Perception of effort in manual materials handling

GAMBERALE F. Perception of effort in manual materials handling. Scand J En viron Health 1990;16(suppl I):59-66. Measurement of subjective work load has emerged as a central topic of interest in the study of physical work and effort. From the practical point of view, subjectiveestimates of physi cal effort can contribute greatly to the assessment of work loads in physical activitiesand to the design of manual handling tasks. The rationale of this approach is that subjectiveestimates reflect the individual perception of the amount of physicaland motivational resources that thesubjects invest to meetthedemand imposed by the physical task. The present paper provides an illustration of some of the techniques used in measuring the perception of physical effort and reviews and discusses the main results obtained with these techniques in the assessment of manual materials handling.

K ey terms: maximum acceptable work load, perceived exertion, physical work.
From the ph ysiological po in t o f view, indi vid ual capacity to per form physical work is limited by central factors associated with the pulmonary, circulato ry, an d meta bolic systems an d by local factors concerning muscula r strength, joints, and the skeleto n . O n the basi s of these con siderations many biomechanical and physio logical met hod s have been deve loped to predict the wor k capacity of indi vid ual s or to assess the ph ysical dema nds of specific ma nual handli ng tasks.
From the psychological point of view, motivational factors play an important role in the perception of effort and , con sequently, in determining the ind ividual capacity for physical work .
Subjective reactions to physical work, such as perceived effort or perceived work load , have often been foun d to correlate with wo rk intens ity and wo rk performance. Until recent ly, however, subjective mea sures of fatigue have not been seriously considered as co nstituting a pos sib le basis for criteria in the asses sment of manual materia ls handling. T he majo r reason for the neglect of subjective rea ctions in fa vor of ph ysiologi cal ind icators of fat igue is t hat these reactio ns hav e been difficu lt to define a nd measu re. Anot her rea so n has bee n the lack of fam iliarity with the use of "sophisticated" psychophysical methods on the part of the ph ysio logist. However, so me of the difficulties encountered in the use of subjective ind icators of fatigue are fundamental and are connected with the defi nition of the subjective dimensions involved and wit h the na ture of the measurem ent itself. Th us subjecti ve reac tio n to physical wor k can on ly be measured indirec tly through the use of self-repo rted tec hniques, and sometimes it is no t possible to ascertain to what extent the response obtained reflects per cept ual or cognit ive processes. Th e suita bility o f su bjective indica tors of fa tigue as criteria in the assessment of ma nual materia ls ha ndlin g will always depend on a series of elements which affect the reliabilit y a nd validity of the measurements, eg, (i) the type of subje ctive reaction o bserved, (ii) the way in which the reaction is observed and reco rded , (iii) the extent to which the rea ction var ies systematica lly in di ff erent work opera tio ns, (iv) ho w well the reaction correla tes wit h work int ensit y and work performance, (v) how well the react ion correlates with the physiological and neurological event s, and (vi) the extent to which the reaction ca n be predicted from knowled ge of the individua l and of th e work operations.
An anal ysis of the literatur e clearly shows that in recent years the st udy of the subjective reactio ns to physical wo rk has reli ed almost excl usively on a few type s o f methods, eg, (i) magnitude estimation, (ii) ratings of perceived exertion, (iii) ratings of effort expended, and (iv) rati ngs of acceptable load. Th e psychop hysica l met ho ds, mainly magnit ude est imatio n, ha ve bee n mo st ly used in basic resea rch with the aim of studying th e rela tion between the physical or objective work load and the subjective wo rk load. [See the paper by Borg (I)] . T he ot her me thods have, however , been more wide ly used in applied ergonomics, and the se will be described in th e pre sent paper, in which the main res u lts obtained throug h their applica tio n wiII also be reviewed.

Magnitude estimation
In magnitude estimation the subject is req uired to mat ch a num ber directl y to the perceived intensity of a phy sica l exercise in reference to a sta nda rd in tensity which has a given numerical magnitude. T he basic assumption of thi s direct-sca ling met hod is that the subject is able to ma tch his perception with nu mbers. Numero us empirical studies ha ve shown that su bjects

Magnitude estima t ion (MEl
To give a more co ncrete example of the applica tion of ratio scaling to the study of perc eived exertio n in manual materials handling, some of the result s of an experiment performed in the laborator y (8) will be referred to in more detail.
In this experiment five brewer's dra ymen with several years of occupational experience were stu died in the labora tory durin g horizontal lifting work , which simulated an ord er filling task at a brewery. Th e ta sk consisted of lift ing a case (36 x 36 x 24 ern) app roximat ely 1.1 m hor izontally at a height of 60 ern. In order to facilitate repeated lifting , a treadm ill was used to return the case after each lift. The work was performed in cycles o f three conse cut ive lifts fo llowed by a rest period . The subjects worked at a given lifting rate with cases weighing 67, 100, 133, 167, and 200 % of a reference case weight, which they had indi vidu ally chosen as constituting the maximum amount that they could lift without strain and discomfort. (See the section Acceptable Load.) After 11 min of work at each load, each subject was asked to give a magnitude estimation of the work load . The estimations were made by having the subject indicate a number equ ivalent to th e perceived work load in relation to the immediately preceding work load , which was always assigned the value 100. Th e seq uence of work load s was chosen in such a way that when a subject had comp ared each load with the immediately preceding load , he had performed all possible pair wise comparison s between the load s. Since the sequence of the work load s was also perform ed in reverse order, a complete ma tri x of relati ve comparisons between case weights was obtained for each subject. On the basis of these matrices, a scale of perceived work load was calcul ated fo r each subject acco rding to Ekman (9). With the same methods, a subjective scale was also calculated for all the subjects o n the basis of a ratio matrix of geometric mean values. The results of this part of the investigation are illustr ated in table I and figure I. As shown, the relation between the magnitude estimates of perceived work load and case weight (objective load) is satisfactorily described by the psychophysical power function. The exponents of the individual subject fun ctions varied between 1.44 and 2.45, indi catin g th at a doubling of the case weight would result in an approximately three-to fivefold increase in perceived work load .
The same techniq ue of collecting magnitud e estimation s and produ cing subjective scales o f physical work load was also used by my co-workers and I (10) in an experiment in which a compl ex, diagon al lift ing ta sk was studied. (See the section Acceptable Load. ) In this experi ment th e subjective work load was also a po wer function of the ph ysical work load , as illustrated in figure 2.
In summar y, th e application of the psychophysical method of magnitud e estimati on to the study of physical work can provid e a detail ed descript ion of the relationship between the perce ived and the ph ysical level are able to perform this task and tha t the subjective scale obtained is a ratio scale.
The results of several psychophysical stud ies using magnitude estimation clearly indicate that the perceived inten sity of muscular effort is a power function of the ph ysical work load. [See also the paper by Borg (I)] . Thus Stevens & Mack (2) obtained a power function with an exponent of 1.7 when scaling subjective force, as exerted in the squeezing of a handle. Borg & Dahlstrom (3) investigated muscular work carried out on a bicycle ergometer and obt ained an expon ent of 1.6. Exponents of the sam e magn itude were also found by Eisler (4) in relating subjective force to physical force both for work involving large mu scle groups, ie, exerting a force against a foot pedal, and for work involving small muscle groups, ie, sq ueezing a handgrip. A power function with an exponent of 1.6 was also obtained by my co-workers and r (5) in an investigation of the subjective experi ence associated with external resistance to breathing.
Perceived effort depends not onl y on the intensity of the ph ysical work performed , but also on the duration of the work . When investigating the combined effects of the two factors, Stevens & Cain (6) and, later, Caffarelli et al (7) found that the increase in effort with time also followed a power funct ion . The exponent of this function was shown to depend on the type of work investigated and on its inten sity.  of exertion. There is strong empirical evidence that perceived exertion is a positively accelerated function of work load, irrespective of the type of work performance or the muscle groups involved.

Perceived exertion
By far the most important contribution in the study of the psychological aspects of physical work has been made by Borg (11,12), who is also the author of the most frequently used scale for rating the degree of perceived exertion during physical work, the so-called RPE scale (rating of perceived exertion scale) (13). The scale, printed in a quarto format, is as follows: The RPE scale has been developed on the basis of empirical data from work on the bicycle ergometer. "In fairly young to middle-aged people (25 to 45 years old), working at moderate to high intensity levels, the heart rate roughly corresponds to 10 times the RPE value [p 341]" (14). Thus, according to Borg, the relation between RPE and heart rate for work on the bicycle ergometer is linear. Consequently, RPE is also a linear function of the physical work load. This conclusion should not be interpreted as conflicting with the results of several psychophysical studies, in which perceived exertion was found to be a positively accelerated function of the work load. It is important to observe that the form of the relation obtained when RPE values are plotted against the corresponding values of heart rate for the same work load depends largely on the specific characteristic of the rating scale itself, ie, the number of categories, the verbal definition, etc. One of the objectives in the construction and development of the scale was in fact to obtain a linear relationship between the RPE and work load. This objective was achieved by a careful choice of verbal categories.
The RPE scale has been the most frequently employed method for the assessment of subjective qualities during physical work. Thus a substantial volume of literature has become available concerning the significance of perceived exertion as measured by the RPE scale. Since RPE values could be used as a complement to circulatory responses during exercise, and in fact the scale was constructed with this objective, it is of particular interest to observe circumstances when changes in physiological responses are not followed by corresponding changes in RPE.
In a study by Pandolf et al (15), changes in temperature were found to affect heart rate but not RPE. In another study (16) the heat gain resulting from wearing an unventilated gas-protective suit during simulated firefighting and gas accident practice, as well as during exercise on a bicycle ergometer, brought about a considerable increase in heart rate which was not followed by a corresponding increase in RPE. These results suggest that heat load does not have an impact upon perceived exertion that is comparable to the physiological strain. it produces. Thus heat load may create conditions in which a person will overestimate his or her physical endurance. An overestimation of this kind could be a determinant factor in the occurrence of cases of exhaustion and collapse observed in firefighting and other extreme situations (16). Provided that exercise on the bicycle ergometer is performed at the same work load, eg, 100 W, different pedaling rates produce the same heart rate. However, Henriksson et al (17) and Pandolf & Noble (18) have reported that pedaling at a rate of 30 or 40 revolutions/min resulted in a higher RPE than pedaling at a rate of 60 or 80 revolutions/min. Ekblom & Goldbarg (19) and Sjoberg & Frankenhaeuser (20) used autonomic blocking drugs to affect heart rate during physical work. While heart rate changed in the expected direction as a result of the drugs, RPE was unaffected. Pandolf et al (15) and Morgan (21) reported that while heart rate was unchanged during the course of work on a bicycle ergometer, RPE tended to increase after 5 min of work. Similar results were obtained by Ljungberg et al (8) during horizontal lifting work. RPE collected after both 4 and 13 min of work showed a significant increase, while heart rate did not display any noticeable differences. Finally, my co-workers and I (22) showed that, at the same heart rate, exercise on a bicycle ergometer was perceived by a group of women as more demandir.g during menstruation than during either the premenstrual or postmenstrual phase. The changes in RPE during the menstrual cycle were interpreted as due to motivational factors.
Many investigations have shown that the relationship between RPE and heart rate is highly dependent upon the type of physical task involved. At a given heart rate, RPE is lower for running than for work on the bicycle ergometer (19,23) or for walking (24). At the same heart rate or at the same oxygen uptake, arm exercise also gives a higher RPE than cycling or pushing a wheelbarrow 09, 2'5). Changes in RPE were also associated with increases in blood lactate concentration, which led to the formulation of the following hypothesis: "The higher the blood lactate concentration an exercise produces as compared to oxygen uptake, the higher will be the level of the overall perception of exertion ... [p 553]" (25). In addition to the RPE scale, other rating scales have been used for measuring perceived exertion. However, not one of these scales has shown the same versatility, parsimony, and validity as the RPE scale.

Expended effort
In several investigations subjective fatigue during prolonged physical exercise and endurance tasks has been studied while the subjects rate the degree of effort expended over time. The technique used was developed by Caldwell (26) for use with an isometric handgrip task. The procedure consists of letting the subjects indicate when, in an endurance task, they per-ceive the expenditure of one-fifth, two-fifths, etc, of their available effort or when, in the same self-paced fashion, pain intensity attains values of 1-5.
Using this technique, Caldwell (26), Caldwell & Smith (27), Menzer et al (28), and Lloyd et al (29) obtained a linear relationship between time on an isometric task and subjective estimates of both pain intensity and perceived effort expended. With the use of the same technique, subjective estimates of effort were also found to be linearly related to endurance time during treadmill performance (30).
It has been suggested (31) that the self-paced rating procedure used in the aforementioned studies might have contained a timing artefact, in that the subjects might have been judging elapsed time rather than effort expended. The relationship between subjective estimates and time on the task would attain spurious linearity if the subjects of these studies were matching intervals on the time dimension.
The suggestion of spurious linearity is indirectly supported by the results of a recent experiment by Kilborn et al (31), in which static exercise was performed as a 900 elbow flexion at 25 070 of maximal voluntary contraction force, either until exhaustion or for a duration of 20-80 % the of maximal endurance time. In this experiment the subjects were requested to rate effort expended after the exercise had been interrupted by the investigator. Furthermore, the subjects rated effort expended only once at the end of each submaximal test, and only one submaximal test was performed each experimental day. This procedure was adopted to minimize the possible effect of time on the ratings. Under these conditions, the relation of perceived effort expended and time on task was not linear. Thus 15 out of 18 subjects overestimated effort expended as related to endurance time, or, in other words, they underestimated their maximal static endurance capacity during the task.
As has already been mentioned, a category scale with numerical values from I to 5 has been used in some investigations to collect estimates of the intensity of pain felt during prolonged physical work. In these studies the researchers also made use of the self-paced technique already described. The value' of 1 was defined as constituting the first noticeable pain, and the value of 5 as intolerable pain, ie, the pain perceived at the end of the endurance task. In the experiment by Ljungberg et al (8) reviewed previously, subjects were also required to rate pain intensity on a 5-point rating scale ranging from "no pain" to "intolerable pain." In this experiment no subject rated perceived pain using the highest pain intensity of the scale at maximal endurance time, which clearly indicated that perceived pain in the local muscles was not the primary factor limiting endurance in the task. Considering these results, it seems clear that perceived pain intensity, especially if assessed at submaximal endurance time, would be a poo r predictor of the static enduran ce capacity of the ind ividual.
The use of the rating seales for expended ef fort and per ceived pain in the study of per ceived exertion during ph ysical exercise has produced result s so mewhat less con vincing than tho se obtained by the use of ratioscaling techniques or the RPE . The relation between effort expended and time on ta sk in prolonged physical work or endurance tasks has not been as well explored and documented as, eg, the relation betw een perceived exertion and work load .

Acceptable work load
Three ma in approaches can be used to determ ine how much load a worker can handle sa fely. Th ey are based on estimations of biomechanical, physiological, and perceptual stre sses.
The biomechanical models (32-34) emphasize the load s on the musculoskeletal system during the liftin g tas k and pro vide estimations o f the forces and torques generated during the lifting task . The limitations of the biom echanical models are primarily th at th ey do not dire ctly predi ct work capac ity , that the y are static in nature, and that the y ha ve onl y considered work on the sag itta l plan e.
The ph ysiolo gical models (35,36) are based on predictions o f the metabolic energy expenditure in lifting and in other manual materials handling tasks. As in the case of the biomechanical models, the ph ysiological models are restricted to lifting in the sagittal plane alone. Moreover , the se models make assumptions with regard to the additivity or comparability of energy costs for different tasks which ha ve not been validated.
As an alt ernat ive or a compleme nt to the biomechan ical and ph ysiological mod els, lift ing capacity prediction models ha ve been developed on the basis o f a psychophysical approach . These mod els are among the few available for determining the maximum acceptable level of work load in manual mat erials handling ta sks.
Th e meth od used resembles in some respects o ne of the basic psychophysical methods , ie, the method of adjustment , used to determine perceptual th reshold (37). The procedure is to ask th e subject to adjust his or her work load , ie, the weight o f the material to be handled or the frequ ency of the op eration , to the maximum he or she can perform witho ut fallin g into ove rexert ion or excessive fati gue . The subject is told to " imagine you are on piece-wo rk, gett ing paid for the amount of work you do, but working a normal 8-hour shift that allow s you to go home without feeling bushed, unusually tired or weakened ." Following the procedure used in the method o f adjustment, the subject will start with a high work load on on e occasion and with a low work load on another. The two res ults ar e th en averaged . Generally the subject is given 20-40 min to adjust the work load.
Undoubtedly the psychophysical approach is very attractive in its simplicity and deser ves serious attention . The applicability of the method relies on the assumption that the su bject is able to estimate with some accur acy his or her highest work load. Furthermore, it is assumed that the work loads accepted by subjects under conditions which simulate real work are lighter than the loads leading to manual handling injuries among workers. However, as yet, it is not known to what extent these assumptions are valid.
Using this technique, some researchers (38)(39)(40)(41)(42)(43) ha ve systematically collected estimates of work load during different manual materials handling activities standard ized and simulated in the laboratory. The results of a series of studies using the psychophysical approach ha ve been summarized and integrated by Snook (44). In his report Snook presents a series of tables indicating the maximum weight s predicted to be acceptable to 10, 25, 50, and 90 % o f the male and the fema le working populations. Similar tables have also been published by Ayoub et al (45), who combined their own data with the dat a published by Snook (44).
According to Snook (44), a proper use of these tables can reduce the occurrence of musculoskeletal injuries more effe ctively than selecting the wor kers for the job or training the workers to lift properly. Currently, however, there is no general agreement concerning the validity of these estimate s. Thus Mital (46), studying lifting work, was not able to validate the assumption that people can estimate the amount of weight they can actu ally lift 8 h a day based on the effort they perceived in 20-30 min of work . He also noted that the psychophysical method tends to overestimate the maximum acceptable weight of lift.
Recent studies (8, 47) ha ve sho wn that subjects tend to select higher wo rk loads when adjusting the frequency o f the op er at ion performed than when adju st ing the weight of the object the y have to handle. In one of my own studies using thi s psychophysical method (8), subjects working with horizontal lifting chose about half as heavy a case as subjects working with vertical lifting at a comparable lifting rate. Similar result s ha ve also been reported by Garg & Badger (48). Th ey found that an increase in the as ymmetry of a lift ing task was alwa ys follo wed by a decrease in th e work load preferred by the subject.
To test the reliability and validity of thi s psychophysica l method, my colleagues and I have perform ed three experiments (8,10). The concept of ma ximum acceptable work load is presented in the following discuss ion in light of the results of these experiments.
The fir st experiment (8) has already been referred to in connection with th e illustration given on the use of the method of magnitude estimation . In this experiment , following the described procedure, five brewer's dra ymen estimated the maximum acceptable work load of a horizontal lifting task lasting 36 and 18 s with two workplaces.
The subjects had no difficulty in followin g the instructions. Thus it only took about 10 min fo r them to make their choice of maximum acceptable weight. The choices were cons istent irrespective of whether a subject had started with 5 or 25 kg in the case. The intersubject rankings with regard to the maximum acceptable weight were identical for all the trials and therefore indicated that the reproducibility of the results was very sati sfactory. However , the interindividual variation was conside rable and did not appear to be related to the subjects' physical characteristics, eg, body size, muscle strength, and aerobic power. Thus the lowest weight (5.3 kg) was selected by a relatively powerfully built person with considerable muscle strength and high aerobic power, whereas the heaviest weight (17.0 kg) was chosen by a much shorter and lighter subject with less muscle strength and lower aerobic power.
Other results of this experiment should also be taken into con sideration when the va lidity of the psychophysical approach is assessed. First, when the lifting rate was doubled, the average weight selected was only lower by II 0,10, although the subjects performed 74 % more work . This result indicates that physiological factors may not be the primary determinants of the work loads selected. Second, in a study of vertical lifting with the same instructions and comparable lifting rate (49), the subjects preferred to work with about twice as heavy a case as in the present study. This result clearly indicates that horizontal lifting, which is very common in real materials handling situations, imposes a greater strain than vertical lifting. Thi s fact would suggest tha t the values of maximum acceptable work load s collected in standardized lifting situations only have a very limited representativity.
Two groups of subjects participated in the seco nd experiment (10). One group consisted of eight warehouse workers with occupational experience of lifting work, and the other group consisted of eight office employees without such experience. A diagonal lifting task was used in this experiment. It consisted of four consecutive lifts. The first lift was from a table on the left side (knuckle height) to a shelf on the right side (knee height) . The second lift was mo ving the case back to the table. The third lift was from the table on the left side (knu ckle height) to a shelf on the right side (shoulder height). The fourth lift, concluding the lifting cycle, was mov ing the weight back to the table.
During the first part of the experiment the subjects performed the lifting task repeatedly using five weights and five work paces. On this occasion the aim was to study the relation between the objective and subjective work loads. As expected from the results of the prev iou s experiment, the percei ved work load for thi s type of lifting ta sk was found to be a po sitively accelerated function of the objective work load , as defined by the weight lifted and by the work pace.
During the second part of the experiment th e subjects selected acceptable weight s for this lifting exercise at three different work pac es and acceptable work paces for three different weight s. An analysis of the results obtained revealed that there were large and systematic differences in the selection of work loads between two groups of subjects who had been instructed by two different experimenters . A thorough investigation afterwards revealed that this bias was due to the fact that during the lifting task one of the instructor s had reminded the subjects to adjust the wor k load if they felt like it. No such action was taken by the other experimenter. Thus reminding some of the subjects that they could adjust the work load any time they wanted produced a systematic increase in the maximum acceptable work load finally selected by the subjects . This result clearly indicates that the data collected by the use of the psychophysical approach are highly dependent on the procedure adopted for instructing the subjects. It seems evident that thi s characteristic of the method must con stitute a definite limitation to its applicability.
In the third experiment (10) two new groups of subject s were chosen following the same criteria as in the pre vious experiment. The lifting task and the general pro cedure of thi s experiment was the same as in the pre vious experiment. However the instructions given the subjects were further standardized and all the subjects were instructed by the same experimenter. The subjects selected acceptable weights at three different work paces and acceptable work paces for three dif ferent weights. They also selected an acceptable work load by changing both the weight and the work pac e. For each work load selected , the subjects rated perceived exertion according to the RPE scale (13). The entire pro cedure was then repeated, the replication of each selectio n of work load bein g retained.
On the whole, the reproducibility of the selected work loads was sati sfactory also in this experiment. However, it was apparent that selecting the work pace was an easier and more reliable task than selecting a weight.
Two feature s of the results of the present experiment ar e relevant for th e assessment of the validity of published data co ncern ing ma ximum accept able work load s collected with the psychophysical approach. One feature is that occupational experience of lifting work had a negati ve effect on the selection of work loads. Thu s the maximum acceptable work load s selected by the warehouse workers were systematically lower than those selected by the office employees. This ph enomenon was also clearly evident when subjects were left free to select both the weight and the work pace in the same trial. Under these conditions the warehouse workers selected a 78 % lower weight and 87 % lower work pace on the average. Although they selected lower work loads, the warehouse workers con sistently rated overall percei ved exertion and the perceived exertion on the arms and shoulders, the back and the legs higher than the office employees. The other feature of significance, with regard to the validity of the method, is that the subjects' physical characteristics, eg, body size, muscle strength, and aerobic power, showed no relation to the work loads they selected.
To summarize, the results of the three experiments raise questions concerning the representativity of psychophysical data collected in the laboratory regarding maximum acceptable work loads for manual materials handling operations. Furthermore, as the experiments have shown, it is difficult to reach a satisfactory understanding of the significance and implication of the concept of maximum acceptable work loads for lifting tasks. The results indicate that the process of selecting a work load might be governed both by cognitive and motivational factors and by perceptual stimuli.

Concluding remarks
Independent of the specific psychophysical method used to measure it, the perception of effort should be interpreted as constituting a "summing up" of the influence on the organism from all structures under stress during physical work. Although there is no objective counterpart to this perceptual phenomenon, the perception of effort during physical work not only has a psychological validity, but it also reflects real conditions such as the interplay between the requirements of the physical task and the capacity of the individual. No single technique, of those described in this paper as available for measuring subjective reactions during physical work performance, can be considered generally superior to the others.
Undoubtedly, a ratio scaling method such as magnitude estimation is particularly suitable for studying the relation between the perception of effort and the objective work load for a specific physical task. However, in practical settings where an assessment of the physical requirements of different work situations is needed or when interest is focused on interindividual or intraindividual comparisons, the use of a rating scale of perceived exertion can have clear advantages over magnitude estimation.
Rating scales of effort expended have been applied to the study of physical work almost exclusively to collect estimates of effort or pain in prolonged exercise or endurance tasks. As yet the reliability and validity of the measurements of effort expended have not been satisfactorily evaluated.
The use of the method of selecting an acceptable work load seems to be restricted to situations in which the objective is to evaluate manual materials handling activities and to generate norms aimed at the prevention of manual handling injuries. Although the validity of the method in this respect has not yet been demonstrated, the method deserves serious attention. There seems to be good reason to investigate the possibility 5 of combining the psychophysical approach with the biomechanical and physiological ones to develop an integrated model capable to predict, and therefore to prevent, musculoskeletal injuries.