand used and arm posture. In addition, the subjects had a tendency to report the subjective ratings of power grip exertion higher than the actual % MVC values. 5. Conclusion An experiment was conducted to test the grip force of male subjects at four perceived exertion levels on the CR-10 scale under handedness and postural conditions. Generally speaking, the subjects tended to apply a higher power grip force (%MVC) than they perceived at levels 2, 5, and 7 on the CR-10 scale. It was found that the grip forces between dominant and non-dominant hands at low CR-10 levels were negligible. The grip forces were significantly different between the two hands at level 10. Similar results were found for the arm posture conditions. The grip forces between the two postures at low CR-10 levels were negligible. At level 10, the grip forces at 180 posture were significantly higher than those at
K.W. Li, R. Yu / Applied Ergonomics 42 (2011) 929e933
the 90 . The overall correlation coefficient between the CR-10 rating and the grip force was significant (r ¼ 0.92; p < 0.0001). Eighty regression models have been established between the CR-10 rating and the grip force. The subjects reported higher subjective ratings than the estimations which were based on their own perception and grip force data, even though the regression models had very high R2 values. Acknowledgments This research was financially support by the National Science Council of the ROC under grant NSC 98-2221-E-216-013. The author thanks Shang Wei Tseng and Tsung-Yu Pei, graduate students with the Department of Industrial Management at Chung-Hua University, for their assistance in the study. References Bao, S., Silverstein, B., 2005. Estimation of hand force in ergonomic job evaluations. Ergonomics 48, 288e301. Bao, S., Spielholz, P., Howard, N., Silverstein, B., 2009. Force measurement in field ergonomic research and application. Int. J. Ind. Ergon. 39, 333e340. Borg, G., 1970. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med. 2 (3), 92e98. Borg, G., 1982. Psychophysical bases of perceived exertion. Med. Sci. Sports. Exerc. 14 (5), 377e381. Borg, G., 1990. Psychophysical scaling with applications in physical work and the perception of exertion. Scand. J. Work. Environ. Health. 16 (10), 55e58. Buchholz, B., Park, J.-S., Gold, J.E., Punnett, L., 2008. Subjective ratings of upper extremity exposures: inter-method agreement with direct measurement of exposures. Ergonomics 51 (7), 1064e1077.
933
Casey, J.S., McGorry, R.W., Dempsey, P.G., 2002. Getting a grip on grip force estimates: a valuable tool for ergonomic evaluations. Prof. Saf., 18e24. Chung, H.C., Wang, M.J., 2001. The effects of container design and stair climbing on maximal acceptable lift weight, wrist posture, psychophysical, and physiological responses in wafer-handling tasks. Appl. Ergon. 32, 593e598. Deeb, J.M., 1999. Muscular fatigue and its effects on weight perception. Int. J. Ind. Ergon. 24, 223e233. Grant, K.A., Habes, D.J., Putz-Anderson, V., 1994. Psychophysical and EMG correlates of force exertion in manual work. Int. J. Ind. Ergon. 13, 31e39. Hjelm, E.W., Winkel, J., Nygard, C.H., Wiktorin, C., Karlqvist, L., 1995. Can cardiovascular load in ergonomic epidemiology be estimated by self-report? Stockholm MUSIC 1 study group. J. Occup. Environ. Med. 37, 1210e1217. Ho, R.W.D., Chang, S.Y., Wang, C.W., Hwang, M.H., 2000. Grip and key pinch strength: norms for 15- to 22-year-old Chinese students. Chin. Med. J. 63 (1), 21e27. Koppelaar, E., Wells, R., 2005. Comparison of measurement methods for quantifying hand force. Ergonomics 48, 983e1007. Li, K.W., Yu, R-f., Han, X.L., 2007. Physiological and psychophysical responses in handling maximum acceptable weights under different footweardfloor friction conditions. Appl. Ergon. 38, 259e265. Li, K.W., Yu, R-f., Gao, Y., Maikala, R.V., Tsai, H.-H., 2009. Chinese workers performing combined manual materials handling tasks. Int. J. Ind. Ergon. 39, 422e427. Moore, J.S., Garg, A., 1994. Upper extremity disorders in a pork plant: relationship between task factors and morbidity. Am. Ind. Hyg. Assoc. J. 55, 703e715. Park, D., Yun, M.H., 1996. An application of psychophysical ratings to external force estimation. Comput. Ind. Engng. 31, 675e680. Petersen, P., Petrick, M., Connor, H., Conklin, D., 1989. Grip strength and hand dominance: challenging the 10% rule. Am. J. Occup. Ther. 43 (7), 444e447. Silverstein, B.A., Fine, L.J., Armstrong, T.J., 1986. Hand wrist cumulative trauma disorders in industry. Br. J. Ind. Med. 43, 779e784. Spielholz, P., 2006. Calibrating Borg scale ratings of hand force exertion. Appl. Ergon. 37, 615e618. Wang, M.J., Chung, H.C., Chen, H.C., 2000. The effect of handle angle on MAWL, wrist, posture, RPE, and heart rate. Hum. Factors. 42 (4), 553e565. Wiktorin, C., Selin, K., Ekenvall, L., Kilbom, A., Alfredsson, L., 1996. Evaluation of perceived and self-reported manual forces exerted in occupational materials handling. Appl. Ergon. 27, 231e239.