Journal of Hazardous Materials 142 (2007) 730740

sl s

006

Abstract

One of th envculture. Typ regulexhortations netwas also exte exteetc., which influerosion in s uteaccident ana (Curdent analysi a), H(Canada) an ed crinsights: the biggest source of motivation that causes deviant behavior leading to accidents is Faulty Value Systems. The second biggestsource is Enforceable Trust. From a management control perspective, deterioration in safety culture and resultant accidents is more due tothe action controls rather than explicit cultural controls. Future research directions to enhance the models utility through layering are addressedbriefly. 2006 Else

Keywords: Sa

1. Introdu

The saindividuterns ofstyle anprogram

Man issocial b

This pacharacteristhe valuesmakes theHuman Faalmost impthe social c

E-mail ad

0304-3894/$doi:10.1016/jvier B.V. All rights reserved.

fety; Social factors; Social capital; Socio-management; Accident

ction

fety culture of an organization is the product of theal and group values, attitudes, competencies and pat-behavior that determine the commitment to, and thed proficiency of, an organizations health and safetymes [1]., at one and the same time, a solitary being and aeing Albert Einstein [2].radoxical human nature of retaining individualtics but still getting dynamically influenced by, thoughts, feelings and actions of those around,

management of safety culture and the study ofctors in organizational safety challenging. It isractical to remove a person from the influences ofontext in which he (she) exists and study his (her)

dress: sumanashokrao@yahoo.co.in.

motivations and behavior in isolation. Any attempts to finduniversal laws governing human nature seem very difficult toformulate and almost an improbability. We have hardly beenable to formulate any laws of human nature which are of thesame rigor and standards as the laws of nature which havebeen formulated in the hard sciences. . . [3]. But it is preciselythis elusive human nature, which determines the organizationalsafety culture to a large extent, that organizations striving toachieve the highest workplace safety standards are compelled tomanage.

As is evident from the above, any solution to the above organi-zational challenge of establishing and sustaining a robust safetyculture would cut across the fields of both management and soci-ology. Hence safety culture, as well as accident analysis method-ologies that inform prudent safety management, both needto be approached from a socio-management perspective. Thispaper discusses one of the many possible socio-managementapproaches to safety culture and accident analysisthroughmanaging the safety social capital inherent in organizationalnetworking relationships. It also discusses and demonstrates the

see front matter 2006 Elsevier B.V. All rights reserved..jhazmat.2006.06.086Safety culture and accident analysiapproach based on organizationa

Suman RaoRisk Analyst, India

Available online 28 June 2

e biggest challenges for organizations in todays competitive businessically, the key drivers of safety culture in many organizations areto comply with safety norms, etc. However, less evident factors like

nded networking relationships and social trust of organizations withconstitute the safety social capital in the Organizationseem to alsoafety social capital cause deterioration in safety culture and contriblysis models do not provide answers to these questions, CAMSoCs model, is proposed. As an illustration, five accidents: Bhopal (Indid Exxon Valdez (USA) have been analyzed using CAMSoC. This limitA socio-managementafety social capital

ironment is to create and preserve a self-sustaining safetyation, audits, safety training, various types of employeeorking relationships and social trust amongst employees,rnal stakeholders like government, suppliers, regulators,ence the sustenance of organizational safety culture. Canto accidents? If so, how does it contribute? As existingtailing Accidents by Managing Social Capital), an acci-yatt Regency (USA), Tenerife (Canary Islands), Westrayoss-industry analysis provides two key socio-management

S. Rao / Journal of Hazardous Materials 142 (2007) 730740 731

applicability of CAMSoCa socio-management accident anal-ysis model.

1.1. Huma

There arin understaaccidents. Gattributed tof researchHuman Faclearning intels have beaccidents, eFactors Anpell and Wiinvolved inTheoretic A

These macts) in acaccident sption in theprocesses a

But theeasily predcircumstanmay reacting under pnature, discidentified pFor examplacunae in tantee thatwould not

Some ofof knowledpart of an eerror (devinorms or thety) [5]. Therrors at a hhuman behsafety?

1.2. Needsociologica

Assumiknowledgeing to dealThis behavnetworkedsocial normthe devianttraced to foor groups obe? For ex

were magnified due to refusal of Union Carbide to shift the MICfacility to a government designated industrial area for hazardous

als. Ie an

instepubhavetuatim a

ociol

appis nos havle opy citiolol ac

the manizgicanizaLessEssch ta, sucs ant shalysal ccru

righat soug? Ift parlse c

fortoryls, cry ped exl. Thzatiocticeow eyeesgularsuashed

ork

ocio

rdernetwemen Factors analysis models

e many evolving approaches that help organizationsnding Human Factors relating to safety culture andenerally 6080% of normal accidents are said to be

o Human Factors [4]. Over the past few decades, a lotefforts and resources have been invested in analyzingtors associated with accidents and incorporating theo workplace safety promotion initiatives. Many mod-en established that discuss human factor analysis in.g. Reasons Swiss Cheese Model, HFACS (Human

alysis and Classification System) developed by Shap-egmann, Classifications of Socio-Technical Systemssafety control by Rasmussen, STAMP (Systems-ccident Model and Processes) of Leveson, etc.odels analyze the root causes of human error (unsafe

cidents: tracing the path from active failures at theot, all the way up to latent defects in the organiza-form of improper management of people, systems,nd culture (Organizational influences).fact remains that human nature and behavior is notictable. Employees may react differently to similarces in a working environment. The same employeedifferently to the same circumstances when work-ressure. Given this element of dynamism in humanrete organizational steps taken, as a response to theast latent defects may not be sufficiently effective.le, if a particular human error was traced back tohe selection and placement process, there is no guar-a person selected as per the new improved processcommit this or some other related new error.the key contributors of human errors/factors are lackge and/or lack of skill and/or deviant behavior on themployee (or a group of employees) committing theant behavior is behavior that does not conform toat does not meet the expectations of a group or soci-is leads us to the question: should we analyze humanigher level of granularity or focus on specific deviantaviors to make a quantum impact on organizational

for extending the scope of accident analysis intol domains

ng safety processes and training can address lack ofand lack of skill, what remains immensely challeng-

with is deviant human behavior that causes accidents.ior could be on the part of an individual or a group ofindividuals. Quite evidently, there must have been athat the individuals (groups) did not internalize for

behavior to result. Can this lack of internalization bermal or informal inter-relationships between peoplef people that did not work the way it was meant to

ample, the consequences of the Bhopal gas tragedy

materiCarbidtrust (est forwouldThis sitors fro

1.3. S

Thesafetyologistpossibquentlby socnorma

givenare orgsocioloto orgabook tralianapproafactorsfailureto meethis antechnic

Thefactorssuch ttheir thcultureare juswhat e

Theregulamanua

the vefelt, anmanua

organiIn praonly hemplowith reThe peestabliin his w

2. A s

In otionalmanagf there had been a strong relationship between Uniond the Government of Madhya Pradesh, bonded byad of lobbying) and motivated by a common inter-lic safety (instead of profiteering/personal gains), it

led the plant to be sited in this less populous area.on, then, necessitates examining accident causal fac-sociological perspective.

ogy and organizational safetylication of sociological theories in organizationalt new. In the context of organizational safety, soci-e hitherto been primarily concerned with evaluatingerator (human) errors and nonconformities. A fre-

ed reference in sociological circles is the work donegist Charles Perrow on normal accidents (the termcidents refer to those failures that are inevitableanner in which human and technological systems

ed) [4]. Attempts have also been made to apply otherl concepts, e.g. Banduras Socio-Cognitive Theorytional safety [6]. Sociologist Andrew Hopkins, in hisons from Longford (2000) has analyzed the Aus-o accident causes further up the social chain. Hiskes the reader through the full range of contributingh as management system inadequacies, regulatory

d even the cost cutting pressure driven by the needareholder dividend demands. The principle behindis is that if the organizational factors are right, theauses of accidents will not come into play [7].cial question then is, how to get the organizationalht? How does one manage organizational featuresafety becomes the second nature of employees inhts, action and behavior and results in a healthy safetyone were to agree that regulation, audits and trainingt of the solution and not the entire solution itself, thenould be done to narrow the gaps in solutioning?mal safety norms/features are documented incompliance requirements, process and training

odes of conduct and employee role definitions, butrsuasive informal safety norms/features have to beperiencedthey cannot be easily documented in anyey reflect the collective thinking and behavior of then and are the organizational safety values in action., this collective thinking and behavior affect notmployees relate to one another but also affect howrelate to people external to the organization, e.g.tory and inspection authorities with respect to safety.siveness of collective thinking and behavior is wellby the eminent sociologist Durkheim-particularlyon social facts [8].

-management framework for safety culture

to manage such intangible but yet potent organiza-orks and relationships, we need a focussed socio-

nt framework.

732 S. Rao / Journal of Hazardous Materials 142 (2007) 730740

The ideal socio-management framework for safety will con-sist of:

(1) a theorageme

(2) an acti

Organizsafety valuexhorting tever, increanot the panthe relationto organizinvestmentdistributedthe hope ofa focus onthat organiof social c

2.1. Organsocio-mana

Definitithe social cing Alonethis concepsocial netwworks to doof social orthat facilita[10].

It followof safety, iencouragedto favorabland cooperachieve the

2.1.1. SafeIn order

one needsBased on aworks genecapital wel

Help focstakeholture. A rlarge ext

Facilitatees andstockholwell as idegree o

Determine and positively increase the extent to which employ-ees bond with each other and do things for each other in the

restsble ecienc

otetiontion

rief rere ichescludpos

s fouanceive msociit fdispcultu

n acing

ial nral gies o

hilally,ieve

on d) to

ngfue morks a

Mottes [tic anvaluend 13].

aluers (quiroundroupon fmpa

ationuses

ecipchaey eate hetical foundation with roots in sociology and man-nt,onable framework for managing social capital.

ations have traditionally tried to generate and sustaines in their employees through training interventions,he employees to comply with safety norms, etc. How-singly organizations are finding that these alone areacea for all problemsthey also need to focus onship between individual employees as a key driver

ation wide culture-change initiatives. Tremendouss are made in collaborative, networking technologies,databases and knowledge management systems withachieving better safety standards and practices. Suchthe interrelationships between employees demands

zations look more closely at the sociological conceptapital.

izational safety social capitalagement theoretical foundation

ons of social capital abound, but this paper is based onapital premise of Robert Putnam, whose book Bowl-(2000) contributed considerably to the popularity oft. Social capital refers to the collective value of allorks and the inclinations that arise from these net-things for each other [9]. It consists of the features

ganizations such as networks, norms and social trustte coordination and cooperation for mutual benefit

s that if organizational social capital in the context.e. safety social capital is managed, developed and

in the organization, we can indeed look forwarde organizational features that facilitate coordinationation amongst employees and other stakeholders tohighest standards of workplace safety.

ty Social capital and safety cultureto manage organizational safety social capital better,to first have insights on how social capital works.n application of Putnams views on how social capitalrically [11],it emerges that managing safety social

l can:

us on the relationship between employees and otherders in the interests of establishing a safety-first cul-elationship based on trust and high safety values to aent determines the inherent safety in the system.e safety related information flows among employ-

other stakeholders including regulators anddersthese information flows would be formal asnformal and will in turn depend on the structure andf relationship.

inte Ena

defi Prom

nizacrea

A bthat thapproaally, inwith acationimportventatsafetyabove,very insafety

2.2. Amanag

Soca natustrateg[12]. Winformnot achcohesi(actorsmeaniage thnetwo

2.2.1.Por

altruistwo: mer; alatter [

(i) Viore

(ii) Bgm

co

u

ca

(iii) Rex

thipof safety.mployees to take collective action if they find safetyies.a sense of solidarity amongst everyone in the orga-in ranking safety as firstthis will facilitate theand sustenance of a safety first culture.

eview of current literature in Safety Culture suggestss no consistent definition of Safety Culture. Many

to safety culture are being discussed internation-ing socio cognitive approaches [7]. Organizationsitive safety culture are characterized by communi-nded on mutual trust, by shared perceptions of theof safety, and by confidence in the efficacy of pre-easures [1]. Considering some of the ways in which

al capital can work for an organization as discussedollows logically that safety social capital is indeedensable to Organizations in order to have a positivere.

tionable socio-management framework forsafety social capital

etworks (in which social capital resides) are notiven and must be constructed through institutionalriented to the institutionalization of group relationse network relations can be created formally andmere authority upon the networking actors can-their sustained continuation. The sustenance of this

epends on the motivation of the networking playersbind together. It follows that if one is looking at anyl management of safety social capital, one has to man-tivational sources that bind the socio-organizationalnd the associated social capital first.

ivational sources of safety social capital12] classified sources of social capital into two types:d instrumental and each is further broken down intointrojection and bounded solidarity for the for-

simple reciprocity and enforceable trust for the

Introjection refers to internalized norms and behav-e.g. people willingly abiding by safety reportingements, preventive maintenance schedules, etc.).ed Solidarity refers to identification with ones own

, sect or community. It is an emergent product of com-ate (e.g. worker unions, from two different chemicalnies, by being thrown into a common hazardous sit-, learn to identify with each other and support theof each other).

rocity Exchange refers to a generalized system ofnge in which social actors help others not becausexpect immediate repayment, but because they antic-elp being extended to them at some future date (e.g.

S. Rao / Journal of Hazardous Materials 142 (2007) 730740 733

sharing safety alerts and best practices amongst a commu-nity of nuclear power producers across the globe).

(iv) Enforcbecauto law(e.g. athe ma

Thoughnotation ofin this paplectivism. Amerits of tscope of th

2.2.2. ThePutnam

positive inside to it toagainst regtion [14]. Sinclude: nethe group,the networexistence oessarily meof Strongculture wilthe underlysafety equiof the primin Canada.based on ain for eachExxon Valdfor an alcounderline thin the orga

2.2.3. Manoriented co

The depnetworks itrust. Thissocial capion relationculture throvation forcommon inbalances (cfrom behavented contr

The objachieve gotrol is obvioimpossibleantee goodcontrol wh

that no major, unpleasant surprises will occur [15]. In the inter-ests of achieving good control, Merchant [15] proposed four

ontroew oh conr to ttion

tionionaviostra

ties.Prds,physrovacessiew.ckliuirers totrol

rsonnininguiretestspertrol

iningcon

cificnicajob dlturaed rand

exam

entivns)ry eyea

contsultpro

form

rderes a

undrioraana

e CAcapi

caus

ideneable Trust refers to trust that exists preciselyse obligations are enforceable, not through recourseor violence but through the power of the communitycompany may voluntarily withdraw its drugs fromrket fearing community reprisals).Portess model is based on an individualistic con-social capital unlike Putnams collectivist approach,er, Portess model has been modified to reflect col-

detailed discussion regarding the differences andhe two approaches to social capital is beyond theis paper.

importance of managing negative social capitals interpretations of social capital have been largelynature. However social capital has a potential down-o and authors like Portes and Landolt have cautionedarding social capital as a cure all for every situa-ome of the negative aspects attributed to social capitaltworks strong in social capital excluding others fromrestricting the freedom of certain individuals withink who may not wish to conform, etc. Also, meref strong social capital within a network need not nec-an benefits accrue to the memberse.g., existence

safety social capital that supports a safety firstl not translate into any use for the organization ifing technology is unsafe, e.g. the grossly inadequatepment and gear provided by the mine owners was oneary causes for the Westray coal mine disaster (1992)A high social capital present amongst colleagues

strong social network may lead colleagues to stand-other, even in deviant behavioras it happened in theez case where tired, overworked colleagues stood inholic captain. The potential upsides and downsidese need for safety social capital to be managed well

nizational context.

aging Social capital through behaviorallyntrolsletion/malfunctioning of formal and informal socialn organizations leads to an erosion of norms andinvariably leads to depletion of the underlying safetytal and affects safety adversely. Networks are baseds and relations are dynamic. Consistent safety-firstughout the organization requires, apart from moti-

formal and informal groups to bind together in theterests of safety, a method of installing checks andontrols) that prevent the dynamic networks/groupsing deviantly. This necessitates a behaviorally ori-ol system.ective of a behavior-oriented control system is tood control and not perfect control. Perfect con-usly not a realistic expectation because it is virtuallyto install controls so well designed that they guar-behaviors. What is more realistic is to aim for good

ere an informed person can be reasonably confident

main coverviof eacto refedescrip

i. Acactbehcon

duwo

ofappprorev

chereqnee

con

ii. Petrareqityjobcon

traingspemu

ofiii. Cu

basingan

incPlaeve

in 3of

iv. Reees

per

In opreservtant toto detesocio-m

3. Thsocial

Allan accls: action, personnel, cultural and result controls. Anf what constitutes these controls and a few examplestrol class is provided below. Readers are requested

he book as cited Ref. [18] for further information ands.

controls comprise of behavioral constraints, pre-reviews, action accountability and redundancy:ral constraints include constraints like physical

ints, administrative constraints and separation ofoka-yoke type of process constraint, Computer pass-

restricted access to certain hazardous areas are formsical /administrative constraints. The necessity to seekl of shift supervisor when following a non-standardor a process-deviation is an example of pre-actionA system of maintaining compulsory on-the-job

sts is a form of action accountability control. Ament to maintain mirror sites/stand-by service engi-

reduce system downtime is a form of redundancy.el controls comprise of selection and placement,, job design and provision of necessary resources:

ment for interview candidates to undertake personal-and then map these personality traits to the desired

sonality profile is a form of selection and placement. Requirements for employees to undergo refresher

at pre-determined intervals are an example of train-trols. Requiring each role in the organization to haveally identified KRAs (Key Result Areas) and com-ting the same to role-holders in writing is an exampleesign controls.l controls comprise of codes of conduct, group-ewards and inter-organizational transfers: maintain-

reinforcing a written code of ethics at work isple of code of conduct control. Apart from team

es, schemes like ESOPs (Employee Stock Optionare examples of group-based rewards. Requiringmployee to have a change in job-profile at least oncers is an example of intra-organizational transfer moderols.controls aim at maximizing the chances of employ-ducing the results the organization desires: pay-for-ance is a prominent example of results control.

to install a system of organizational controls thatnd enhances safety social capital, it is first impor-erstand how erosion in safety social capital can leadtion in safety culture and accidents. This requires agement approach to accident causal analysis.

MSoC (curtailing accidents by managingtal) model

al factors, including technical factors, identified int causal analysis have the potential to get examined

734 S. Rao / Journal of Hazardous Materials 142 (2007) 730740

further for the overlying social factorsafter all technology isalso created and used by humans. This underscores the need forapplying sociological concepts in cohesion with managementprinciples in accident analysis-in other words, using a socio-management approach to accident analysis.

Such an analysis needs a critical evaluation of Social Factorscontributing to accidents, which in turn necessitates a frameworkthat

should to be valid in different accident scenarios and in dif-ferent industries,

should have a basis in sociological/socio-psychological theo-ries as well as serve as a practical tool for industrial accidentanalysis.

The framework would typically comprise of the followingelements: the social actors involved, the relationships amongthem and reasons for their deviant behavior that led to theaccident. The output of such a framework would be to iden-tify motivation source type problems and specific organizationalcontrols that broke down and led to accidents. This would thenhelp inform the actionable socio-management framework formanaging organizational safety social capital as discussed inSection 2.2. In all, such an approach to accident analysis woulddeliver value to society by helping to systematically extractsocial lessons learned from accidents, which information, inturn can badvancing

3.1. Lendi

Overlytem that pr

The omnipresence of social factors overlying all organizationalsystems and situations including accidents is well evident. Thescope of the term social factors is very wide in sociology. Inthis paper social factors would mean and include all aspects ofthe organizational actor groups environment that involve morethan one person or that has been, in the past the product of peo-ple [17]. It follows that if one were to analyze accident causalsocial factors in greater depth, then one needs to look at whatwent wrong with the underlying organizational socio-featureslike networks, norms and trust in the safety contextin otherwords how and why was the organizational safety social capitaleroded. This entails focusing on deviant behavior (as discussedin Sections 1.1 and 1.2) that resulted in erosion of safety socialcapital, thereby leading to the accident. The next question iswhy the actor groups, especially the Key Actor Group indulgedin this deviant behavior. The process through which we seeksuch information is called attribution in social psychology.Many Social psychologists have contributed to theories on attri-bution, e.g. Graham and Folkes, Heider, Pittman, etc. However,this paper, deals with the question: did others behavior stemmainly from internal causes (their own traits, motives, inten-tions); mainly from external causes (some aspect of the socialor physical world) or from a combination of the two? [18]. Indeveloping the CAMSoC model, the internal causes for deviantbehavior of the Key Actor Group is analyzed by seeking the faultin the source of motivation and the external causes for deviant

or iszatio

diszatiotionstep

layeree utilized by organizations and regulators alike fortowards a better safety culture.

ng a structure to the CAMSoC model

ing every technical or civil system is a social sys-ovides purpose, goals, and decision criteria [16].

behaviorganiifest. Aorganiin Sec

Theare:

Fig. 1. A pictorial representation of the singleanalyzed by seeking to identify the break-down ofnal controls that allowed the deviant behavior to man-cussion of the implications of motivation sources andnal controls for social capital has already been made2.2.s followed in the CAMSoC accident analysis (Fig. 1)

d CAMSoC model.

S. Rao / Journal of Hazardous Materials 142 (2007) 730740 735

(i) Identify the accident causal factorsthis could be from var-ious sources ranging from primary accident investigationreportsdataba

(ii) Once tof acciand So

(iii) The Kdominaeach S

(iv) The sovationais idenattribuered.

(v) Failurecontrolin comled to tattribu

The abocomplexitiwere to inc major andmotivationmajor and mmodel, suitthen the soin terms ofzational colayered morequired to

Social Sthe classificor minor, ththe classifione accideconsistentin developiclassificatio

In thisanalysis hareferencesany authorof these clato demonslessons usident causinin past accthe thorougmore sociain this studtreated as ainvestigatio

In live swould be u

lators/judiciary to help identify the Key Actor Group(s) respon-sible for the accident, provide insights into their motivations,

so identify the failure in specific organizational controlsd tos anzatio

io-mCAM

rdercapitcon

re an

eogvarionts isMS

nts.ed bynal s

brie

ase nintroeader to

Themidfaced0 tonromillinallyil anse fg Tad thuencis ue cuany

tentlattituafetyC.

Thehe eaprommen

nt wcaus

so chand interviews to secondary accident informationses, press reports, etc.he causal factors (either human or technical factors)dents are identified, their social context is examinedcial Factors(SFs) are thus extracted.ey Actor Group (the group which contributed pre-ntly to the decline in social capital in the context of

F, by indulging in deviant behavior), is identified.urce of motivational problem (using Portess moti-l classifications) for this groups deviant behaviortified. If there is more than one motivational sourcetable, the predominant motivational source is consid-

s of controls are mapped from a behaviorally oriented(Merchants organizational control model), which

bination with the motivational problem had actuallyhe accident. If there is more than one controls classtable, the predominant control class is considered.

ve is a basic single layered model of CAMSoC. Thees with the CAMSoC modeling will increase if onelude more layers in the analysis e.g. all actor groups

minor-for each causal factor in (ii) above, all causalal sources in (iii) above and all causal control failures,

inor in (v) above. In such a multi-layered CAMSoCable aggregation methods have to be employed andcial lessons from the accidents have to be extractedidentified flawed motivational sources and organi-

ntrols. In this paper, however, only the basic single-del of CAMSoC is elaborated. Further research isdevelop such a multi-layered CAMSoC model.ciences are not exact sciences. It is quite possible thatation of the Key Actor Group as in (ii) above as majore source of motivation problem as in (iii) above and

cation of controls as in (iv) above may differ fromnt investigator/analyst to another. In order to ensureresults from the model, further research is requiredng templates that would formalize the process of thisn and reduce subjectivity to a reasonable extent.

paper, such classifications in the ensuing accidentve been made by the author of her own, based onas cited and have not been independently verified byity. Readers may please be aware that the purposessifications and the consequent accident analysis is

trate the utility and possibilities of deriving socialng the CAMSoC model and not to fix/assign acci-g responsibilities on any of the Key Actor Groups

idents. Although the author has taken every care forhness of the detailing, it is possible that there are

l factors existing than those that have been identifiedy. The accident analysis as found in this paper may ben illustration of the models applicability to accidentns.afety/accident investigations, the CAMSoC modeltilized by the accident investigation teams and regu-

and althat leinsightorgani

4. Socusing

In osocialcausedity, wewere gcover

accidethe CAaccidefollownizatio

4.1. A

Plea brieftive. Rin orde

4.1.1.At

Indiatory. 4#610 ftory, kphysicthe soate cauenterinruptureconseqBut aswas thsown m

consisligentbasic sthe MI

4.1.2.In t

mine,all 26accidesparkswas althe accident. This in turn would be useful to provided inform socio-management approaches to betteringnal safety.

anagement analysis of ve major accidentsSoC

to lend practical insights into the subject of safetyal, five major accidents from different sectors, whichsiderable damage to life and property in their vicin-alyzed. Care was taken to ensure that these accidentsraphically distributed across the world in order tous populations and cultures. A brief outline of thegiven below, followed by an illustration of the how

oC model is applied to causal social factors of theseResults arising out of this analysis are presented,

a summary of socio-management learning for orga-afety from the CAMSoC analysis.

f description of the accidents [19,20]

ote that the description given below is only aimed asduction to the accident and are by no means exhaus-rs are requested to refer to the sources as mentionedobtain more information about the accidents.

Bhopal Gas Tragedy [2123]night of 2 December 1984, the city of Bhopal inthe worst ever disaster in chemical accident his-

nes of Methyl Isocyanate (MIC) escaped from Tankan Insecticide (Sevin) making Union Carbide fac-

g more than 10,000 people and leaving thousandsand mentally impaired. It did everlasting damage tod ground water in the plant vicinity. The immedi-or the disaster was MIC release triggered by waternk #610 and the subsequent pressure build-up thate tank. The water itself had entered the tank as ae of a supervisor ordered water-washing operation.

sually the case in many accidents, this accident toolmination of a series of factors, whose seeds wereyears ahead and whose growth was encouraged by

y inferior and inadequate safety practices. The neg-de of Top Management and the telling absence of aculture were responsible for the disaster as much as

Westray coal mining disaster [24,25]rly morning hours of 9 May 1992 the Westray coaloted by Curragh Inc., in Nova Scotia blew up killingworking underground. The immediate cause of the

as a methane explosion most probably ignited due toed by a continuous miner. Much like Bhopal, Westrayaracterized by an utter lack of safety culture.

736 S. Rao / Journal of Hazardous Materials 142 (2007) 730740

4.1.3. The Tenerife Air Tragedy [26,27]On a foggy early evening of 27 March 1977 two Boeing

747sPan Am 1736 and KLM 4805 collided on the runwayof Los Rodeos Airport on the Tenerife Island. Five hundredand sixty people lost their lives, over 60 were injured and theairplanes were completely charred. The immediate cause of theaccident was the impact of KLM 4805 on the Pan Am 1736 duea take-off operation by the KLM captain, without obtaining thefinal clearance from ATC. That non-compliance, catalyzed by ageneral communication gap and an unexpected thick fog finallyresulted in the air tragedy.

4.1.4. The Hyatt Regency Walkway collapse [2831]On the 17 July 1981, a dance party turned into a nightmare

for the 2000-odd guests assembled at the Hyatt Regency Hotel,in Kansas City. Over 114 people were killed and at least 200injured when the second and fourth level walkways crashed.The immediate cause of the accident was the split of the boxbeam and the support rod pulling through, causing the secondand fourth level walkways to collapse. Unplanned doubling ofthe structural load onto the box-beam supporting the fourthfloor due to an unauthorized change in design and a faultyoriginal design were the prime factors that made the collapseinevitable.

4.1.5. TheOn 24 M

Bligh reef acrude oil into the fishermore thanthere werewere lost dwas the inasteer rightfactors of a

to one of the worlds worst oil-spill disasters despite a voyage-friendly atmosphere throughout.

4.2. Applying the CAMSoC model for accident analysis

4.2.1. Methodology adopted for accident analysis usingCAMSoC modeli. The social context of each of the causal factors was identi-

fied and captured as SFs accident wise. In this, the followingwas used as a general rule for analysis: examine and tracethe causal factor up to the point when more than one personis involved and collective thinking/relationships with otheractors or the lack of it is detected. This gives the causalsocial factor. As an illustration (see Table 1), in the caseof the Hyatt Regency, inadequate and faulty design was akey cause for the collapse of the walkway. Whereas thiscould have been traditionally traced to factors like a fail-ure of design supervision, faulty design engineering skills orplain lack of communication between the two actor groups,viz. the Design Engineering Firm and the Fabricator, a soci-ological perspective suggests that this be seen from a net-work/relationship point of view where the role of more thanone actor comes into play. Consequently, if one were to trace

faulups

t to bsigne nerks/nric tthis

ne onis ofin th

m andes

Table 1Illustration of

Social factor:Key Actor Gr is cas

taininthe f

ing prrigina

The motivatio ar asety hable enther thack oect coriginding cneerin

Behavior orie actionaps. T

tt RegkingbegaExxon Valdez oil-spill disaster [32,33]arch 1989 the tanker Exxon Valdez piled into the

nd spilled some 10.8 million gallons of North Slopeto the seas. This act caused immense long-term lossesies, tourism and the sea ecology apart from fouling up

1000 miles of beach in south central Alaska. Thoughno deaths directly as a result of the accident, four livesuring the clean up operation. The immediate causedequate response of the ship to a course change (a

) command given by the Third Mate. The underlyingn alcoholic Captain and crew change violations led

thegroou

Deii. Th

wo

fabfordobaspleFirthe

CAMSoC model applied to a causal SF of Hyatt Regency

lack of design ownership between fabricator and design engineering firmoup Design Engineering Firm (DEF) In th

mainwithmakof o

n source type problem Enforceable Trust As fsocireliawhethe lprojthe oBuilengi

nted control type failure Pre-action reviews Pre-mishHyachectheyty design to the point of relationships between actorin the Hyatt Regency case, the social factor emergese lack of design ownership between Fabricator andEngineering Firm.

xt step is to analyze what went wrong with the net-orms/trust that was supposed to hold the safety social

ogether and which Key Actor Group was responsibledepletion in safety social capital. This analysis wasthe basis of either contractual obligations or on thegenerally accepted normative principles. For exam-e Hyatt Regency case, both the Design Engineeringd the Fabricator were responsible for not aligning

ign with the fabrication. However, contractually, the

e DEF contributed to a decline in social capital by notg appropriate formal or informal channels of communication

abricator (a collapse of extended networking relationships), andesumptions that the fabricator would have fabricated on the basisl design (collapse of misplaced trust)the DEF was concerned, it violated the trust the owners andd placed in them as professionals with a license to delivergineering design. In a multi-million dollar contract, not checkinge Fabricator was delivering as per the design speaks volumes of

f care and diligence with which the DEF had approached thenstruction. To add to this, even had the fabricator delivered as peral design, the design by itself was in violation of the Kansas Cityode-again a violation of trust placed by society in a professionalg firmreviews are an important form of control for avoiding suchhis should have been applicable at least at two levels-owners of

ency independently validating construction design and DEFif the fabricators are proceeding with the right design beforen the actual fabrication

S. Rao / Journal of Hazardous Materials 142 (2007) 730740 737

Table 2Bhopal-Key Actor Groups SF wise

Key Actor Group No. of SFs

Bhopal (see Chart 1)Top management 29Plant management 14Design engineers 5Regulators/law makers 2Total SFs analyzed 50

licensed Design Engineering Firm was expected to validatethe designs and take ownership for the designs-hence thismakes

iii. Then thbehaviomotivatSectionclassific

iv. Finallyof deviaof orgafor furt

See Tab(iii) and (iv4.2.2. AppWalkway c

An examparticular s(Table 1); lthe other 12in the same

Causal fof the Hyadispute witricator had

5. Resultsorganizati

In all 1accidents usummarize

5.1. Insigh

A summthe five acc

Table 3Westray-Key Actor Group SF wise

Key Actor Group No. of SFs

Westray (see Chart 2)Top management 14Plant management 7Coal miners 1Regulators/law makers 8Total SFs analyzed 30

Table 4Tenerife-Key

or Gr

(seedesigncontroanagcaptain 3crew 4

M crew 2SFs analyzed 16

ey Actor Group SF wise

or Group No. of SFs

ee Chart 4)n engineer 5cator 1anagement 5

Miss investigators 1SFs analyzed 12

aldez-Key Actor Group SF wise

or Group No. of SFs

aldez (see Chart 5)in 11mate 1

crew 1anagement 2

SFs analyzed 15

lmost all the accidents (except Tenerife) (Charts 15),st frequent disturbance to safety social capital is causedlty Value Systems (FVS) of the Key Actor Group. Thiswed by (violations of) Enforceable Trust (ET), lopsidedocity Exchanges (RE) and misplaced Bounded Solidarity

Table 7CAMSoC Re

Motivational

Value IntrojecBounded SoliReciprocity EEnforceable TTotal motivatithe DEF the Key Actor Group for this social factor.e source of motivation for this actor groups deviantr was examined and categorized (based on Portessional sources of social capitalsee discussion in2.2.1 for further information on source of motivationations).the specific failure of controls with identified typesnt behavior was mapped using the Merchant model

nizational controls (see discussion in Section 2.2.3her information on control classifications).

le 1 for an illustration of the reasoning underlying the) classifications.lication of CAMSoC model in the Hyatt Regencyollapsean illustration

ple of how the CAMSoC model was applied to oneocial factor of Hyatt Regency accident is illustrated inimitations of space prevent an elaborate discussion of2 SFs across all the five accidents that were analyzedmanner.

actor. Inadequate and Faulty Design was a root causett walkway collapse. This was later a point of legalh the Design Engineering Firm claiming that the Fab-never sought their approval for design modifications.

of the accident analysis and implications foronal safety

23 SFs (Tables 26) were analyzed across the fivesing CAMSoC. Key findings from this analysis ared below.

ts on motivation sources

ary of the motivation source type problems acrossidents is given in Table 7.

Key Act

TenerifeATCATCTop mKLMKLMPanATotal

Table 5Hyatt-K

Key Act

Hyatt (sDesigFabriTop mNear-Total

Table 6Exxon V

Key Act

Exxon VCaptaThirdShipTop mTotal

In athe moby Fauis folloRecipr(BS).sults-Summary of Motivation Source Type Problems

Source Type Problem(Single Layered CAMSoC) Bhopal Westraytions (Faulty Value Systems) 20 12darity 3 4xchange 9 4rust 18 10on source type problems 50 30Actor Group SF wise

oup No. of SFs

Chart 3)ers 1llers 4ement 2Tenerife Hyatt Exxon Total

5 5 7 494 6 17 4 177 3 2 40

16 12 15 123

738 S. Rao / Journal of Hazardous Materials 142 (2007) 730740

Chart 1.

Chart 2.

Chart 3.

Motivatsociologications and R

Chart 5.

motivation source type problems, as obtained in the above man-ner, helps set regulatory and management priorities when dealingwith issues relating to safety culture.

A graphical analysis based on the CAMSoC model is pre-sented in Charts 15.

5.2. Trends in failure of behavior oriented controls

A summary of failures of controls is given in Table 8.Within the scope of this accident analysis, the following are

the control related observations (Table 8) that have significantsociological and management related implications:

On an aggregated basis, failures of explicit cultural con-trols are only on 15 occasions out of a total of 123 SFs.Especially in accidents like Bhopal and Westray where the

rioraof e, re

Table 8CAMSoC Re

Failure of con

Action controBehavioralPre-actionAction acc

Total action

Personnel conSelection aTrainingJob design

Total perso

Cultural contrCodes of coGroup base

Total cultur

Result or outpTotal controldetetion17%Chart 4.

ional problems of Key Actor Groups have significantl and management implications for both Organiza-egulators alike. The frequency-based ranking of the

of contremphasitrols (thocontribuorganiza

In almoaction co

sults-Summary of Failure of Controls Classes

trols class(Single Layered CAMSoC Model) Bhopal Westraylsconstraints 11 3

review 10 2ountability 7 3

control failures 28 8

trolsnd placement 2

3 13

nnel control failures 3 6

olsnduct 5 3d rewards 2

al control failures 5 5

ut controls 14 11failures 50 30tion of safety culture was very evident, the propor-xplicit cultural control failures were low 10% andspectively as compared to failure of other formsols like Action Controls and Result Controls. Thiszes the role of other forms of organizational con-ugh not recognized explicitly as cultural controls) in

ting to the deterioration of the safety culture in thetion.st all accidents (except Westray (27%)), failure ofntrols seems to account for the highest proportion of

Tenerife Hyatt Exxon Total

5 3 222 2 16

5 7 22

7 7 10 60

1 37 1 121 2 6

8 2 2 211 2 2 132

1 2 2 15

1 1 2716 12 15 123

S. Rao / Journal of Hazardous Materials 142 (2007) 730740 739

control failures. Bhopal (56%), Tenerife (44%), Hyatt (58%),Exxon (67%). This emphasizes the need for well-establishedformal opre-actioform a2.2.3).

Faulty Vsource o

respondirate of cof conduvalue syonly a mtural conexplore

6. Conclu

There awhile workcusses a s

safety cultvide valuabmanagemeing Accidemethod anthe erosionand the Keythe combincontrols thaanalysis wadents: Bhoinsights obvation andsummarizemanagemenorms andsafety socisafety cultu

Acknowled

I thank tance on theto the infinelder Broth

Reference

[1] S. Gadd,of safethttp://ww

[2] A. Einsteers, 2003

[3] R.R. SarEaston a

[4] R.T. Sch611.

[5] D.E. Hebding, L. Glick, Introduction to SociologyA Text with Readings,Newbery Award Records Inc., 1987, p. 405.

[6] M.D. Cooper PhD, Towards a Model of Safety Culture. An AbridgedrsionessedSafeicol,

d Displosiolicy ULukesSociop://ww

. Putpubli. Put

vided. Pu

ough/soc

Portenu. SoYanagelopmerica.Porte

n Prn andwInte

. MersonLeveety S

p://suntrodu2.ccc. Bacholo3.

Rao,.), Pr

Accid. Ru

nagemD. Chevier,se No

:\rmBrownw.bhss Acationsite/anscritia.estrayt info

aymenRantcraft68, ateneriers Rm httpurseNeilault.hrganizational relationships that define behavioral,n review and action accountability controls, whichpart of the Action Control portfolio (see Section

alue Systems has emerged as the prime motivationalf disruption in safety social capital (49/123 SFs). Cor-ngly one would have expected to find a high failureultural controls, which include adherence to codesct, and similar controls that seek to guide/monitorstems. Actually, as per the results of this analysis,inor proportion of the control failures is due to cul-trols (15/123 SFs). Further research is suggested to

and resolve this paradox.

sions

re many different approaches that can be followeding towards a better safety culture-this paper dis-

ocio-management approach for achieving a betterure. Behavioral insights from past accidents pro-le lessons for safety culture. In this paper, a socio-

nt accident analysis model called CAMSoC (Curtail-nts by Managing Social Capital) is proposed. Thisalyses accident causing deviant behavior by tracingin safety social capital, the networking relationshipsActor Group responsible for this erosion, and finally

ation of motivation sources and failed organizationalt led to the accident. The CAMSoC model of accidents demonstrated by applying the model on five acci-

pal, Westray, Tenerife, Hyatt and Exxon Valdez. Thetained from the analysis in terms of sources of moti-failure of organizational controls was discussed andd. Organizations and Regulators can use these socio-nt insights to better manage organizational networks,values in the context of safety, i.e. the organizationalal capitaland thereby move towards bettering there.

gements

he reviewers for their valuable suggestions and guid-previous draft of this paper. This paper is devoted

ite affection bestowed on me by my dear Father ander.

s

A.M. Collins, Safety culture: a review of the literature, definitiony culture as suggested by ACSNI, HSE, 1993. As accessed atw.hse.gov.uk/research/hsl pdf/2002/hsl02-25.pdf.in, Ideas and Opinions by Albert Einstein, Crown-Rupa Publish-, p. 153.ana, Behavioralism and political theory-Contributions of Davidnd Lucian Pye, Rawat Publications, 2001, pp. 5657.aeffer, R.P. Lamm, Sociology, Tata Mcgraw-Hill Edition, 1999, p.

Veacc

Of[7] J. N

forExPo

[8] S.ofhtt

[9] R.DAs

[10] R.Dpro

[11] R.Dthrcom

[12] A.An

[13] T.devAm45

[14] A.icadreNe

[15] K.APea

[16] N.Safhtt

[17] Iniws

[18] R.APsyp. 5

[19] S.(Edof

[20] A.Gma

M.Els

[21] Caat C

[22] J.ww

[23] Claporold

[24] TraSco

[25] Wdenwr

[26] H.Air22sec

[27] Ayfro

[28] CoMcdefpublished in Safety Science 2000, vol. 36, pp. 111136. Asat http://www.behavioral safety.com/articles/Towards A Model

ty Culture/.Have Australias Major Hazard Facilities Learnt From The Long-aster? An Evaluation Of The Impact Of The 1998 Esso Longfordn On Major Hazard Facilities In 2001. Sponsored By The Publicnit, The Institution Of Engineers, Australia, 2001. p. 49.

(Ed.), translated by W.D. Halls, Emile Durkheim, The Ruleslogical Method, Free Press, 1982, pp. 5059. As accessed atw.generation-online.org/p/fpdurkheim1.htm.

nam, Social Capital: what is it? What does Social Capital mean?shed at http://www.bowlingalone.com/socialcapital.php3.nam, Bowling alone, J. Democracy 6 (1) (1995) 6578, Accessby Stony Brook University.tnam, Social Capital: What is it? Social Capital worksmultiple channels: as published at http://www.bowlingalone.

ialcapital.php3.s, Social capital: its origins and applications in modern sociology,c. Rev. Sociol. 24 (124) (1998) 3.ihara, Role of social capital and external actors in grassrootsent, in: Prepared for Delivery at the 2001 Meeting of the Latin

n Studies Association, Washington, DC, September 6, 2001, pp.

s, P. Landolt, The downside of social capital. The Amer-ospect Online, 7 as quoted from smart library on chil-

families, 1996, courtesy: http://www.children.smartlibrary.org/rface/segment.cfm?segment=2526.rchant, Modern Management Control Systems Text and Cases,Education, First Indian Reprint, 2002, p. 9.son, A New Accident Model for Engineering Safer Systems,cience, vol. 42, no. 4, 2004, pp. 237270. As obtained fromnyday.mit.edu/accidents/safetyscience-single.pdf.

ction to SociologyProfessor Larry Stern as accessed at http://cd.edu/lstern/sociology 1301 class notes 01.htm.ron, D. Byrne, Social Perception: Understanding Others Socialgy, 8th ed., Prentice Hall of India Private Limited, 1997,

Safety investigationsa human perspective, in: C. Kirchsteigeroceedings of the JRC/ESReDA Seminar on Safety Investigationents, 2003, pp. 308317.shton, Lessons learned from past accidents, risk assessment andent in the context of the Seveso II directive, in: C. Kirchsteiger,ristou, G.A. Papadakis (Eds.), Industrial Safety Series, vol. 6,1998, pp. 8386, 9196.

. 233,Case Name: Bhopal Disaster, TED Case Studies, as accesseddownload\ec\bhopal\Bhopal Disaster.htm.ing, Union Carbide: Disaster at Bhopal, as accessed at http://

opal.com/pdfs/browning.pdf.tion suit Ref: 99 Civ. 11329 (JFK) against Union Carbide Cor-and Warren Anderson as accessed at http://www.bhopal.net/

mendedplaint.html.pt of Westray Mine Public Inquiry Commission Stellarton, Nova

Coal Mine Disaster Pictou County, Nova Scotia Acci-rmation as accessed at http://www.littletechshoppe.com/ns1625/u.html.

er, Report by the Secretary of Civil Aviation, sourced fromAccident Digest (ICAO Circular 153-AN/56), 1978, pp.

s accessed from http://www.panamair.org/OLDSITE/Accidents/fe.htm.eports Pan am 1736 KLM 4805, January 12, 2003 as accessed://www.airsafetyonline.com/.overview, EMCH 200 Fall 2001, Section 001, Instructor:as accessed from http://www.me.sc.edu/fs/mcneill/EMCH200/tml.

740 S. Rao / Journal of Hazardous Materials 142 (2007) 730740

[29] Bidding is not the Solution: Case Studies in Bidding as accessedfrom the website of American Council of Engineering Companies http://www.acec.org/.

[30] Kansas Citys Hyatt Tragedy: 20 years Later: Skywalk collapsestill offers lessons to engineers, inspectors, Rick Montgomery TheKansas City Star Date: 07/15/01 as accessed from http://www.kcstar.com/projects/hyatt/.

[31] Case Discussion: Engineering Ethics: The Kansas City Hyatt RegencyWalkway Collapse, Texas A&M University, as accessed at http://ethics.tamu.edu/ethics/hyatt/hyatt1.htm.

[32] Excerpt Reprinted Final Report of Alaska Oil Spill Commission,the State of Alaska, 1990 byas accessed from http://www.evostc.state.ak.us/History/excerpt.htm.

[33] D. Postman, S. Toomey, NTSB: Captain Too Drunk To Run Ship,Article citing the investigations of National Transportation SafetyBoard, Anchorage Daily News accessed at http://www.adn.com/evos/stories/EV84.html.

Safety culture and accident analysis-A socio-management approach based on organizational safety social capitalIntroductionHuman Factors analysis modelsNeed for extending the scope of accident analysis into sociological domainsSociology and organizational safety

A socio-management framework for safety cultureOrganizational safety social capital-a socio-management theoretical foundationSafety Social capital and safety culture

An actionable socio-management framework for managing safety social capitalMotivational sources of safety social capitalThe importance of managing negative social capitalManaging Social capital through behaviorally oriented controls

The CAMSoC (curtailing accidents by managing social capital) modelLending a structure to the CAMSoC model

Socio-management analysis of five major accidents using CAMSoCA brief description of the accidents [19,20]The Bhopal Gas Tragedy [21-23]The Westray coal mining disaster [24,25]The Tenerife Air Tragedy [26,27]The Hyatt Regency Walkway collapse [28-31]The Exxon Valdez oil-spill disaster [32,33]

Applying the CAMSoC model for accident analysisMethodology adopted for accident analysis using CAMSoC modelApplication of CAMSoC model in the Hyatt Regency Walkway collapse-an illustration

Results of the accident analysis and implications for organizational safetyInsights on motivation sourcesTrends in failure of behavior oriented controls

ConclusionsAcknowledgementsReferences