5
ACCIDENT PRONENESS C0t.w CAMEROX r\ustralian Rod Research Board. Vermont 1. INTRODUCTIOS Vie.. Australirt Review papers dealing with accident proneness have appeared at regular intervals for many years. Among the most recent. and one of the best. is Ferguson’s contribution to the Seminar On Occupational Injuries organized by the Royal Australasian College of Surgeons, in March of this year [Ferguson. 19733. A symposium of five papers on the topic was published in the i?Ic&-t~l Jo~rt~~l of Ausfrtrlicl as recently as October I97 I. For the past 20 years or so. reviewers have concluded. without exception. that il~dividL1~il s~lsceptibility to accidents varies to some degree. but that attempts to reduce accident fie- quency by eliminating from risk those who have a high susceptibility are unlikely to be oficctive. The reduction which can be achieved in this way represents only a small fraction of the total. Attempts to design safer man--machine systems are likely to be of considerably more value. The snfcty profession has taken up this alternative approach with con~n~eli~~~~bi~ enthu- siasm. and has been supported by othor professional disciplines, notably cnginecring, mcdicinc and psychology. The legal profession has more rcccntly lcnf its support. and actions in which substantial dumagcs have been award4 againsl industrial corporations and public au thoritius on the grounds of un.safc &sign. poor maintcnanca or uncontrolled risk to persons have begun to appear in our civil courts quite froqucntly. Trade unions have c~lrnp;li~s~~d cfliictivcly to s~‘curc enforcement of safety r~~ill~lti~)rls in the b~~iI~~i11~ in- dustry, to take an example which reccivcd some publicity in Mclhournc recently. It looks as though the 50-yrar campaign to lay Ihc ghost of accicjcnt proncncss is nearly over. In 1971 Pcrgamon Press issued Volume I I in tticir prestigious Intcrniltio1~al Scrims of Monographs in Experimental Psychology. It is cntitlcd ,4cciJ~f f’rorrc~uc~.s.s and tk itl~thors arc Lyncttc Shaw and Herbert Sichcl, both well known for their joint rcscarch into accident c~tu~tion among bus drivers in South Africa. The book has ~~;~t~;r~~lly attracted a good deal ofattention; thcrc has been no other attempt to examine the concept in such detail, and at such length. The purpose of this paper is to examine the cvidencc prcscntcd and the arguments dcvcl- oped by Shaw and Sichel, since they draw conclusions about the importance of accidcnl proneness which arc at variance with those of most other writers on the subject. If they are correct, then the emphasis on safety by design and protection which has ~tl~~r~~~t~ri~~d our efforts in rcccnt years may be misplaced. A rc-examination of our methods ma): be necessary. 2. THE NATURE 01’ AC(‘II>liNT I’KONIINL:.SS The book begins by reviewing the history of rcscarch on accident proneness since Greenwood’s pioneering work, and tracing the rise and subsequent dcclinc of enthusiasm for the concspt as an explanation for ditTerenGal accident involvcmcnt. An important point is made -accident proneness means different things to dif7iircnt pcoplc. The view of acci- dent proncncss taken by early workers on the subject was that certain attributes rendtzrcd some individuals more likely to bccomc involved in accidents than others. The distribution of accidents among the pop~llation at risk was not random. They suggcstcd that if thcsc attributes could be identified and measured it might be possible to reduce accident occur- rence by screening out persons whose relatively high rating on these attributes indicated a potential susceptibility to accidents. This view, as stated, is statistically sound. but later research has encountered ditliculty in identifying and measuring the relevant attributes.

Accident proneness

Embed Size (px)

Citation preview

ACCIDENT PRONENESS

C0t.w CAMEROX

r\ustralian Rod Research Board. Vermont

1. INTRODUCTIOS

Vie.. Australirt

Review papers dealing with accident proneness have appeared at regular intervals for

many years. Among the most recent. and one of the best. is Ferguson’s contribution to

the Seminar On Occupational Injuries organized by the Royal Australasian College of

Surgeons, in March of this year [Ferguson. 19733. A symposium of five papers on the topic

was published in the i?Ic&-t~l Jo~rt~~l of Ausfrtrlicl as recently as October I97 I.

For the past 20 years or so. reviewers have concluded. without exception. that il~dividL1~il

s~lsceptibility to accidents varies to some degree. but that attempts to reduce accident fie-

quency by eliminating from risk those who have a high susceptibility are unlikely to be

oficctive. The reduction which can be achieved in this way represents only a small fraction

of the total. Attempts to design safer man--machine systems are likely to be of considerably

more value.

The snfcty profession has taken up this alternative approach with con~n~eli~~~~bi~ enthu-

siasm. and has been supported by othor professional disciplines, notably cnginecring,

mcdicinc and psychology. The legal profession has more rcccntly lcnf its support. and

actions in which substantial dumagcs have been award4 againsl industrial corporations

and public au thoritius on the grounds of un.safc &sign. poor maintcnanca or uncontrolled

risk to persons have begun to appear in our civil courts quite froqucntly. Trade unions

have c~lrnp;li~s~~d cfliictivcly to s~‘curc enforcement of safety r~~ill~lti~)rls in the b~~iI~~i11~ in-

dustry, to take an example which reccivcd some publicity in Mclhournc recently. It looks

as though the 50-yrar campaign to lay Ihc ghost of accicjcnt proncncss is nearly over.

In 1971 Pcrgamon Press issued Volume I I in tticir prestigious Intcrniltio1~al Scrims of

Monographs in Experimental Psychology. It is cntitlcd ,4cciJ~f f’rorrc~uc~.s.s and tk itl~thors arc Lyncttc Shaw and Herbert Sichcl, both well known for their joint rcscarch

into accident c~tu~tion among bus drivers in South Africa. The book has ~~;~t~;r~~lly

attracted a good deal ofattention; thcrc has been no other attempt to examine the concept

in such detail, and at such length.

The purpose of this paper is to examine the cvidencc prcscntcd and the arguments dcvcl-

oped by Shaw and Sichel, since they draw conclusions about the importance of accidcnl

proneness which arc at variance with those of most other writers on the subject. If they

are correct, then the emphasis on safety by design and protection which has ~tl~~r~~~t~ri~~d

our efforts in rcccnt years may be misplaced. A rc-examination of our methods ma): be

necessary.

2. THE NATURE 01’ AC(‘II>liNT I’KONIINL:.SS

The book begins by reviewing the history of rcscarch on accident proneness since

Greenwood’s pioneering work, and tracing the rise and subsequent dcclinc of enthusiasm

for the concspt as an explanation for ditTerenGal accident involvcmcnt. An important point

is made -accident proneness means different things to dif7iircnt pcoplc. The view of acci- dent proncncss taken by early workers on the subject was that certain attributes rendtzrcd

some individuals more likely to bccomc involved in accidents than others. The distribution

of accidents among the pop~llation at risk was not random. They suggcstcd that if thcsc

attributes could be identified and measured it might be possible to reduce accident occur-

rence by screening out persons whose relatively high rating on these attributes indicated

a potential susceptibility to accidents. This view, as stated, is statistically sound. but later

research has encountered ditliculty in identifying and measuring the relevant attributes.

although limited success has been achieved in selected groups for whom a high rejection

rate was acceptable. This view of accident proneness. according to the authors, remains current in most European countries. and attempts to develop methods of predicting acci- dent proneness are continuing. A particularly active area is traffic safety.

A contrasting view of accident proneness which gained currency in U.S.A. during the 1930’s. is that accident prone individuals contribute a large proportion of the accident total. That is, a small number of accident repeaters are involved in a very large number of accidents. If this “hard core” can be identified and eliminated. the accident problem can be greatly reduced. Shaw and Sichel (or rather Shaw. since Sichel appears to have contri- buted only a single chapter on statistical techniques) point out that the “hard core” view of accident proneness has never been seriously held except in U.S.A., and cannot be sup- ported by the evidence. It is not surprising. in Shaw’s view. that accident proneness is now considered in U.S.A. to be of very minor importance; the concept gained currency in such

an extreme form that it was bound to be discredited by the observed facts. American studies of accident repeating drivers indicated that eliminating these drivers from the driv- ing population would produce only a small reduction in total accidents during the follow- ing year; the great majority of accidents occurring in any one year were found to be first accidents. The American view of accident proneness was so far modified by the year 1962 that a leading authority could say “. . . the statistical evidcncc is quite convincing that the intlucncc of a proneness fktor. whcrc it can bc dctectcd in the accidents occurring within Iargc groups. is quite limited.” [McFarland. 19621 This period. in the area of highway

safety. wascharactcrixd by active endeavours to cnsurc that safety was built into vehicles and highways. and has icd to the dcvclopmcnt and installation of scat belts. impact- absorbing front and rear panels. “collapsihlc” steering columns, crash barriers. frangible

supports for road signs and street lamps. and. in short. to the grcatcst advances yet made

towards the reduction of in,jury and death on the highway. A loss of intcrcst in accident proncncss scums of little account in comparison with thcsc achicvemcnts.

Shilw has done a useful scrvicc in distinguishing thcsc two versions of accident pronc-

ncss. one of which is totally unsupported by the facts. The more reasonable view. that acci-

dents arc not uniformly distributed across the population at risk. is tcnablc, but in the abscncc of rcliablc methods of identifying those who are going to have accidsnts bcforc

the accidents occur. acccptancc of this view dots not take US very fk. Indeed Shaw’s view of accident proneness. that “even when exposed to the same conditions. some people arc

inhcrsntly more likely to have accidents than others” is a mere statistical platitude. Shaw insists that pcoplc who “arc inhcrcntly more likely to have accidents” can bc idcn-

tificd by means of pcrformancc and personality tests, togcthcr with some kind of clinical

appr;lisal. tlcr own work with bus drivers in South Africa is drawn upon to support the

argument. and is imprcssivc within its limitations. The group of drivers with whom she

worked. howcvcr, appears to bc a highly unusual one, c ,lnd she may have been drawn into

the error which she attributes to American authorities, that of generalizing from a specific

set ofcircumstanccs to a gcncral condition. The clinical descriptions of accident prone bus

drivers which she includes in her book are descriptions of people who would bc rcgilrded as unsuitable for most kinds of employment. In other words. the attribute which Shaw

Iabcls accident proncncss is a composite ofothcr attributes which could bc idcntificd with- out invoking the idea of accident proncncss at all. The gcncral value of Shaw’s mcrhod ofprcdiction has not been fully demonstrated in her studies. although it may be quite satis-

factory for her specific purpose.

3. -l-Ilk! El--FECT OF LiIVE.L OF RISK

An important thcmc running through the book is that of the general level of risk, which

has been found in many stud& to bc a significant variable controlling the presence or abscncc of a proncncss factor. The point was noted by Farmer and Chambers [ 19393 and

is used by Shaw to explain why many research workers have failed to find a proneness component in accident statistics. She points out that in all the studies reporting the prcs- cnce of it proncncss filctor the gencrnl Icvcl of risk (that is. the probability of having an accident) has been comparatively high. When the overall risk is low. proneness “cannot

.Accidrnt proneness 51

operate”. to use Shaw’s own terminology. Statistical demonstration of this phenomenon

is fairly convincing. but it is hard to accept logically. One might expect that reduciny the

overall level of risk (that is. making the system as safe as possible) would tend to reduce

the random component in the overall distribution of accidents. thus throwing into promi-

nence those individuals who have a high “inherent susceptibility” to accident. Yet the

reported studies suggest the opposite. and Shaw makes considerable capital out of the fact. She points out that the American view that accident proneness is not important has been

developed mainly from road accident studies. Since American roads are among the safest in the world (that is. a low level of risk) they do not allow the accident prone driver to

reveal himself as different from his fellows. She does not appear to realise that this is exactly the same as saying that he does not exist. or at least does not exist as a problem. The best way of dealing with the accident prone driver. on this kind of evidence, is to make

the road system as safe as possible for all drivers.

Curiously enough. Shaw’s own results are not consistent with the general statement on

which she bases her argument. Quoting the Sichel statistic 5. which is the average number of days between accidents. she says: “It will be noted that the spread of t’s (for different

drivers) is narrowest . . .where the traffic conditions are very difficult and widest. . . where

operating conditions are much easier.” Large differences in i should make it easy to iden-

tify the bad risks. while small differences should make it ditficult. and the reported result

seems to argue strongly against the notion that accident proneness “operates” when risk

is high. Perhaps enough has now been said to indicate the general lint of Shaw’s arguments in

this book. But in summary she is saying that the importance of proneness has been under- cstimatcd for two reasons. The first is the exaggerated view of proncncss which gained cur-

rcncy among American rcscarch workers at a critical period in the history of accident

rescnrch. This caaggcratcd view was ncvcr consistent with the obscrvcd facts. The second is that gcncral improvcmcnts in the safety of man-machine systems, including road sys-

tcms. have led to a gcncrally lower Icvel of risk and thus to rcduccd opportunities for acci-

dent proncncss to bc dcmonstratcd.

4. Ttlfi INI~IVIOUAL NATURE 01: ACCtD11NT t’KONLiNI:SS

Shaw cmphasiscs the need for studies of individuals over a long period of time, if their susceptibility to accidents is to bc dctcrmincd. The cross-sectional studies favoured by stu-

tisticians arc not appropriate for studies of proneness. The point is valid, and the authors of this book have done valuable pioneering work in the development of appropriate statis-

tical tcchniqucs for the necessary longitudinal studies. They have also demonstrated that in rcstrictcd “captive” groups it is possible to predict with some degree of accuracy the

relative susceptibility of their individual members to accident. They have not demonstrated that accident proneness is an identifiable characteristic (or group of characteristics) in the

population gcncrally. nor that it makes an appreciable contribution to the accident total.

In particular. they have not demonstrated that proncncss is an enduring characteristic, un-

atfectcd by training and distinct from youth. incxpcrience, sex. high exposure, alcohol or

other variables which have been identified as inlluential in other studies. In attempts to

identify the “causes” of accidents. it appears to be more profitable to assess the contribu- tion made by specific factors which arc known to bc statistically associated with accident

occurrence than to employ a blanket concept of proncncss which does not distinguish these factors.

5. THE ACCIDENT PKONENIISS CONCEPT

A disturbing feature of the book is the niive way in which the term accident proneness is used. It is rcpcatcdly rcfcrred to as a “concept”. which suggests that it is a general notion describing a certain class of events: the term does not imply an explanation of events in

any causal SCIISC. ht the author rcpcatcdly refers to accident proncncss as an explanation of individual difTercnccs in accident involvcmcnt. DifTcrences are said to be “due” to acci- dent proneness. This is not an explanation. as the term is usually understood in scientific logic. It is like attributing occasional toothache to pcriodontic neuralgia. or a persistent

sore throat to chronic pharyngitis. It is not an explanation. it is a descriptive label for a statistical phenomenon. and takes us no Farther towards identification of the factors contri- buting to the phenomenon itself.

There is cause for concern, also, in Shaw’s frequent references to those who “believe in- or “have faith in” the accident pronenessconcept and to those who do not. The latter come in for very severe criticism. and their numbers include so many who are distinguished in the accident research field that it might well be regarded as an honour to be selected as a target. ~~cFarland~ Haddon Suchman and Klein (joint authors of the standard text on accident research); Conga-: Goldstein. and many others of similar distinction are triti- cized. But the disturbing feature of Shaw’s treatment is her attitude that the accident prone- ness concept is an article of faith; either one believes in it or one does not. This is not what concepts are for. A concept is a logical device for organizing one’s thoughts about a particular topic. for groupin, L 41 together a number of observations for further analysis. Research into the causes and prevention of accidents has progressed to the point where specific mriabks can be isoiated for study. and there seems no need for a blanket concept under which the effects of these specific variables can be subsumed. The “accident prone- ness” of earlier studies can now be broken down so that the eflects of age. sex, enpcri- encc, exposure to risk. stresses such as Ioss of sfcep. alcohol and drugs. long hours of work and high rates of work, and so on can be assessed in isolation and in interaction with one another. Shaw dots not seem to realise that each specific variable which is progrcssivcly itlcntificd as a contributor to accident involvcmcnt Icavcs less to bc accounted for by it nebulous proncncss f:lctor. The succt’ss with which the procuss of idcntitication has been carried out may hc il third important reason for a declining iritorcst in ilCCi&!nt proncncss.

The term accident proncncss appears to hc used in the book in two distinct ways. One dcrivcs from tllc untiisputctl fact that some pcopic hnvc more iKcidcnts thll others under the same col~diti~~r~s of risk. The fact is totally ~~r~rcr~~;lrk~~bic. since the only other possible state of the univcrsc is the much less Iikcly alternative that cvcryhody has the same number of accicicnts. In this scflsc it is quite Icgitimatc to “say that some pcoplc ilfC more prone to accitlcnrs than others. anti the complclt group of ktctors contributing to their higher involvcm~nt may reasonably. if unproli~ahly. bc I~ihellcci “accident proncncss”. I~CSCilKh

tllay then bc tlircctcd tO\VilrClS ttic isolation and control of the factors which arc contrihut- ing.

The second scnsc in which the term is usccl is much less acceptable. Accidents arc spoken of as “due” to proncncss: a proneness IktCX is said to “oporatc” under conditions 01 modcratc to high risk; tests and clinical assessments are used to dctcrminc the degree of prollcl~essc~hibited by it~cliviciuafs:pcrsttnsor groups ofpcrsons arc said to be .‘intlcrcl~tly” and s~)rl~etii~ics “inna tcly” accident prone. To WC the term in this way suggests a fitilurc to distinguish hctwccn it lab& which is convcnicnt for dcscriptivc purposes. and an CXpl~llliltiOIl. which implies sonic causal relationship.

6. C‘ONCLUSLON

This paper wits not intondcd as a rcvicw of research on accident proneness: thcrc art plenty of those avsilablc. It has dcvcloped in the writing into a rcvicw of a book --a book which rcprcscnts the major contribution. in sheer magnitude. to the literature 011 tlw topic for many years. The book is a serious attempt to re-instate accident proneness iiS an e~pl~~~l~ltio~l for a signilicant proporti~~rl of the dil~erenti~il accident i~ivolve~~~e~~t which is oftctt ohscrvcci \vitliin specific groups. It carries with it the implication that sclcction, train- ing. control and cnfbrccmcnt ilW cffcctivc methods Of reducing i tccidcnt frequency. The implication may lx. to some cstcnt, correct. but methods can ix cfTcctivc kvithout being cost clfcctivc. All the evid~ncc WC have so far tends to the conclusion that proper dcsigrl ol’ the man- machine system. with proper attention to the sources of human error, and prqwr protcctivo dcviccs for those who arc unavoidably usposcd to some risk. is more cfTcctivc and gives bcttcr value for money. Indeed. Sha~v’s own argument. that accidm proncncss can only “opcr;lfe*’ when the Icvel of risk is high. contains its own contradiction. Safety is hsst scrvcd byen~ic~lvollrji~~ to rcducc the clement of risk; with low risk the prob- km ofaccidcnt proncnoss disappears altogcthcr.

Accident proneness

REFERENCES

Farmer. E. and Chambers E. G. [ I9391 A stud) of personal qualities in accident proneness and proficient!. f/r- dusrriul Heulth Rcwarch Bourd. Rqort No. 55. London.

Frrguson. D. [I9731 -Icci&nt Pronr’l Semmar on Occupatlonul InJurIes. Royal ,Australawn Cokge of Surgeons_ Melbourne.

Greenwood, hl. and Woods. H. ht. [IYIY] The incidence of industrial accidents upon mdibldwls uith special reference to multiple accidents. Inrl. F&JW Rzs. Ed. Report No. 1. London.

McFarland. R. A. [ 19623 The epidamiolog> of motor vehicle accidents. J. .4m. rnrd. .-ts. 180, (41. XY- 300. Shaw. L. und Stchel. H. [ lY7 I] .Icci&nt Prortrntw. Resuorch in the Occrtrrt~~we. Cmt.wfrort urrd Prtwr~rr~~r~ I)/’ Rtwd

Arcidtwts. International Series of Monographs in Experimental Psycholog. Volume I I. Prrgamon Pras. Oxford.

Abstract-Review oF research on accident proneness have appeared nt intervals for man> )-ears. In the past 20 yzars or so. reviewers have concluded without exception that the contribution or the accident prone to the total accident problem is small. and attempts to control accidents by attention to man mxhine design and the provision of approprixtc safety equipment are more effcc- rive than attempts to identify- those people with a high swsceptrbrlity to accidents. r\ recent book presents the case for the latter approach at great lenpth and with considerable force. This paptx ermines the arguments prewntcd and concludes that they are not well tiiundcd.