14
This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Occupational accident experience: Association with workers’ accident explanation and definition

Embed Size (px)

Citation preview

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Occupational accident experience: Association withworkers’ accident explanation and definition

Claudia Niza a,*, Sılvia Silva b, Maria Luısa Lima b

a London School of Economics (LSE), Operational Research Department, Portugal Street, London, UKb Centro de Investigac!ao e Intervenc!ao Social (CIS)/Instituto Superior de Ciencias do Trabalho e da Empresa (ISCTE),

Departamento de Psicologia Social e Organizacional, Av. Forc!as Armadas, Ed. ISCTE, 1649-026 Lisboa, Portugal

Abstract

Problem: The experience of an occupational accident, as a harmful and unexpected event, may elicit a process of con-struction of meaning. However, research has provided inconsistent results regarding to the role of such an experience inposterior sense-making. This article aims at understanding the association of an occupational accident experience with theexplanation and definition given by workers for such events.

Method: Fifty-six semi-structured interviews were conducted with workers from several sectors. This data was subjectto content analysis and HOMALS (Multiple Correspondence Analysis).

Results: An accident experience was associated with defensive explanations (focused on causes external to workers) anddefensive definitions (highlighting the sudden nature of accidents and organizational weaknesses).

Discussion: The study of the construction of meaning about accidents and its variability is highly relevant for under-standing posterior preventive behavior and the accuracy in the report of future occupational accidents.

Impact on industry: Organizations should be aware of discrepancies in conceptions and interpretations among employ-ees about occupational accidents when planning prevention programs or improvement of accident data registration.! 2007 Published by Elsevier Ltd.

Keywords: Occupational accident; Accident experience; Lay explanation; Definition; Prevention

1. Problem

Occupational accidents represent a serious problem to society, with statistical data alerting to the high fre-quency and severity of these events every year (e.g., ILO, 2002). Research on occupational accidents hasexposed the negative impact these hazards have on their victims, families and co-workers (e.g., Dembe,2004), with both consequences for cognition and behavior (see Gonc!alves et al., 2008).

The role of a personal accident experience in the way individuals think about these hazardous events is notwell-established in literature. On one hand, some studies have shown that an accident experience has an

0925-7535/$ - see front matter ! 2007 Published by Elsevier Ltd.doi:10.1016/j.ssci.2007.11.015

* Corresponding author. Tel.: +351 217903079; fax: +351 217903962.E-mail addresses: [email protected] (C. Niza), [email protected] (S. Silva), [email protected] (M.L. Lima).

Available online at www.sciencedirect.com

Safety Science 46 (2008) 959–971

www.elsevier.com/locate/ssci

Author's personal copy

important impact on individuals’ subsequent sense-making. Shaver (1970), based on the work of Walster(1966), developed the defensive attribution hypothesis which claimed that the victims of accidents tended toexplain the event in a way that personal responsibility is minimized. Further supporting results (e.g., Salminen,1992) showed that, when explaining an accident, workers tended to focus on the circumstances of the accident,whereas supervisors tended to blame the employees as the source of the deviation from the normal work rou-tine. This attribution di!erence is the most replicated result regarding the impact of a personal accident expe-rience on sense-making (e.g., Kouabenan, 2000c, 2001b). Gyekye and Salminen (2006) demonstrated that thistendency is also shown by coworkers who empathize with the victim of the accident. The authors identifiedthree categories of coworkers in their relation to the accident (situationally relevant, personally relevantand non-relevant) and concluded that witnesses who had some perception of personal or situational similaritywith the victim attribute less responsibility to the hazarded worker.

However, other empirical findings failed to find an association between accident occurrence and reasoning.Gyekye (2003) in a study about the attribution of causes for occupational accidents in both dangerous (min-ers) and non-dangerous (textile employees) work environments, concluded that victims perceived workers’personal characteristics to be primarily responsible for accident occurrences. On the other hand, Kouabenan(1998) in a study about the beliefs and perceptions of risk related to accidents concluded that ‘‘the true e!ect ofaccident experience (. . .) remains to be established” (p. 251) as he found no connection between a prior acci-dent history and specific beliefs related to accidents. Moreover, Girasek (1999) explored the meaning of theword accident with the purpose of understanding the lay notions of preventability and predictability relatedto accidents [e.g., when you hear the work accident, do you usually think that what happened could have beenpredicted (prevented)?]. Within the socio-demographic information collected from participants (e.g., sex, age),the author included data from prior accident experience but concluded that previous victims did not di!er sig-nificantly from non-victims in the interpretations given to the word accident.

Understanding the construction of meaning individuals make of their experience can be related to thestudy of lay epistemology or the process of knowledge acquisition and organization of laypeople (E"er,1984). In the case of hazardous events, the attribution of causes is the usual framework (Silvester and Chap-man, 1997). However, causal attribution is included in a larger scheme of sense-making of the world and thesearch of causes to events is just one of the forms through which individuals understand their social envi-ronment (Pernanen, 1993). As exposed by Bar-Tal and Kruglanski (1988), the theory of lay epistemologyhas two distinct but interrelated categories: on one hand, identification and classification of concepts, andon the other hand, the connection between concepts, where the search of causality is included. Followingthis perspective, the definition of concepts or searching for ‘‘what is it” and not just ‘‘why is it” is anotherimportant way of constructing meaning (e.g., Condit et al., 2004). Moreover, as these two categories arelogically related (Bar-Tal and Kruglanski, 1988), it is plausible to think that these two processes may be con-nected in a way that the causes individuals give to explain an action or event may be coupled with the def-inition they provide for it.

We intend to clarify this relationship, focusing on the analysis of the impact of a worker’s accident expe-rience on both the causal attributions and definitions of occupational accidents.

1.1. Causal attribution for occupational accidents

Attribution activity is the process of sense-making whereby the individuals attempt to identify the causesthat are present in their daily lives (Deschamps and Clemence, 2000) either about feelings, behaviors or events(e.g., Kelley and Michela, 1980).

The foremost findings in attribution literature are the actor/observer distinction (Jones and Nisbett, 1971)and the fundamental attribution error (Ross, 1977). Jones and Nisbett (1971) exposed the attribution di!er-ences between actors and observers to an event, and presented the explanation that distinct personal involve-ment and access to information tended to make actors more focused on external causes (associated with thecontext of the situation) and observers on internal causes (connected to the individuals). Regarding the otherimportant finding, Ross (1977) defined what was called the fundamental attribution error, characterised by thefact that people generally tend to attribute the causes of events to the individuals involved in them, neglectingthe importance of situational factors.

960 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

This causal attribution framework has been applied to the study of hazardous events in the workplace,namely burnout (Moore, 2000), alcohol-related accidents (Pernanen, 1993) or occupational accident (e.g.Melia et al., 2001) because these kind of rare or overwhelming events at work make employees stop andask why (Silvester and Chapman, 1997). Particularly regarding occupational accidents, the causal search refersto the process of determining why an accident has occurred. The outcome of this search is the causal attribu-tion: the individual’s explanation of what caused the accident.

Despite the consistency in the aforementioned attribution di!erences between victims and observers of acci-dents, Kouabenan (1999, 2001a) mentioned the existence of other factors that could influence the lay expla-nations of accidents. On one hand, age was shown to have an influence in employees’ explanations tooccupational accidents (e.g., Melia et al., 2001), being older workers more likely to attribute the event to exter-nal factors and younger workers to refer internal attributions. On the other hand, Gherardi et al. (1998)referred that a higher educational level was associated with internal attributions, by contrast with externalattributions given mainly by low literate individuals. Kouabenan et al. (2001) showed that the sex and the hier-archical level of individuals played a significant role in the attributions for an accident. These authors con-cluded that subjects with a high hierarchical position made more attributions blaming the victim than didsubjects in a subordinate level. Moreover, sex interacted with the hierarchical position in a way that malesupervisors made more internal attributions but there were no di!erences among women of di!erent levels.At an organizational level of analysis, Hofman and Stetzer (1998) referred safety climate as a moderator ofthe common defensive attribution pattern found in victims. Safety climate as the perceptions of events, prac-tices and behaviors related to safety in the organization (Silva et al., 2004) was shown to attenuate the ten-dency for defensive attributions, resulting that individuals working in a positive safety climate tended tomake more internal attributions than employees in a negative safety climate organization. Regarding organi-zational culture, research has shown that di!erent communities of practices and shared knowledge tend to givedi!erent explanations to occupational accidents. In a study on the construction sector, Gherardi et al. (1998)concluded that engineers explained accidents as abnormal events resulting from the disrespect of safety rules,while foremen justified the occurrence of accidents as being an expected result of the dangerousness of thework environment. Lastly, Kouabenan (2001) focused on the impact of national cultures on lay explanations,concluding that western individualistic societies present more internal attributions to accidents, while morecollectivist societies tend to o!er more explanations related to external factors.

The existence of several variables that have been shown to play a role in the attribution process suggests thatthe attribution of specific causes to accidents may vary according to personal characteristics, organizationalcontext or cultural variables. It is a dynamic and purposive process that provides individuals with an internalsense of security and order. This subject is of extreme importance because, regardless of the actual causes ofparticular accidents, it is the explanation given by individuals that determine their emotional, cognitive andbehavioral responses to the accident (Woodcook, 1995). The application of attribution theory to occupationalaccidents is highly relevant because identifying the causes of accidents is a way of structuring the situation andrelating it to one’s motivations and needs (Pernanen, 1993) and a fundamental prerequisite for preventiveaction (e.g., Kouabenan, 1998). There are several studies that acknowledge the importance of understandingthe attribution of causes to events (e.g., Butchart et al., 2000; Kouabenan, 2000a) for prevention. Individualsadjust their future behavior to the perceived causes of accidents and the belief of preventability is central toprevention actions (Poole, 1987). Moreover, if naıve causal explanations for accidents are communicated, acci-dents countermeasures could be better understood and more easily accepted because they would integrate thecognitive and emotional functioning of those directly involved in their execution (Kouabenan, 2000b).

1.2. Definition of occupational accident

From the perspective of international organisms, ILO (2002), in its worldwide report on Recording andNotification of Occupational Accidents and Diseases, concluded that there was a great variety of occupationalaccidents definitions among countries, ranging from a simple reference in legislation to accidents occurring inthe workplace (e.g. Botswana, UK) to countries with a more precise definition (e.g. USA). Similarly, Eurostatin its Questionnaire on National Declaration Systems of Accidents at Work, identified variability in accidentdefinition as a factor that could compromise the comparability of data on national occupational accidents.

C. Niza et al. / Safety Science 46 (2008) 959–971 961

Author's personal copy

From an academic perspective, there are very few studies that have analysed the importance of occupa-tional accidents definition. In a study on hospital service workers, Weddle (1996) found that of all workersrecalling having been injured in the previous year, about 40% did not report one or more injuries. For mostof these cases, subjects said that they did not consider that particular harmful event to be an occupational acci-dent. On the other hand, Conroy and Sciortino (1997) evaluated whether di!erent definitions used for occu-pational deaths would identify di!erent cases and proposed a definition for occupational injury. According tothem, injury is consensually defined by researchers as ‘‘tissue or body damage or loss of function of a bodypart” (p. 274) but the problem is the occupational prefix as it ‘‘is more di#cult to define and there is no stan-dard definition used consistently by researchers. Definitions vary widely and depend upon the data source andreason for identifying cases of occupational injuries” (p. 274). These authors provided evidence suggesting thatcounts of occupational injury fatalities based on surveillance systems using restricted definitions di!er fromsystems using broader definitions for occupational injury death. In a study by Saldana et al. (2003) there wereexamined the di!erences and similarities between several possible ways to define occupational accidents. Theauthors mentioned that there exist two di!erent views on what constitutes an occupational accident: the firstintegrates all events, with or without injuries, whereas in the second, occupational accidents are only those thatresult in injury consequences (naming the events that do not result in injuries as incidents). These authorsdefended the use of a wide definition that should cover all events with potential to cause damages (personalor organisational) as a more valid analysis framework than a narrow definition of occupational accidents. Assuggested by Pernanen (1993), the lack of interest or awareness about the importance of the definition of theconcept accident (reflected in the scarce studies on this subject) may be due to the fact that most researchersthink of occupational accident as a well-defined expression, with an equally shared meaning and find no needfor further clarification. However, a study on the definitions presented in occupational health and safety lit-erature (Niza et al., 2006) showed that there are a number of di!erent definitions in use, depending on theauthors’ scientific domain and nationality. Moreover, there are noteworthy examples from studies aboutthe public sense-making (e.g., Condit et al., 2004; Chapman et al., 2003) that acknowledge the importanceof studying the meaning of concepts to individuals because there is evidence of a significant variety of inter-pretations for the same words.

These discrepancies in the definition and interpretation of what is an occupational accident may obstruct ane!ective communication between workers, employers and governmental organizations. The definition used byinstitutional entities may not be shared by organizations, and organizations may have an o#cial definitionthat is very di!erent from what its employees think an occupational accident is. There is a potential spacefor concept misunderstanding that may create gaps between the number of accidents that occur and the num-ber present in o#cial statistics.

1.3. Aim of the study

Insofar, research has shown that di!erent people may perceive and interpret the same accident in di!erentways and that an accident experience has some impact on posterior sense-making. Although there is a consis-tent amount of evidence about the type of explanations victims of accidents give, it is not clear the impact thisexperience has on the definition of the event and how these two sense-making processes may be connected. Asshown in literature (e.g., Bar-Tal & Kruglanski, 1988) the combined study of these subjects is fundamental tounderstand the overall lay reasoning about occupational accidents.

Given these considerations, in the present study we aim at understanding the association of a personaloccupational accident experience with the causal attributions and definitions given by workers.

2. Method

2.1. Sample

The sample was composed by 56 subjects (29 women and 27 men), with ages ranging from 20 to 64 yearsold (average 36.98). The participants worked in the services (63.6% of which 17.9% in health and 8.9% in edu-cation) or industry (35.7% of which 7.1% in construction) sectors. Seniority in the organizations had a min-

962 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

imum of 6 months and a maximum of 32 years, with an average of about 8 years. Regarding literacy, about15% of the subjects were at the sixth grade level and about 18% had completed the ninth grade. More than halfof the subjects (57.1%) completed the 12th grade and a quarter of the sample was at the university level.

Almost all subjects were employed at the time of the interview (85.7%) and about half had some experiencein the safety area (e.g., training in workplace safety procedures). In the past 3 years, a third of all participantshad an occupational accident (33.9%) and about 20% of all subjects had experienced sick leaves as a conse-quence of occupational accidents. Almost half of the subjects confirmed a personal occupational accidentexperience at some point in their lives and about two thirds reported having had incidents (e.g., near missesor small harmless accidents) in the company they were working at the time of the data collection.

2.2. Data collection

This study was based on a convenience sample of interviews performed by Psychology students. The super-visors gave training to the students in the interview protocol, followed all data collection during individualtutorial sessions (three for each interviewer) and at the end the quality of their work were assessed. Each stu-dent made a previous contact with the participants in order to book the future appointment. The majority ofthe interviews were carried out outside the participants’ organization (80%) and the remaining in a workplacecontext.

The interview schedule was developed by the authors of this paper. The interviews performed were semi-structured, including questions related to the participants’ employment situation, lay conceptions about occu-pational accidents and participants’ personal experience of hazardous events. The questions analyzed for thisstudy were questions: ‘‘In your opinion, what is an occupational accident?” and” In your opinion, what are thecauses for occupational accidents?” The first question was related to the definition of occupational accidentsand the second was about the causes of occupational accidents.

All interviews were recorded and subjected to fully transcription.

2.3. Data analysis

A content analysis (with development of a system of categories) and a Homals (multiple correspondenceanalysis) were the procedures used to analyze the data.

The content analysis used to study the causal attributions for occupational accidents was adapted fromGyekye (2003, p. 537). This author created a questionnaire divided into external and internal causal factorsthat influenced or contributed to occupational accidents, both presenting 15 subcategories to which the sub-jects answered in a five point scale. In our study, these subcategories were not rated in a scale but marked asabsent/present in the content analysis.

Although using the same strucure, there were only included the subcategories from Gyekye (2003) that werementioned by the subjects. Ten external and eight internal subcategories were removed from the analysis.Regarding external attributions, the subcategories unsafe equipment, inadequate training, pressure from man-agement and poor housekeeping were kept. It was necessary to add three external subcategories (mentioned byour subjects and not present at the original study): lack of inspection, lack of supervision and bad luck/des-tiny. Concerning the internal related subcategories, there were included lack of skill, attention lapse, miscon-duct, inexperience, carelessness, bad day and fatigue. In this case, it was not necessary to include further sub-categories related to internal factors.

This study introduced a new variable in the analysis, designated by primary attribution and defined as thefirst type of attribution mentioned by the subjects. Subjects commonly refer several factors to explain occu-pational accidents, mentioning both internal and external causes. This variable was related to the first causalfactor referred by the subjects, conceptualised as their more immediate explanation.

For a summary of the system developed and category dictionary see Table 1.The content analysis performed on the question about the definition of occupational accidents was devel-

oped in Niza et al. (2006), in a study about the definitions and operationalizations of work-related accidentsused in empirical studies for a 10-year period (1995–2005). This system of categories is structured aroundfour main categories: (1) occupational accident context, (2) occupational accident sudden characteristic,

C. Niza et al. / Safety Science 46 (2008) 959–971 963

Author's personal copy

(3) occupational accident causes and (4) occupational accident consequences. The first category was related towhat was happening at the time or where the accident occurred, and was divided in to three subcategories: (1a)at the workplace, (1b) during the performance of a task, and (1c) commuting to or from work. The secondcategory was connected to the sudden or unexpected characteristic of the accident event. The third categorywas related to the causes of the occupational accident and was divided into three subcategories: (3a) lack ofsafe equipment, (3b) lack of working conditions, and (3c) lack of safety training. The last category was aboutthe consequences of the accident and was splat into (3a) physical consequences and (3b) psychologicalconsequences.

This data analysis procedure was validated through an inter-judge consensus, in which two independentjudges classified a random sample of answers according to the developed categories. There was an overallinter-judge consensus of 85%, suggesting the suitability of the system produced.

In the last step, a Homals analysis was performed including the subcategories developed in the contentanalysis related to causal attributions and definitions, in addition to socio-demographic and accident experi-ence variables. Homals is a Multiple Correspondence method for qualitative variables that finds patterns ofassociation (proximity) in the data. This procedure organizes the information along two orthogonal dimen-sions, and the intersection of these two dimensions provides a combined pattern in four quadrants of aggre-gated data.

3. Results

3.1. Content analysis

The frequencies for each subcategory related to causal attribution are shown in Table 1.The majority of the primary attributions made by the subjects were internal (33), and overall, the explana-

tions mentioned were more internal (70) than external factors (54). This first result supports the fundamentalattribution error previously mentioned. Within the internal factors, the causes more frequently referred weremisconduct (51.8%), carelessness (30.4%) and attention lapse (26.8%), whereas the more frequent external fac-tors pointed were poor housekeeping (30.4%) and unsafe equipment (28.6%). Taking all factors together, theresults showed that there were five main causes referred by the subjects, of which three were internal andtwo external: misconduct (I), carelessness (I) and poor house keeping (E), unsafe equipment (E) and attentionlapse (I).

Table 1Dictionary and frequencies related to causal attribution categories (adapted from Gyekye, 2003)

Categories Subcategories Dictionary Absolute frequency Relative frequency

Primary attribution Internal Causes associated with the worker 33 58.9External Causes associated with the context 18 32.1

Internal attribution Lack skill Deficient knowledge about task 2 3.6Attention lapse Incorrect course of action selected 15 26.8Misconduct Failure to use protective equipment 29 51.8Inexperience Lack of adequate ability 4 7.1Carelessness Exceeded prescribed limits 17 30.4Bad day Unusual misbehavior 2 3.6Fatigue Reduced alertness 1 1.8Total 70

External attribution Lack inspection No external safety assessment 2 3.6Lack supervision No management control 6 10.7Bad luck/destiny Inevitability of accidents 6 10.7Unsafe equipment Faulty utensils and tools 16 28.6Inadequate training Deficient preparation and guidance 6 10.7Pressure Excessive work pace 1 1.8Poor housekeeping Lack of workplace tidiness 17 30.4Total 54

964 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

The frequencies of each subcategory connected to definition are shown in Table 2.The most frequently mentioned categories were context (29) and consequences (23), followed by causes (16)

and the sudden characteristic (6) of an accident. In the context category, the subcategories with higher fre-quencies were task performance (21) and workplace (16), while in the consequences category the resultsshowed a focus on the physical aspect of the accidents (20). Regarding the category of causes, lack of safeequipment (12) and working conditions (9) were the subcategories more frequently mentioned.

3.2. Homals

As previously mentioned, Homals provides an organization of the data along a two orthogonal dimensionsstructure (Table 3).

Dimension 1 was associated with the definitions given, personal history of accidents and activity sector(Eigenvalue 0.142). On the other hand, Dimension 2 was connected to the recent accident experience andsocio-demographic characteristics (Eigenvalue 0.111). Explanations for occupational accident were dividedbetween the two dimensions. Table 4 provides a description of the coordinates that allows the placement ofeach variable in the analysis in this two dimensional space.

The intersection of these two dimensions presents a four quadrants structure of the data. Fig. 1 presents anorganized synthesis of the results found. The position of the variables correspondent to the coordinates pre-sented in Table 4 is presented in structured representation.

The definitions on the left-side quadrants in Fig. 1 were designated defensive because were focused almostexclusively on external causal factors, minimizing the personal responsibility of the employee. On the otherhand, attributions named internal were situated in the right-side quadrants, with primary internal attributionand mentioning only one external factor, emphasising the personal role of the employee in the occupationalaccident. On a di!erent perspective, the upper quadrants were characterised by simple attribution styles,because they only focused on a maximum of two causal factors. Finally, the lower quadrants reflected morecomplex attribution approaches, including both locus of control (internal and external) and severalexplanations.

Regarding the nature of the definitions, they were classified as defensive (referring the causes and suddencharacteristic of the accident) definitions or neutral (based on the accident’s consequences and context). Wheninterpreting the definitions regarding their focus point, they could be centred in the fact that an occupationalaccident was most of all an accident (definitions based on consequences and characteristic) or in the perceptionthat an occupational accident is specially something that happens at work and reporting it to the workplace(definitions related to the context and causes).

Results showed that an occupational accident experience was associated with defensive explanations basedon external factors, supporting Shaver’s (1970) defensive hypothesis. The results also showed that an occupa-tional accident experience was associated with defensive definitions and defensive attributions (left-side

Table 2Frequencies related to definition categories

Categories Subcategories Absolute frequency Relative frequency

Context 29 51.8Workplace 16 28.6Task 21 37.5Commuting 4 7.1

Sudden characteristic 6 10.7

Causes 16 28.6Lack of safe equipment 12 21.4Lack of work conditions 9 16.1Lack of safety training 2 3.6

Consequences 23 41.1Physical 20 35.7Psychological 5 8.9

C. Niza et al. / Safety Science 46 (2008) 959–971 965

Author's personal copy

quadrants). This result confirmed an association between explanations and definitions forms of sense-making.On one hand, workers who believe accidents are due to bad luck consider an occupational accident somethingthat is sudden and unpredictable. On the other hand, workers who explain these accidents as a result of orga-nizational weaknesses (e.g., unsafe equipment) defined an occupational accident in the same way: by its causes.Furthermore, the results also revealed a connection of definitions focusing on the accident with a simple attri-butional style, and definitions focusing on work with a more complex attributional style.

Fig. 1 also showed an association of each profile of attribution and definition (providing real examplestaken out of interviews) with the correspondent socio-demographic groups. The results propose that di!erentwork experiences and practices may be connected with diverse ways to define and explain occupational acci-dents. Workers at the beginning or end of their professional careers seem to understand accident in a simplerway than individuals at the peak of their work lives. The di!erent work environments of services, constructionor industry also appear to influence the perception of occupational accidents. This result strengthens the per-spective of occupational accidents sense-making being an active and purposive process that relates to individ-uals according to their needs and motivations.

Table 3Discrimination measures in both dimensions

Dimension

1 2

Context 0.353 0.012Context_workplace 0.185 0.015Context_task 0.291 0.044Context_commuting 0.081 0.011Sudden characteristic 0.017 0.012Causes 0.452 0.094Causes_lacksafety 0.437 0.091Causes_lackconditions 0.443 0.112Causes_lacktraining 0.070 0.029Consequences 0.467 0.179Consequences_physical 0.408 0.176Consequences_psychological 0.125 0.023Primary attribution 0.135 0.147E_lackinspection 0.042 0.014E_lack supervision 0.044 0.028E_badluck 0.013 0.040E_unsafeequipment 0.090 0.029E_inadequatetraining 0.006 0.027E_pressure 0.060 0.007E_poorhousekeeping 0.011 0.048I_lackskill 0.047 0.021I_attentionlapse 0.016 0.012I_misconduct 0.044 0.061I_inexperience 0.127 0.006I_badday 0.009 0.134I_carelessness 0.100 0.133I_fatigue 0.002 0.013

Accident ever 0.217 0.194Accident last 3 years 0.126 0.218Activity sector 0.289 0.253Sex 0.051 0.083Age 0.094 0.508Literacy 0.082 0.282Seniority 0.184 0.471Employment situation 0.037 0.184

Eingenvalue 0.142 0.111

966 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

4. Discussion

This study aimed at understanding the association of a personal occupational accident experience with thecausal attributions and definitions given by workers, using interview data from a group of 56 participants.

The results suggested that an occupational accident experience influences workers’ sense-making about thissubject and this could be found in both causal attributions and definitions.

With respect to causal attributions, an occupational accident experience was associated with defensiveattributions (focused on factors external to workers), supporting the defensive attribution hypothesis. The

Table 4Coordinates of variables in Dimension 1

Dimension

1 2

Context 0.633 !0.354Context_workplace 0.652 !0.957Context_task 0.776 !0.279Context_commuting 1.038 !2.341Sudden characteristic !0.195 0.519Causes !1.196 !0.184Causes_lacksafety !1.436 !0.036Causes_lackconditions !1.577 !0.004Causes_lacktraining !1.740 !0.054Consequences 0.867 0.509Consequences_physical 0.906 0.625Consequences_psychological 1.165 1.808Primary internal attribution 0.177 !0.113Primary external attribution !0.345 0.275E_lackinspection !0.684 !2.266E_lack supervision 0.368 !0.358E_badluck !0.460 0.202E_unsafeequipment !0.398 !0.319E_inadequatetraining !0.215 !1.118E_pressure !1.883 !0.042E_poorhousekeeping !0.248 !0.382I_lackskill 1.130 2.030I_attentionlapse 0.234 !0.218I_misconduct 0.119 !0.007I_inexperience 1.008 0.114I_badday !0.904 2.165I_carelessness 0.580 !0.237I_fatigue 0.451 !0.558

Accident over 3 years !0.088 0.429Accident last 3 years !0.305 !0.146Education 0.120 !0.048Health 0.367 !0.171Other Services 0.223 !0.027Men 0.131 0.111Women 0.119 !0.04820–25 years old 0.203 !0.29125–40 years old 0.092 !0.12140–45 years old !0.078 0.35245–65 years old !0.032 0.022Seniority less than 2 years 0.217 !0.416Seniority 2–5 years 0.037 !0.337Seniority 5–10 years !0.123 !0.168Seniority over 10 years !0.174 0.100Unemployed 0.407 1.028Retired !0.813 0.127

C. Niza et al. / Safety Science 46 (2008) 959–971 967

Author's personal copy

consistency of this hypothesis suggests that understanding the possible ways this pattern of attribution maytake is also very important. Although all victims seem to share the same style of explanation, there were dif-ferences in the attributions’ levels of complexity. This result proposes that victims undertake distinct defensivestrategies or process the information di!erently according to their specific characteristics (age, seniority) anddistance from the event (recently or more than 3 years ago). Moreover, the results reflected a predominance ofinternal factors, supporting other results about the fundamental attribution error and the internality norm(e.g., Beauvois and Dubois, 1988). Overall, subjects mentioned more internal factors when explaining occupa-tional accidents, regardless of their past accident experience (victims and non-victims).

This defensive pattern was also found in the definitions of occupational accidents. The defensive definitionsare characterized by reference to the sudden nature of the accident or organizational weaknesses that caused it.They were also named defensive (as attributions) because they reflect a particular focus of workers on the fea-tures of accident that go beyond their personal control. This description of occupational accidents is associ-ated with defensive attributions, shedding light on the combined sense-making process of defining andexplaining events. The recognition of the association between these two reasoning forms may prevent negative

Systematic representation of the different attribution and definition styles and their associated

socio-demographic groups

SIMPLE EXPLANATIONS ACCIDENT RELATED DEFINITIONS

DEFENSIVE EXPLANATIONS DEFENSIVE DEFINITIONS

COMPLEX EXPLANATIONS WORK RELATED DEFINITIONS

INTERNAL EXPLANATIONS NEUTRAL DEFINITIONS

Definition: Consequences “An occupational accident can be from a small scratch to death”

Definition: Characteristics “An accident is something unpredictable”

Attribution: lack of skill (I), inexperience (I) “Accidents happen because people don’t know their job”

Attribution: Bad luck (E), Bad day (I) “An accident just happens”

Attribution: Misconduct (I), Carelessness (I), Attention Lapse (I), Fatigue (I), Lack supervision (E) “Accidents happen because people don’t pay attention and are irresponsible. But there can be other reasons”

Definition: Context “An occupational accident is an accident that happens in the workplace”

Attribution: Unsafe equipment (E), Management Pressure (E), Inadequate Training (E), Poor Housekeeping (E), Lack inspection (E) “There are several causes to accidents but basically accidents happen because there is no safe equipment and facilities”

Definition: Causes “An accident is something that happens because of lack of working conditions”

No accident Unemployed9th/ 12th grade

No accident Recently employed

20-40 years old Women

Services Sector

Accident less than 3 years ago Construction Sector

6th grade Seniority 5-10 years

Accident over 3 years ago Retired

40-65 years old Men

Industry Sector 4th grade

Seniority >10 years

Fig. 1. Systematic representation of the di!erent attribution and definition styles and their associated socio-demographic groups.

968 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

information cycles in the organizations: employees who believe an accident is something unpreventable anddue to bad luck will be less likely to take protective measures and to accurately report the event. If the needfor prevention is not raised, workers may keep thinking that accidents are caused by fate or organizationalflaws and will not report accidents caused by their own misconduct.

This study presents one major limitation. Despite the structured questions format, the interviews were per-formed by several Psychology students, more inexperienced that senior researchers in the collection of thedata. However, possible biases arising from this data collection procedure were minimized through the com-plete record and transcription of the interviews.

Nevertheless, our results have important methodological, theoretical and practical implications (Ryneset al., 2005).

At the methodological level, our study provided a qualitative analysis of the causal attribution combinedwith the definitions of occupational accidents that was inexistent in the literature. There are several examplesof studies about causal attributions of accidents, but most are based on quantitative scales that are presentedto the participants in a predetermined format. Our study o!ered the possibility to workers to express theirsubjective views without the limitations of a fixed and prearranged response instrument and the Homalsmethod permitted the analysis of the richness of the answers.

At the theoretical level, this study highlighted the variability of the sense-making di!erences amongemployees and suggested that the study of the lay interpretation of occupational accidents should pay atten-tion to these particularities. Furthermore, it included the analysis of the definitions of occupational accidentsand connected them with causal attributions in a larger lay epistemology framework (Bar-Tal and Kruglan-ski, 1988). To our knowledge, this the first study to introduce definitions, combining the analysis of bothprocesses and the results have shown that this was an important advance to research. Moreover, our studysuggested the mentioned possibility to expand the predictions of Shaver’s defensive hypothesis, including dif-ferences in the type and complexity of the victims’ explanations. According to Deschamps and Clemence(2000) there exists a great variability of causes within the internal and external categories. The two groupssharing external attributions (defensive) present some di!erences: the older group has an ‘‘it was just an acci-dent” perspective (Woodcock, 1995) with a (simple and defensive) bad luck explanations, while the othergroup numbered a variety of external factors present in the workplace (complex and defensive). On the otherhand, an occupational accident experience was associated with defensive definitions, focusing on the acci-dent’s characteristics and causes. It was mostly defined as something unexpected by the older and moresenior workers and defined as an event caused by organizational factors by the younger and lower seniorityworkers.

At the practical level, this study has important implications for organizational accidents record and preven-tion programs. The o#cial organizational definition of occupational accident is likely to be more similar tosome groups of employees than others, and this distance may interfere with the company process of collectingand registering occupational accidents data. To an employee that thinks an occupational accident is an eventresulting from his personal misconduct at work, a commuting accident may not be considered an occupationalaccident. Furthermore, if a worker defines an occupational accident as the result of poor working conditions,an injury due to a personal attention lapse will probably not be regarded as an event worth reporting. All thesepossible variations may leave many occupational accidents unrecorded and should be taken into account in aprecise collection of information about these hazardous occurrences. Moreover, the explanations given to theaccident influence worker’s future prevention behavior and the anticipation of this process are important forcorrective measures. The answer employees give to uncertain and dangerous events is congruent with the per-ceived causes and the careful assessment of the variety of attributions among workers can increase the likeli-hood of safety promotion programs success.

Acknowledgements

This study was funded by FCT (Foundation for Science and Technology – Portugal) with referenceFCT.PIQS/PSI/50070/2003. For more on this project see Silva et al. (2006,2007).

C. Niza et al. / Safety Science 46 (2008) 959–971 969

Author's personal copy

References

Bar-Tal, D., Kruglanski, A., 1988. The Social Psychology of Knowledge. CUP, Cambridge.Beauvois, J.L., Dubois, N., 1988. The norm of internality in the explanation for social events. European Journal of Social Psychology 18,

299–316.Butchart, A., Kruger, J., Leboka, R., 2000. Perception of injury causes and solutions in a Johannesburg township: implications for

prevention. Social Science & Medicine 50 (3), 331–344.Chapman, K., Abraham, C., Jenkins, V., Fallowfield, L., 2003. Lay understanding of terms used in cancer consultations. Psycho-

Oncology 12, 557–566.Condit, C., Dubriwny, T., Lynch, J., Parrott, R., 2004. Lay people’s understanding of and preferences against the word ‘‘mutation”.

American Journal of Medical Genetics 130A, 245–250.Conroy, C., Sciortino, S., 1997. Describing patterns of occupational agricultural deaths: the e!ect of case definition. Journal of Safety

Research 28 (4), 273–281.Dembe, A., 2004. The social consequences of occupational injuries and illnesses. American Journal of Industrial Medicine 40, 403–417.Deschamps, J.C., Clemence, A., 2000. L’explication quotidienne. PUR, Rennes.E"er, M., 1984. Attribution theories of lay epistemology. European Journal of Social Psychology 14, 431–437.Eurostat, 2004. Eurostat Questionnaire on national declaration systems of accidents at work. Instituto Nacional de Seguridad e Higiene en

el Trabajo for Eurostat, Madrid.Gherardi, S., Nicolini, D., Odella, F., 1998. What do you mean by safety? Conflicting perspectives on accident causation and safety

management in a construction firm. Journal of Contingencies and Crisis Management 6 (4), 202–212.Girasek, D.C., 1999. How members of the public interpret the word accident. Injury Prevention 5, 19–25.Gonc!alves, S.P.M., da Silva, S.A., Lima, M.L., Melia, J.L., 2008. The impact of work accidents experience on causal attributions and

worker behaviour. Safety Science 46 (6), 992–1001.Gyekye, S.A., 2003. Causal attributions of Ghanian workers for accident occurrence: miners and non-miners perspective. Journal of

Safety Research 34, 533–538.Gyekye, S.A., Salminen, S., 2006. The self-defensive attribution hypothesis in the work environment: co-workers’ perspectives. Safety

Science 44 (2), 157–168.Hofmann, D., Stetzer, A., 1998. The role of safety climate and communication in accident interpretation: Implications for learning from

negative events. Academy of Management Journal 41, 644–657.International Labour Organisation (ILO), 2002. Recording and notification of occupational accidents and diseases and ILO list of

occupational diseases. International Labour O#ce, Geneva.Jones, E.E., Nisbett, R.E., 1971. The actor and the observer: Divergent Perceptions of the Causes of Behavior. General Learning Press,

New York.Kelley, H.H., Michela, J.L., 1980. Attribution theory and research. Annual Review of Psychology 31, 457–501.Kouabenan, D.R., 1998. Beliefs and the perception of risks and accidents. Risk Analysis 18 (3), 243–251.Kouabenan, D.R., 1999. Explication naive de l’accident et prevention. PUF, Paris.Kouabenan, D.R., 2000a. Explication ordinaire des accidents, perception des risques et srtategies de protection. Pratiques Psychologiques

1, 85–97.Kouabenan, D.R., 2000b. Naive causal analysis and accident prevention strategies. Urban Transportation and Environment 14, 881–886.Kouabenan, D.R., 2000c. Decision, perception du risque et securite. In: Bernaud, J.L., Lemoine, C. (Eds.), Traite de Psychologie du

Travail et des Organisations. Dunod, Paris, pp. 279–321.Kouabenan, D.R., 2001a. Culture, perception des risques et explications des accidents. Bulletin de Psychologie 54 (3), 329–342.Kouabenan, D.R., 2001b. Management de la securite : Role des croyances et des perceptions. In: Levy-Leboyer, C., Huteau, M., Louche,

C., Rolland, J.P. (Eds.), RH Les Apports de la Psychologie du Travail. Editions d’Organisation, Paris, pp. 453–476.Kouabenan, D.R., Gilbert, D., Medina, M., Bouzon, F., 2001. Hierarchical position, gender, accident severity and causal attribution.

Journal of Applied Social Psychology 31 (3), 553–575.Melia, J.L., Chisvert, M., Pardo, E., 2001. Un modelo procesual de las atribuiciones y actitudes ante los accidentes de trabajo: Estrategias

de medicion e intervencion. Revista de Psicologia del Trabajo y de las Organizaciones 17 (1), 63–90.Moore, J.E., 2000. Why is this happening? A causal attribution approach to work exhaustion consequences. Academy of Management

Review 25 (2), 335–349.Niza, C., Silva, S., Lima, L., 2006. Work accidents in the empirical literature: implications for the future. In: Guedes Soares, C., Zio, E.

(Eds.), Safety and Reliability for Managing Risk. A.A. Balkema Publishers, Rotterdam, pp. 801–808.Pernanen, K., 1993. Causal attributions in the explanation of alcohol-related accidents. Addiction 88, 897–906.Poole, G., 1987. Estimates of preventability and their relation to health behaviour. In: Paper presented at the Annual Meeting of the

Western Psychological Association, Long Beach, April, pp. 23–26.Ross, L., 1977. The intuitive psychologist and its shortcomings. In: Berkowitz, L. (Ed.), . In: Advances in Experimental Social Psychology,

vol. 10. Academic Press, NY.Rynes, S., Hillman, A., Ireland, R., Kirkman, B., Law, K., Millar, C., Rajagopalan, N., Shapiro, D., 2005. Everything you always wanted

to know about. Academy of Management Journal (AMJ) 48 (5), 732–737.Saldana, M., Herrero, S., Manzanedo del campo, M., Ritzel, D., 2003. Assessing definitions and concepts within the safety profession. The

International Electronic Journal of Health Education 6, 1–9.Salminen, S., 1992. Defensive attribution hypothesis and serious occupational accidents. Psychological Reports 70, 1195–1199.

970 C. Niza et al. / Safety Science 46 (2008) 959–971

Author's personal copy

Shaver, K.G., 1970. Defensive attribution: E!ects of severity and relevance on the responsibility assigned for an accident. Journal ofPersonality and Social Psychology 14, 101–113.

Silva, S., Baptista, C., Lima, L., 2004. OSCI: an organizational and safety climate inventory. Safety Science 42 (3), 205–220.Silva, S.A., Lima, M.L., Castro, P., Pinto, A.M., Melia, J.M., Correia, R., Gonc!alves, S., Niza, C., 2006. Prevenir e melhor que remediar:

um olhar sobre os impactos dos acidentes de trabalho [Prevention is better than remedy: one look on the impact of work accidents].CIS, Lisboa, ISBN 989-95211-0-8.

Silva, S.A., Lima, M.L., Castro, P., Pinto, A.M., Melia, J.M., Correia, R., Gonc!alves, S., Niza, C., 2007. Relatorio Final do Projecto:Impactos dos acidentes de trabalho: suas valencias ao nıvel social, organizacional e individual (PIQS/PSI/50070/2003) [Final Report ofthe Project: Impacts of work accidents: valences at social, organizational and individual level]. CIS, Lisboa.

Silvester, J., Chapman, A.J., 1997. Asking ‘‘why” in the workplace: causal attributions and organizational behavior. Journal ofOrganizational Behavior, 1–14.

Walster, E., 1966. Assignment of responsibility for an accident. Journal of Personality and Social Psychology 3, 73–79.Weddle, M., 1996. Reporting occupational injuries: the first step. Journal of Safety Research 27 (4), 217–223.Woodcock, K., 1995. Bias in real world accident cause-finding. In: Bittner, A.C., Champney, P. (Eds.), Advances in Industrial Ergonomics

and Safety. Taylor & Francis, London, pp. 907–911.

C. Niza et al. / Safety Science 46 (2008) 959–971 971