This article was downloaded by: [UQ Library] On: 21 November 2014, At: 10:33 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
Discriminability of load heaviness: implications for the psychophysical approach to manual lifting a
a
b
W. KARWOWSKI , C. SHUMATE , J. W. YATES & N. PONGPATANA
a
a
Center for Industrial Ergonomics , University of Louisville , Louisville, Kentucky, 40292, USA b
Exercise Physiology Laboratory , University of Louisville , Louisville, Kentucky, 40292, USA Published online: 31 May 2007.
To cite this article: W. KARWOWSKI , C. SHUMATE , J. W. YATES & N. PONGPATANA (1992) Discriminability of load heaviness: implications for the psychophysical approach to manual lifting, Ergonomics, 35:7-8, 729-744, DOI: 10.1080/00140139208967359 To link to this article: http://dx.doi.org/10.1080/00140139208967359
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Discriminability of load heaviness: implications for the psychophysical approach to' manual lifting W. KARWOWSW,~ C.SHUMATE,' 1. W.Y A ~ E Sand ,~
N. PONGPATANA~
'Center for Industrial Ergonomics 'Exercise Physiology Laboratory University of Louisville Louisville, Kentucky 40292, USA
Downloaded by [UQ Library] at 10:34 21 November 2014
Keywords: Weight discriminability; Load heaviness; Psychophysics; Manual lifting.
'
The main objective of this study was to investigate human ability to discriminate between different levels of load heaviness in manual lifting. Twelve male college students participated in the laboratory experiment. Twen t y-eigh t sequences of five boxes that weighed from 5 to 64 Ibs (2-27-29-1 kg) were used. The subjects were asked to arrange boxes in each sequence in order of the perceived (increasing or decreasing) heaviness, i.e., from lightest to the heaviest box, or from heaviest to the lightest box. The subjects were also asked to assign linguistic descriptors of perceived load heaviness to each box in the sequence, and to indicate the confidence levels regarding correctness of the assigned box order and assignment of linguistic values. The independent variables included magnitude of weight and load differential between the successive weights in a sequence. The number of sequential ordering errors, assignmen1 of linguistic variables, and estimated confidence levels were highly dependent on the load differential and weight range. It was concluded that in order to assure reliable results of the psychophysical approach to determining the values of maximum acceptable weight of lift, the . adjustment process for male subjects should require using small weights of at least 4 Ibs ( 1 43 kg) to be added or removed from the lifted box. The r e d ts of this study also suggest that the error rate in load discriminability can be controlled below the 10% level, if the relative difference in weight between successive boxes lifted is at least 12%. Given the above findings, it is suggested that usefulness of some of the recommendations for setting safe limits for manual lifting tasks, which were reached based on the psychophysical approach and broadly reported in the past, may need to be carefully re-examined. Finally, this study showed that the Weber Jraction for load heaviness over the range of lifted weights from 8.6 to 29. i kg is between 0-03and 0.04.
1. Introduction
The psychophysical approach to determining the maximum acceptable weight of lift (MAWL) has been extensively used over the past 25 years (Snook and Irvine 1967, Ayoub 1976, Snook 1985). This approach aims to quantify human lifting capacity based on subjective perception of exertion, under the assumption that workers are able to determine with some accuracy I he highest acceptable workload (Gamberale 1985): The experimental procedure for the psychophysical approach to manual lifting proposed by Snook (1978), requires the subjects to imagine working on the incentive basis, as hard as they can (lifting as much as they can), without straining themselves or becoming unusually tired, weak, overheated, or out-of-breath. Typically, the subjects are given control of the weight (or force) handled, and are asked to adjust the weight of the box (force) over a period of 30-40min, up to the maximum level they are willing to accept for an 8 h shift at given frequency of task repetition. 0014-0 139192 53.00 8 1992 Taylor & Francis Ltd.
Downloaded by [UQ Library] at 10:34 21 November 2014
730
W. Kanvowski et al.
Although the psychophysical approach proved very useful in the past efforts to establish safe limits for manual handling of loads (Ayoub et al. 1980, Troup and Edwards 1 985, Millar 1988, Genaidy et al. 1990, NIOSH 198 1, Putz-Anderson and Waters 1991), its practical application should be critically examined in view of several shortcomings reported in the literature (Gamberale et al. 1987, Troup et al. 1987, Garg 1989). Ciriello and Snook (1983) showed that the psychophysical technique leads to an overestimation of the maximum acceptable weights and forces for tasks with very high frequencies. Mital(1983) reported that at the end of the 8 h, females were lifting 85% of the weight which they initially selected during the psychophysical experiment, while males lifted only 65Oh of the original MA WL values. Karwowski and Yates (1986) concluded that the psychophysical approach should not be used to set lifting standards for frequencies higher than 6 liftslmin. Recently, Ciriello et a/. (1990) studied the effects of task duration on the maximum acceptable weights and forces. They concluded that the psychophysical methodology is suitable for determining maximum acceptable weights for task frequencies of 4.3 liftslmin or slower. According to Gamberale (1 98 5), the perception of exertion during physical work nor only has psychological validity, but it also reflects real conditions such as the interplay between the requirements of the job and the capacity of the individual. Unfortunately, the human decision process underlying selection of a given weight of load as the maximum acceptable one for an 8 h day has not yet been clearly explained (Karwowski and Ayoub 1984, Karwowski 1988). An important, but mainly overlooked consideration in such a process, is the human perception of load heaviness, and human ability to successfully discriminate between different levels of load heaviness during the experimental adjustment procedure. As reported by OSHA (1982), among the contributory factors perceived by industrial workers who sustained back injuries while lifting were: ( I ) too heavy objects (reported by 36% of workers); and (2) underestimating the weight of objects before lifting (claimed by 14% of workers). Karwowski (1 988) conducted a study to investigate subjective perceptions of load heaviness by male subjects. Seven linguistic categories of load heaviness, i.e., very light, light, less-than-medium, medium, more-than-medium, heavy, and very heavy, were used in three laboratory experiments: (1) classification of the pre-weighted boxes using the linguistic values of load heaviness; (2) selection of weights that correspond to a given linguistic category of load heaviness; and (3) determination of the MAWL values using the psychophysical methodology. Subjects were asked to lift unmarked boxes of different weights, ranging from 4-5 kg to 27.3 kg. The majority of males selected MAWL values, which were independently judged by them as either very heavy or heavy, to be lifted infrequently during an 8 h shift. Kanvowski ( 1 99 1) also studied subjective perception of load heaviness by female college students. The loads selected as maximum acceptable weights for an 8 h shift were independently judged by the subjects as being in the more-than-moderateor heavy weight categories. Comparison of the results for females and males led to a conclusion that female subjects were more realistic, with respect to the perceived load heaviness, in selecting . the MAWL values, than were the male subjects. 2.
Background
As pointed out by Carlson et al. (1 977), although the measurement of human ability
to discriminate between lifted weights has a long history, there is a lack of information
Downloaded by [UQ Library] at 10:34 21 November 2014
Discriminability of load heaviness
73 1
about human abilities in discrimination of loads which are typically lifted in industry. An important consideration in the experimental weight adjustment process for determining MAWL values based on the psychophysical approach, is the magnitude of a constant weight that represents a proportion by which a standard stimulus must be different from a standard weight in order to discriminate between the two. Unfortunately, even today this issue still remains unexplored (Carlson el a/. 19 77). Previous research concent rated on the concept of the just noticeable differeence (jnd), or a threshold weight value, defined as the amount of difference between a stimulus weight and a standard weight when the difference between these two weights can be noticed 50% of the time. This threshold value divided by the stimulus weight gives a weight fraction called the Weber fraction. Weber ( 1 834) postulated that the size of jnd increases linearly with the size of the standard weight. In other words, larger standard weights require greater weight differences between them and in order to notice a difference 50% of the time. the ~eight,~stimulus The value of a Weber fraction for load heaviness was estimated b y Teghtsoonian ( 1 97 1) to be 0.02. The average Weber fraction for weights below 3 kg was reported to be about 0.088 (Woodworth and Schlosber 1965). Carlson et al. (1977) used the psychophysical methods of constants to measure the difference threshold for lifted weights in the range from 10 to 20kg, and reported an average Weber fraction of about 0.04. The Weber fraction values for the three standard weights of 10, 15, and 20kg were 0-052, 0.034, and 0.043, respectively. Recently, Wang et al. (1991) reported a Weber fraction of about 0.04 for two standard weights of 8.2 and 20 kg. In one of the early studies dealing with human perception of load heaviness, Ljundberg et al. (1982) used a standard stimulus based on the subjects' selection of acceptable weight, and applied the ratio-rating method to study the perceived workload for horizontal lifting of loads weighing between 67% and 200% of the acceptable weight. This method of acceptability scaling was also adopted by Baxter et a!. (1986), who used it to quantify the Steven's power function as a model of perceived load heaviness in vertical lifting. Baxter el al. ( 1986) have shown the importance of perceived load heaviness on thc selected maximum acceptable weight of lift. Karwowski er a!. ( I 989) based on the results of a pilot study concluded that the human ability to discriminate weight differences between lifted boxes is severely impaired as the load difTerentia1 decreases below 4 1bs ( 1-8kg). Recently, Hamrick (1 990) compared major muscle trunk activity during lifting tasks at the weight level below, at, and above the MAWL values. He concluded that the relationship- between changes in muscle activity, relative to external load, is non-linear, and there seems to exist critical biomechanical thresholds at which the subject judges the load unacceptable. Hamrick (1990) also suggested that the use of psychophysical approach to determine the maximum acceptable weight of lift may result in different perception of effort when adjusting the box weight. Specifically, the change in perceived effort appears to be greater when reducing the box weight which is above the MAWL level, than when increasing the box weight which is below the MAWL level. Another important factor that affects the magnitude estimation of weight is the order by which the weights are presented and then judged. The phenomenon of psychophysical hysteresis proposed by Stevens (1 975), indicates that if the standard weights are presented in ascending order in terms of weight magnitude, the judgement of the associated stimulus weight will be larger, than if the standard weights are presented in descending order. Bevan et al. ( 1963) concluded that the
W. Kanvowski et al.
732
Downloaded by [UQ Library] at 10:34 21 November 2014
judgement of a stimulus weight with respect to standard weights is affected by whether or not a given standard weight is greater or smaller than the standard weight it precedes o r follows.
3. Decision-based model for determining MA WL values Detection of information about load heaviness is one of the main cognitive functions of the subject involved in weight adjustment process when seeking to establish MAWL values based on the psychophysical approach. Holland and Lockhead ( 1 968), and Ward and Lockhead (1 97 1) showed that the judgement of weight in a current trial is influenced by a subject's feedback from judgement of weight from the previous trial. Therefore, an important issue in assessment of acceptable load is the ability of the subject to discriminate between consecutive weights of the box lifted, especially under the time pressure of high task repetition. This unexplored issue is very important in view of the complexity of the human decision-making process during the weight adjustment period, and its potential consequences for the quality of MAWL values derived on the basis of psychophysical methodology. It can be postulated that inability of the subject to identify a difference between two weights during the. act of load adjustment will most likely lead to the next adjustment. In other words, if the magnitude of the first and consecutive adjustment(s) is not sufficient for the subject to detect the weight difference (called here the load heaviness discriminability), he/she will continue the adjustment process. In such a case, while the absolute magnitude of weight changes may be significant, it will never be perceived as such by the subject who tends to compare bet ween weights of the successive boxes lifted. This process can be conceptualized through the decision-based model illustrated in figure 1. Decision making refers here to the process whereby subjects evaluate information about load heaviness which is available through the initial perceptual processing (Salvendy and Knight 1987).
I gure 1.
AATlNO OF ACCEPTABLE WORKLOAD
I
+
A decision model-based approach to determining the maximum acceptable weight
of lift (MAWL).
According to the proposed decision-based approach to determining the maximum acceptable weight of lift, failure to discriminate the change in weight through perceived box heaviness at time ti, may lead to the subsequent weight adjustment(s), and, consequently, potential overestimation of the true MAWL values. Such
Discriminability of load heaviness
733
Downloaded by [UQ Library] at 10:34 21 November 2014
overestimation of the MAWL values within limited time exposure to selected workloads and under time pressured decision-making has been well documented for higher task repetitions (Ciriello and Snook 1 983, Karwowski and Yates 1986). 4. Objectives Subjective perception of load heaviness plays an important role in the study of manual handling capacities, and should be considered when establishing guidelines for safe lifting (Karwowski 1988, Karwowski 1991). Previous studies in this area concentrated on determination of values of the Weber fraction (Carlson et a/. 1977), which defines the human ability to successfully discriminate between weights only 50% of the time. From the psychophysical approach to manual lifting point of view, however, it is very important to establish much higher levels of load heaviness discriminability, for example at 90 or 95% level. This would increase reliability of the selected MAWL values, and improve the confidence in results of the psychophysical studies in manual material handling. Therefore, the main objective of this study was to investigate human ability to discriminate between weights of lifted loads within the range from 2-4 to 29.1 kg, as the function of their absolute load magnitude, and the weight difference between two successive boxes in a sequence (called here the
load differential). 5. Pmedures 5.1 . Subjects . Twelve male college students, who were compensated for their time, took part in the laboratory study. Each subject was required to fill out a personal consent form and a data form. Only those subjects that were free of any medical problems were allowed to participate. The subjects were asked to wear comfortable clothes, stretch before the experimentation to minimize any potential for injury, and refrain from any strenuous exercises prior to the experimental sessions. Age and anthropometric characteristics of the subjects are shown in table 1. The isometric strength measurements were made according to the procedures proposed by Chaffin (1 975). The isokinetic strength tesls were performed according t o procedures described by Pytel and Kamon ( 198 1 ).
5 . 2 . Experimental design Within subjects, randomized complete block design was used, with subjects as blocks and the load differential as an independent variable. Each of the 28 box sequences consisted of five compact boxes (36 x 36 x 34 cm) weighing from 5 to 64 Ibs (2.3 to 29.1 kg). This range of weights constituted a range of loads that is common to many material handling tasks performed in industry (OSHA 1982). Within each sequence of weights, there was a constant load differential of either 1, 2, 4, or 8 Ibs (0-45,0.9 1, . 1.6, or 3.63 kg). Table 2 illustrates the distribution of weights in each sequence as a function of load differential. The dependent measures were the number of sequential box ordering errors with respect to perceived load heaviness, subject confidence leveis in the correctness of box sequence order, and confidence levels in assignment of linguistic descriptors to each of the boxes.
5.3. Methods All of the boxes were identical in appearance, and a cover was placed over the weights to prevent subjects from seeing the box content. The boxes were placed in
W.Karwowski et al.
734 Table I .
Age, physical characteristics and static and dynamics strengths (kg) of the subjects (n= 12).
Mean
Variable
SD
Range
Downloaded by [UQ Library] at 10:34 21 November 2014
Age (years)
Body weight (kg) Stature (cm) Acromial height (cm) Arm length (cm) Forearm length (cm) Static arm strength (kg) Static back strength Static composite strength Static leg strength Dynamic arm strength Dynamic shoulder strength Dynamic back strength Dynamic lift strength
Table 2. Arrangements for experimental box ordering sequences.
Sequence number
Load differential
Box
weight
8, 16, 32, 48, 6, 16, 24, 32, 40, 48,
12, 20, 36, 52, 8, 18, 26, 34, 42, 50, 56, 58,
16, 20, 24, 28, 40, 44, 56, 60, 10, 12, 20, 22, 28, 30, 36, 38, 44, 46, 52, 54, 60, 62,
24 32 48 64 14 24 32 40 48 56 64
Number of Total number sequences of trials
4
4
240
2
7
420
Total
28
1680
5, 6, 7, 8, 9 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, 52,
9, 13, 17, 21, 25, 29, 33, 37, 4 1, 45, 49, 53, 56, 57, 60, 6 1,
10, 11, 12
14, 18, 22, 26,
30, 34, 38, 42, 46, 50, 54, 58, 62,
15, 19, 23, 27, 31, 35, 39, 43, 47, 5 1, 55, 59, 63,
16 20 24 28 32 36 40 44 48 52 56 60 64
Downloaded by [UQ Library] at 10:34 21 November 2014
Discriminability of load heaviness random order in marked areas on the floor that were evenly spaced apart. Subjects were first asked to lift each box once to about knee height in order to get an idea of its heaviness. Afterwards subjects were allowed to lift the boxes as many times as they wanted and in any order. Subjects were instructed to lift each box and put it on a table (89 cm high) in the marked areas, either in descending or ascending order of weight. The subjects were also encouraged to compare two boxes at a time, and to rearrange the order of the boxes according to their perceived heaviness at any time. No time restriction or task repetition requirements were made. Upon completion of the ordering procedure, subjects were presented a chart with the following linguistic terms for load heaviness: very light, light, less-than-medium, medium, more-than-medium, heavy, very heavy, and extremely heavy (cannot lift), and asked to assign a linguistic value to each of the boxes. Subjects were told that the value they assign should describe how they feel about heaviness of each load (box) lifted. Before assigning the linguistic values of heaviness, subjects were asked to lift each box at least once again. This was done in order to minimize the chance of assigning a range of weights and linguistic values instead of judging them separately. Subjects were also asked to assess the level of confidence regarding the correct order of box sequence, and the confidence in assigned load heaviness values. A lOcm bipolar scale, marked 'completely ltncertain ' and 'absolutely positive' on its left and right ends, respectively, was used for that purpose. A minimum of a 10min rest break was taken between completion of each box sequence trial.
6. Results and discussion The results were analysed using an analysis of variance (ANOVA). The Newman-Kuels test was used to test differences between means of the response variables. In addition, the Kruskal-Wallis test was used to analyze the subject level of confidence based on the assessment scale lacking the interval scaling properties. 6.1. Number of box heaviness ordering errors Overall, the subjects made a total of 567 errors out of 1680 possible box positions (resulting in the 33.75% error rate) for all ordering sequences. An average error rate was less than two mistakes per sequence. The analysis o f variance revealed highly significant effects (p
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
Discriminability of load heaviness: implications for the psychophysical approach to manual lifting a
a
b
W. KARWOWSKI , C. SHUMATE , J. W. YATES & N. PONGPATANA
a
a
Center for Industrial Ergonomics , University of Louisville , Louisville, Kentucky, 40292, USA b
Exercise Physiology Laboratory , University of Louisville , Louisville, Kentucky, 40292, USA Published online: 31 May 2007.
To cite this article: W. KARWOWSKI , C. SHUMATE , J. W. YATES & N. PONGPATANA (1992) Discriminability of load heaviness: implications for the psychophysical approach to manual lifting, Ergonomics, 35:7-8, 729-744, DOI: 10.1080/00140139208967359 To link to this article: http://dx.doi.org/10.1080/00140139208967359
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Discriminability of load heaviness: implications for the psychophysical approach to' manual lifting W. KARWOWSW,~ C.SHUMATE,' 1. W.Y A ~ E Sand ,~
N. PONGPATANA~
'Center for Industrial Ergonomics 'Exercise Physiology Laboratory University of Louisville Louisville, Kentucky 40292, USA
Downloaded by [UQ Library] at 10:34 21 November 2014
Keywords: Weight discriminability; Load heaviness; Psychophysics; Manual lifting.
'
The main objective of this study was to investigate human ability to discriminate between different levels of load heaviness in manual lifting. Twelve male college students participated in the laboratory experiment. Twen t y-eigh t sequences of five boxes that weighed from 5 to 64 Ibs (2-27-29-1 kg) were used. The subjects were asked to arrange boxes in each sequence in order of the perceived (increasing or decreasing) heaviness, i.e., from lightest to the heaviest box, or from heaviest to the lightest box. The subjects were also asked to assign linguistic descriptors of perceived load heaviness to each box in the sequence, and to indicate the confidence levels regarding correctness of the assigned box order and assignment of linguistic values. The independent variables included magnitude of weight and load differential between the successive weights in a sequence. The number of sequential ordering errors, assignmen1 of linguistic variables, and estimated confidence levels were highly dependent on the load differential and weight range. It was concluded that in order to assure reliable results of the psychophysical approach to determining the values of maximum acceptable weight of lift, the . adjustment process for male subjects should require using small weights of at least 4 Ibs ( 1 43 kg) to be added or removed from the lifted box. The r e d ts of this study also suggest that the error rate in load discriminability can be controlled below the 10% level, if the relative difference in weight between successive boxes lifted is at least 12%. Given the above findings, it is suggested that usefulness of some of the recommendations for setting safe limits for manual lifting tasks, which were reached based on the psychophysical approach and broadly reported in the past, may need to be carefully re-examined. Finally, this study showed that the Weber Jraction for load heaviness over the range of lifted weights from 8.6 to 29. i kg is between 0-03and 0.04.
1. Introduction
The psychophysical approach to determining the maximum acceptable weight of lift (MAWL) has been extensively used over the past 25 years (Snook and Irvine 1967, Ayoub 1976, Snook 1985). This approach aims to quantify human lifting capacity based on subjective perception of exertion, under the assumption that workers are able to determine with some accuracy I he highest acceptable workload (Gamberale 1985): The experimental procedure for the psychophysical approach to manual lifting proposed by Snook (1978), requires the subjects to imagine working on the incentive basis, as hard as they can (lifting as much as they can), without straining themselves or becoming unusually tired, weak, overheated, or out-of-breath. Typically, the subjects are given control of the weight (or force) handled, and are asked to adjust the weight of the box (force) over a period of 30-40min, up to the maximum level they are willing to accept for an 8 h shift at given frequency of task repetition. 0014-0 139192 53.00 8 1992 Taylor & Francis Ltd.
Downloaded by [UQ Library] at 10:34 21 November 2014
730
W. Kanvowski et al.
Although the psychophysical approach proved very useful in the past efforts to establish safe limits for manual handling of loads (Ayoub et al. 1980, Troup and Edwards 1 985, Millar 1988, Genaidy et al. 1990, NIOSH 198 1, Putz-Anderson and Waters 1991), its practical application should be critically examined in view of several shortcomings reported in the literature (Gamberale et al. 1987, Troup et al. 1987, Garg 1989). Ciriello and Snook (1983) showed that the psychophysical technique leads to an overestimation of the maximum acceptable weights and forces for tasks with very high frequencies. Mital(1983) reported that at the end of the 8 h, females were lifting 85% of the weight which they initially selected during the psychophysical experiment, while males lifted only 65Oh of the original MA WL values. Karwowski and Yates (1986) concluded that the psychophysical approach should not be used to set lifting standards for frequencies higher than 6 liftslmin. Recently, Ciriello et a/. (1990) studied the effects of task duration on the maximum acceptable weights and forces. They concluded that the psychophysical methodology is suitable for determining maximum acceptable weights for task frequencies of 4.3 liftslmin or slower. According to Gamberale (1 98 5), the perception of exertion during physical work nor only has psychological validity, but it also reflects real conditions such as the interplay between the requirements of the job and the capacity of the individual. Unfortunately, the human decision process underlying selection of a given weight of load as the maximum acceptable one for an 8 h day has not yet been clearly explained (Karwowski and Ayoub 1984, Karwowski 1988). An important, but mainly overlooked consideration in such a process, is the human perception of load heaviness, and human ability to successfully discriminate between different levels of load heaviness during the experimental adjustment procedure. As reported by OSHA (1982), among the contributory factors perceived by industrial workers who sustained back injuries while lifting were: ( I ) too heavy objects (reported by 36% of workers); and (2) underestimating the weight of objects before lifting (claimed by 14% of workers). Karwowski (1 988) conducted a study to investigate subjective perceptions of load heaviness by male subjects. Seven linguistic categories of load heaviness, i.e., very light, light, less-than-medium, medium, more-than-medium, heavy, and very heavy, were used in three laboratory experiments: (1) classification of the pre-weighted boxes using the linguistic values of load heaviness; (2) selection of weights that correspond to a given linguistic category of load heaviness; and (3) determination of the MAWL values using the psychophysical methodology. Subjects were asked to lift unmarked boxes of different weights, ranging from 4-5 kg to 27.3 kg. The majority of males selected MAWL values, which were independently judged by them as either very heavy or heavy, to be lifted infrequently during an 8 h shift. Kanvowski ( 1 99 1) also studied subjective perception of load heaviness by female college students. The loads selected as maximum acceptable weights for an 8 h shift were independently judged by the subjects as being in the more-than-moderateor heavy weight categories. Comparison of the results for females and males led to a conclusion that female subjects were more realistic, with respect to the perceived load heaviness, in selecting . the MAWL values, than were the male subjects. 2.
Background
As pointed out by Carlson et al. (1 977), although the measurement of human ability
to discriminate between lifted weights has a long history, there is a lack of information
Downloaded by [UQ Library] at 10:34 21 November 2014
Discriminability of load heaviness
73 1
about human abilities in discrimination of loads which are typically lifted in industry. An important consideration in the experimental weight adjustment process for determining MAWL values based on the psychophysical approach, is the magnitude of a constant weight that represents a proportion by which a standard stimulus must be different from a standard weight in order to discriminate between the two. Unfortunately, even today this issue still remains unexplored (Carlson el a/. 19 77). Previous research concent rated on the concept of the just noticeable differeence (jnd), or a threshold weight value, defined as the amount of difference between a stimulus weight and a standard weight when the difference between these two weights can be noticed 50% of the time. This threshold value divided by the stimulus weight gives a weight fraction called the Weber fraction. Weber ( 1 834) postulated that the size of jnd increases linearly with the size of the standard weight. In other words, larger standard weights require greater weight differences between them and in order to notice a difference 50% of the time. the ~eight,~stimulus The value of a Weber fraction for load heaviness was estimated b y Teghtsoonian ( 1 97 1) to be 0.02. The average Weber fraction for weights below 3 kg was reported to be about 0.088 (Woodworth and Schlosber 1965). Carlson et al. (1977) used the psychophysical methods of constants to measure the difference threshold for lifted weights in the range from 10 to 20kg, and reported an average Weber fraction of about 0.04. The Weber fraction values for the three standard weights of 10, 15, and 20kg were 0-052, 0.034, and 0.043, respectively. Recently, Wang et al. (1991) reported a Weber fraction of about 0.04 for two standard weights of 8.2 and 20 kg. In one of the early studies dealing with human perception of load heaviness, Ljundberg et al. (1982) used a standard stimulus based on the subjects' selection of acceptable weight, and applied the ratio-rating method to study the perceived workload for horizontal lifting of loads weighing between 67% and 200% of the acceptable weight. This method of acceptability scaling was also adopted by Baxter et a!. (1986), who used it to quantify the Steven's power function as a model of perceived load heaviness in vertical lifting. Baxter el al. ( 1986) have shown the importance of perceived load heaviness on thc selected maximum acceptable weight of lift. Karwowski er a!. ( I 989) based on the results of a pilot study concluded that the human ability to discriminate weight differences between lifted boxes is severely impaired as the load difTerentia1 decreases below 4 1bs ( 1-8kg). Recently, Hamrick (1 990) compared major muscle trunk activity during lifting tasks at the weight level below, at, and above the MAWL values. He concluded that the relationship- between changes in muscle activity, relative to external load, is non-linear, and there seems to exist critical biomechanical thresholds at which the subject judges the load unacceptable. Hamrick (1990) also suggested that the use of psychophysical approach to determine the maximum acceptable weight of lift may result in different perception of effort when adjusting the box weight. Specifically, the change in perceived effort appears to be greater when reducing the box weight which is above the MAWL level, than when increasing the box weight which is below the MAWL level. Another important factor that affects the magnitude estimation of weight is the order by which the weights are presented and then judged. The phenomenon of psychophysical hysteresis proposed by Stevens (1 975), indicates that if the standard weights are presented in ascending order in terms of weight magnitude, the judgement of the associated stimulus weight will be larger, than if the standard weights are presented in descending order. Bevan et al. ( 1963) concluded that the
W. Kanvowski et al.
732
Downloaded by [UQ Library] at 10:34 21 November 2014
judgement of a stimulus weight with respect to standard weights is affected by whether or not a given standard weight is greater or smaller than the standard weight it precedes o r follows.
3. Decision-based model for determining MA WL values Detection of information about load heaviness is one of the main cognitive functions of the subject involved in weight adjustment process when seeking to establish MAWL values based on the psychophysical approach. Holland and Lockhead ( 1 968), and Ward and Lockhead (1 97 1) showed that the judgement of weight in a current trial is influenced by a subject's feedback from judgement of weight from the previous trial. Therefore, an important issue in assessment of acceptable load is the ability of the subject to discriminate between consecutive weights of the box lifted, especially under the time pressure of high task repetition. This unexplored issue is very important in view of the complexity of the human decision-making process during the weight adjustment period, and its potential consequences for the quality of MAWL values derived on the basis of psychophysical methodology. It can be postulated that inability of the subject to identify a difference between two weights during the. act of load adjustment will most likely lead to the next adjustment. In other words, if the magnitude of the first and consecutive adjustment(s) is not sufficient for the subject to detect the weight difference (called here the load heaviness discriminability), he/she will continue the adjustment process. In such a case, while the absolute magnitude of weight changes may be significant, it will never be perceived as such by the subject who tends to compare bet ween weights of the successive boxes lifted. This process can be conceptualized through the decision-based model illustrated in figure 1. Decision making refers here to the process whereby subjects evaluate information about load heaviness which is available through the initial perceptual processing (Salvendy and Knight 1987).
I gure 1.
AATlNO OF ACCEPTABLE WORKLOAD
I
+
A decision model-based approach to determining the maximum acceptable weight
of lift (MAWL).
According to the proposed decision-based approach to determining the maximum acceptable weight of lift, failure to discriminate the change in weight through perceived box heaviness at time ti, may lead to the subsequent weight adjustment(s), and, consequently, potential overestimation of the true MAWL values. Such
Discriminability of load heaviness
733
Downloaded by [UQ Library] at 10:34 21 November 2014
overestimation of the MAWL values within limited time exposure to selected workloads and under time pressured decision-making has been well documented for higher task repetitions (Ciriello and Snook 1 983, Karwowski and Yates 1986). 4. Objectives Subjective perception of load heaviness plays an important role in the study of manual handling capacities, and should be considered when establishing guidelines for safe lifting (Karwowski 1988, Karwowski 1991). Previous studies in this area concentrated on determination of values of the Weber fraction (Carlson et a/. 1977), which defines the human ability to successfully discriminate between weights only 50% of the time. From the psychophysical approach to manual lifting point of view, however, it is very important to establish much higher levels of load heaviness discriminability, for example at 90 or 95% level. This would increase reliability of the selected MAWL values, and improve the confidence in results of the psychophysical studies in manual material handling. Therefore, the main objective of this study was to investigate human ability to discriminate between weights of lifted loads within the range from 2-4 to 29.1 kg, as the function of their absolute load magnitude, and the weight difference between two successive boxes in a sequence (called here the
load differential). 5. Pmedures 5.1 . Subjects . Twelve male college students, who were compensated for their time, took part in the laboratory study. Each subject was required to fill out a personal consent form and a data form. Only those subjects that were free of any medical problems were allowed to participate. The subjects were asked to wear comfortable clothes, stretch before the experimentation to minimize any potential for injury, and refrain from any strenuous exercises prior to the experimental sessions. Age and anthropometric characteristics of the subjects are shown in table 1. The isometric strength measurements were made according to the procedures proposed by Chaffin (1 975). The isokinetic strength tesls were performed according t o procedures described by Pytel and Kamon ( 198 1 ).
5 . 2 . Experimental design Within subjects, randomized complete block design was used, with subjects as blocks and the load differential as an independent variable. Each of the 28 box sequences consisted of five compact boxes (36 x 36 x 34 cm) weighing from 5 to 64 Ibs (2.3 to 29.1 kg). This range of weights constituted a range of loads that is common to many material handling tasks performed in industry (OSHA 1982). Within each sequence of weights, there was a constant load differential of either 1, 2, 4, or 8 Ibs (0-45,0.9 1, . 1.6, or 3.63 kg). Table 2 illustrates the distribution of weights in each sequence as a function of load differential. The dependent measures were the number of sequential box ordering errors with respect to perceived load heaviness, subject confidence leveis in the correctness of box sequence order, and confidence levels in assignment of linguistic descriptors to each of the boxes.
5.3. Methods All of the boxes were identical in appearance, and a cover was placed over the weights to prevent subjects from seeing the box content. The boxes were placed in
W.Karwowski et al.
734 Table I .
Age, physical characteristics and static and dynamics strengths (kg) of the subjects (n= 12).
Mean
Variable
SD
Range
Downloaded by [UQ Library] at 10:34 21 November 2014
Age (years)
Body weight (kg) Stature (cm) Acromial height (cm) Arm length (cm) Forearm length (cm) Static arm strength (kg) Static back strength Static composite strength Static leg strength Dynamic arm strength Dynamic shoulder strength Dynamic back strength Dynamic lift strength
Table 2. Arrangements for experimental box ordering sequences.
Sequence number
Load differential
Box
weight
8, 16, 32, 48, 6, 16, 24, 32, 40, 48,
12, 20, 36, 52, 8, 18, 26, 34, 42, 50, 56, 58,
16, 20, 24, 28, 40, 44, 56, 60, 10, 12, 20, 22, 28, 30, 36, 38, 44, 46, 52, 54, 60, 62,
24 32 48 64 14 24 32 40 48 56 64
Number of Total number sequences of trials
4
4
240
2
7
420
Total
28
1680
5, 6, 7, 8, 9 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, 52,
9, 13, 17, 21, 25, 29, 33, 37, 4 1, 45, 49, 53, 56, 57, 60, 6 1,
10, 11, 12
14, 18, 22, 26,
30, 34, 38, 42, 46, 50, 54, 58, 62,
15, 19, 23, 27, 31, 35, 39, 43, 47, 5 1, 55, 59, 63,
16 20 24 28 32 36 40 44 48 52 56 60 64
Downloaded by [UQ Library] at 10:34 21 November 2014
Discriminability of load heaviness random order in marked areas on the floor that were evenly spaced apart. Subjects were first asked to lift each box once to about knee height in order to get an idea of its heaviness. Afterwards subjects were allowed to lift the boxes as many times as they wanted and in any order. Subjects were instructed to lift each box and put it on a table (89 cm high) in the marked areas, either in descending or ascending order of weight. The subjects were also encouraged to compare two boxes at a time, and to rearrange the order of the boxes according to their perceived heaviness at any time. No time restriction or task repetition requirements were made. Upon completion of the ordering procedure, subjects were presented a chart with the following linguistic terms for load heaviness: very light, light, less-than-medium, medium, more-than-medium, heavy, very heavy, and extremely heavy (cannot lift), and asked to assign a linguistic value to each of the boxes. Subjects were told that the value they assign should describe how they feel about heaviness of each load (box) lifted. Before assigning the linguistic values of heaviness, subjects were asked to lift each box at least once again. This was done in order to minimize the chance of assigning a range of weights and linguistic values instead of judging them separately. Subjects were also asked to assess the level of confidence regarding the correct order of box sequence, and the confidence in assigned load heaviness values. A lOcm bipolar scale, marked 'completely ltncertain ' and 'absolutely positive' on its left and right ends, respectively, was used for that purpose. A minimum of a 10min rest break was taken between completion of each box sequence trial.
6. Results and discussion The results were analysed using an analysis of variance (ANOVA). The Newman-Kuels test was used to test differences between means of the response variables. In addition, the Kruskal-Wallis test was used to analyze the subject level of confidence based on the assessment scale lacking the interval scaling properties. 6.1. Number of box heaviness ordering errors Overall, the subjects made a total of 567 errors out of 1680 possible box positions (resulting in the 33.75% error rate) for all ordering sequences. An average error rate was less than two mistakes per sequence. The analysis o f variance revealed highly significant effects (p
