af

gs C

offrkof de uionK Conbcotlycan

2012 Elsevier Ltd. All rights reserved.

1. Introduction

of induareneholds,y, 200nitivep. 97)in a vond the

ter in the Gulf of Mexico which killed 11 men and caused the worstoil spill in US history is testimony to the very hazardous nature ofthis industrys activities. Ongoing examination of the causal eventsindicates failures in situation awareness and risk assessment (Re-port to the President, 2011). Analysis of earlier drilling industry

activities. Drilling mud is introduced into the centre pipe in orderto counterbalance internal pressures and to oat the rock cuttingsback to the surface where they are extracted from the hole, makingthe process slippery. Additional sections of drill pipe are added asthe well gets deeper forming the drill string. The drill bit oftenneeds to be replaced due to different rock compositions and in or-der to do this, the entire drill string has to be removed from thewell and the pipes stacked in order to reach the drill bit. This is ahazardous process due to the slippery mud from the cuttings. Oncethis process has been completed, oil or gas ows up through wellby placing a smaller-diameter pipe (tubing) into the casing and a

Corresponding author. Tel.: +44 1224 273210; fax: +44 1224 273211.E-mail addresses: anne.sneddon@innovene.com (A. Sneddon), k.j.mearns@

gmail.com (K. Mearns).1 Present address: Ineos Manufacturing Scotland Limited, Boness Road, Grange-

Safety Science 56 (2013) 8088

Contents lists available at

Safety S

.emouth FK3 9XH, United Kingdom.in the near future.Cognitive skills such as situation awareness are known to be

susceptible to the effects of work-related conditions such as fatigueand stress (Endsley, 1999; Sexton et al., 2000; Tucker et al., 2010)which are common in many high-risk industries, for example inoffshore oil and gas exploration, where personnel work on remoteinstallations, often in time-pressured, dangerous conditions (Flinand Slaven, 1996). The recent Deepwater Horizon drilling rig disas-

They may be based in the drill cabin, using advanced computersystems to provide them with a clear view of the task and real timedata logging. The drill crew have to manually handle heavy equip-ment since the process is not entirely automated. The well isdrilled into the sub-sea oil reservoir or gas eld, which involvescarefully positioning the drill bit, collar and drill pipe into the well.This assembly is then attached to the kelly and rotary table whichrotates, lowers and raises the drill pipe in order to carry out drillingCritical factors in the preventionthe ability of workers to maintain awment, understand the information ittions will develop (Jones and EndsleThe term used in industry for this cogness (SA), dened by Endsley (1988,the elements in the environment withthe comprehension of their meaning, a0925-7535/$ - see front matter 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.ssci.2012.05.027strial accidents includess of the work environ-and predict how situa-0; Stanton et al., 2001).skill is situation aware-as . . . the perception oflume of space and time,projection of their status

accidents, such as the Montara blowout in 2009 off Australia(Hayes, 2012), and on the UK Continental Shelf (UKCS) also identi-ed failures to attend to relevant information in the work environ-ment as a common contributory factor (Sneddon et al., 2006).

Drilling activity is a critical and challenging process in hydro-carbon exploration and production, especially for the increasinglyhazardous deepwater wells (Skogdalen et al., 2011). Drillers haveto maintain control of the well, lead work on the drill oor (some-times involving heavy, physically demanding work), but also dealwith advanced technological equipment and monitoring facilities.to increased participation in unsafe behaviour. Individuals who had previously been involved in a workaccident had signicantly lower WSA scores than those who had not had an accident.Stress, fatigue, situation awareness and s

Anne Sneddon 1, Kathryn Mearns , Rhona FlinIndustrial Psychology Research Centre, School of Psychology, University of Aberdeen, Kin

a r t i c l e i n f o

Article history:Available online 19 July 2012

Keywords:DrillingOffshore industryStressFatigueSituation awareness

a b s t r a c t

Drilling for oil and gas onmaintain high levels of wooped to measure the WSAfactors of stress and fatigudent involvement. A questworking offshore on the Uable internal reliability (Cripation and distraction hadand fatigue were signicanfound to be the only signi

journal homepage: wwwll rights reserved.ety in offshore drilling crews

ollege, Old Aberdeen AB24 3UB, United Kingdom

shore installations is a hazardous occupation, and requires personnel tosituation awareness (WSA). This paper presents a self-report scale devel-rilling personnel, and examines the inuence of the performance shapingpon WSA, and the relationship between WSA, unsafe behaviour and acci-naire designed to measure these variables was completed by 185 drillersontinental Shelf (UKCS). The total WSA scale was found to exhibit accept-achs alpha = 0.86). Sub-scales measuring concentration; attention; antic-efcients between 0.65 and 0.79. Higher levels of stress, sleep disruptionassociated with lower levels of WSA. In a regression analysis, stress wast predictor of WSA. In relation to safety outcomes, lower WSA was related

SciVerse ScienceDirect

cience

lsevier .com/locate /ssc i

years (see Salmon et al., 2008) however at the time this study wasconducted, this literature was not available for consideration.

y SciWe therefore decided to develop a self-report measure whichcould be used with workers in the high risk drilling environmentwithout the presence of an observer, or the interruption of tasks.This measure was intended to indicate an individuals level of SAas a general measure, rather than a transient, task-dependent, statemeasure. Individual differences in situation awareness have beendocumented (Gugerty and Tirre, 2000) and this is an emerginginterest, as two trait-based, self-report measures have recentlybeen developed. The Workplace Cognitive Failures Scale (Wallaceand Chen, 2005), is a 22-item scale developed to assess cognitivefailures in the workplace. The Factors Affecting Situation Aware-ness (FASA) (Banbury et al., 2007) is a measure of a pilots acquisi-tion and maintenance of SA. These scales were not available forconsideration during the design stage of the present study but theycould be useful in the future as measures of SA or to provide asource for testing concurrent validity with the current measure-ment instrument if future work was being conducted in the off-shore drilling sector.

Several general (i.e. not workplace) trait measures of attentionand cognitive disposition were scrutinised, such as the ShortInventory of Minor Lapses (Reason and Lucas, 1984), the Everydaypacker down the outside to form a seal round the tubing. A deviceknown as a Christmas tree is attached at the top of the tubingallowing the drill crew to control the ow from the well. Due tothe pressures at the depths where hydrocarbons are found,blow-outs are possible and are an added risk to the process (Sko-gdalen et al., 2011). Blow-out valves are placed on the seabed in or-der to stabilise the pressure and control the well when necessarybut the potential for a catastrophic situation is very real for thedrill crew.

It is apparent from the description above that SA is a critical is-sue inuencing safety in the drilling industry but little research hasbeen undertaken focusing on SA in this area. In particular, there is alack of offshore specic SA models and measures available andthere is also limited evidence of the role of stress and fatigue inaffecting global SA. The current study was designed to rst developa measure of global SA in offshore drilling crews and secondly touse it to measure the relationships between stress, fatigue and sit-uation awareness in offshore drilling personnel on the UKCS, and todetermine whether situation awareness (SA) is associated withsafety outcomes such as unsafe behaviour, near misses and acci-dent history.

2. Measuring SA

Several methods for measuring SA have been developed, usuallyas a task-based, state characteristic, often using simulators (seeSalmon et al., 2006 for a review). Several of these were consideredfor the current study, including the Situation Awareness RatingTechnique (SART) (Taylor, 1990) where after task completion,respondents rate factors affecting their performance and under-standing to give a global measure of SA. This method was deemedunsuitable as it measures SA for a specic task, while the aim ofthis research was to assess a more global estimate of SA. The Situ-ation Awareness Global Assessment Technique (SAGAT) (Endsley,1988) is used to assess a participants SA when a simulated taskis interrupted. This was also rejected as no drilling simulator facil-ities were available to the researchers at the time this study wasconducted. We appreciate that there have been further develop-ments in SA theoretical underpinnings and measurement in recent

A. Sneddon et al. / SafetAttention Questionnaire (Martin, 1983), and the MindfulnessAttention Awareness Scale (Brown and Ryan, 2003). They were re-jected because of limited validity and reliability data and/or itemswere not appropriate for assessing awareness in a work environ-ment. The Cognitive Failures Questionnaire (CFQ Broadbentet al., 1982) was also reviewed. The CFQ provides a robust measureof everyday attention and lapses but is not work specic and didnot cover issues that would be relevant to a drilling industry situ-ation. It was therefore decided to develop a new trait SA measure-ment technique, the work situation awareness (WSA) scalespecically aimed at measuring general awareness of the drillingwork environment, based on an adaptation of the CFQ.

3. Factors affecting SA and attentiveness

There is limited empirical evidence regarding the factors affect-ing general levels of SA: many studies focus only on specic atten-tional processes such as vigilance, and do not consider awarenessin a broader sense. Two workplace-related conditions that havebeen more widely reported in the literature as impacting SA arestress and fatigue.

3.1. Stress and SA

Increasing levels of stress can result in reduced working mem-ory capacity and diminished attention (Hockey, 1986; Hancock andSzalma, 2008). Stress can result in poor concentration/alertnessdue to an overload on the individuals cognitive resources, and thiscan interfere with the primary perception of the situation, causinginattention to the available information. Consequently, there maybe a narrowing of the individuals attentional eld to incorporateonly a restricted number of core aspects, resulting in peripheralinformation receiving little or no attention. While this cognitivetunnel vision (Tversky and Kahneman, 1974) may be a valuableadaptive strategy in a safety critical environment by preventingoverload, factors outside the central focus of attention may bethose that have most potential to be harmful. Relatively high levelsof occupational stress have been measured in offshore studies,(Mearns and Hope, 2005; Parkes, 1998) and associations betweenstress and offshore accident rates have also been established (Suth-erland and Cooper, 1986, 1996).

3.2. Fatigue, sleep disruption and SA

Fatigue also causes detriments to alertness levels and conse-quently increases the risk of accident involvement (HSE, 2006),as the cognitive resources required are depleted due to physicalexertion or sleep deprivation (Rosekind et al., 1994). Dawson andReid (1997) found that decits in cognitive processing in individu-als with only moderate sleep deprivation were akin to those expe-rienced when blood alcohol levels are over the legal limit fordriving. The effects of fatigue are to generally decrease the speedof cognitive processing, and thus increase reaction times, tunnel vi-sion, inattentiveness, and lower vigilance and concentration(Helmreich et al., 2004). These effects have been reported in themaritime industry (Smith, 2001; Wadsworth et al., 2008), trans-portation (Fletcher and Dawson, 2001), and power generation(Ognianova et al., 1998) and have also been reported for the off-shore oil and gas industry as outlined below.

4. The working environment in the offshore drilling industry

Managers in the offshore oil and gas industry report that lack ofcare and attention is one of the main causes of accidents (ODea andFlin, 2001). This has particular relevance for drilling personnel, who

ence 56 (2013) 8088 81are involved in one of the most dangerous activities, running long,heavy pipes into hydrocarbon reservoirs under the sea bed in a fastoperation. They must be able to continuously monitor and under-

stand the drill oor environment if they are to keep their accidentrisk to a minimum. Occupational stress is a feature of offshore life,originating from the usual sources, the offshore living environment,helicopter travel, and the interface between job and family (Parkes,1998; Sutherland and Cooper, 1986, 1996; Sutherland and Flin,1989). Fatigue and sleep disruption are common. Drilling crews of-ten work 12-h shifts for 14 or more days with no rest days. Manylocations work a shift pattern (known as short change or mid-

nel also may share an accommodation cabin, which can disturb

5. Relationships between performance shaping factors, work SA

tive failures were a predictor of occupational safety behaviour, inthat individuals reporting more cognitive failures also reported in-creased safety non-compliance and more accidents, and Wads-worth et al. (2003) showed that occupational accidents wereassociated with increased cognitive failures.

Hypothesis 2a. WSA will be negatively associated with unsafebehaviour.

company the remainder were employed by the operating oilcompany or another contracting company. A total of 44% were

A self-report questionnaire survey was used to collect data on

and (due to logistical limitations for offshore trips), postal distribu-

82 A. Sneddon et al. / Safety Science 56 (2013) 8088and accidents

Examining the relationships between performance shaping fac-tors such as stress and fatigue and work SA (WSA) may identifytheir relative contribution to accident involvement. The relation-ships being examined in the study are illustrated in Fig. 1 and ex-plained below.

N.B. Only volitional non-compliance was measured, due to thedifculty of measuring non-volitional non-compliance (e.g. forget-ting), as by their very nature, individuals may not be aware ofthem.

Fig. 1 proposes that fatigue and stress will have a detrimentalimpact upon WSA, and that as a result workers with lower WSAwill have more accidents and near-misses, and report more unsafebehaviours, due to their attention and alertness being reduced. It isproposed that WSA is a key part of the explanatory mechanism forwhy stress and fatigue are related to workplace accidents.

5.1. Hypotheses

As part of the validation process (see above) it was predictedthat the Cognitive Failure Questionnaire (CFQ) scores would corre-late negatively with WSA scores. The nature of the scoring on thescales means that higher WSA scores represent better SA whereashigher CFQ scores represent more cognitive failures. This valida-tion was conducted before a series of hypotheses were tested.

Hypothesis 1a. Stress will be negatively associated with WSA.

Hypothesis 1b. Sleep disruption will be negatively associated withWSA.

Hypothesis 1c. Fatigue will be negatively associated with WSA.It is proposed that WSA will have a subsequent effect upon per-

sonal safety outcomes such as unsafe behaviour, accidents andnear-misses. Wallace and Vodanovich (2003a,b) found that cogni-

Fatigue

Sleep disruption

Stress

WSA relaxation time and sleep (Mearns and Hope, 2005).hitch roll over) which involves personnel changing half-waythrough their stint offshore from day-shift to night-shift or vice ver-sa), disrupting sleeping patterns (Gibbs et al., 2005). Conditionsgenerally tend to be noisy due to machinery. There are high num-bers of personnel living and working in a limited area, and person-Fig. 1. Proposed relationships between fatigue, sleep disruption, stressAccidents

Near-misses

Unsafe Behaviour cognitive failures, WSA, stress, sleep disruption, fatigue, unsafebehaviour and accident history. Two modes of distribution wereused: personal offshore site visits by one of the research teamsupervisors, and 74% had worked at their present location (rig)for 5 years or less. Of the respondents, 66% worked a 2-week triprotation pattern (2 weeks offshore then 2 weeks leave onshore),44% worked a rolling shift pattern of one trip of day shift followedby one trip of night shift and 34% had on-call duties.

6.2. ProcedureHypothesis 2b. WSA will be negatively associated with rates ofaccident involvement.

Hypothesis 2c. WSA will be negatively associated with rates ofnear-miss occurrence.

It is also proposed from the above that WSA mediates the rela-tionship between performance shaping factors (fatigue, sleep dis-ruption, stress) on unsafe behaviour.

Hypothesis 4. WSA mediates the relationship between the per-formance shaping factors and unsafe behaviour.

6. Method

6.1. Sample

The sample consisted of drilling personnel (n = 378) based oneight drilling rigs and platforms on the UKCS. All locations werecontracted to operate for one multi-national operating companyat the time of the survey. A total of 185 (49%) questionnaires re-turned were viable for analysis. This is an acceptable response ratefor this remote sector, and is comparable to that found in otherstudies (e.g. Mearns et al., 2006). Respondents included all levelswithin the drilling hierarchy, from roustabout to drilling supervi-sor. Respondents indicated which age group they belonged to,rather than giving their actual age so that anonymity would notbe compromised. The mean age group of respondents was 3544 years. Of the sample, 77% were employed directly by the drillingwith WSA and unsafe behaviour, accident/near miss involvement.

2006). It contained 20 items (Appendix A), and was scored on a

purposes, as it is an established measure for assessing slips of

gue scale was modied to make the 13 items more suitable for

y Sci6.3.4. Sleep disruptionThe sleep disruption scale also came from the AMSA set of mar-

ine measures and contained 14 items, assessing how often sleepwas disrupted offshore (or the onset of sleep delayed). As abovethe offshore work domain. They assessed to what extent boredom,number of days into the trip, working with a crewwho are not fullycompetent, and working a night shift contributed to respondentsfeelings of tiredness, fatigue or decreased alertness. The items wererated on a 5-point scale (ranging from 1 = never to 5 = always).attention in everyday life. It has been shown to have high internalreliability, i.e. the items group together to measure the sameunderlying construct (e.g. Larson et al., 1997; Vom Hofe et al.,1998; Wallace, 2004), is well validated with other scales and hasbeen found to be a signicant predictor of accidents (Wallaceand Vodanovich, 2003a,b). It contains 25 items, with responsesindicating how often particular situations indicating failures ofattention and cognition have happened to the respondent withinthe last 6 months, and is answered on a 5-point scale (4 = very of-ten to 0 = never).

6.3.3. FatigueTo measure levels of fatigue, a scale developed by the Australian

Maritime Safety Authority (AMSA; Parker et al., 1998) was selectedsince it had been developed for a marine environment, similar tothat experienced by the offshore oil and gas industry both with re-gard to the physical environment and the shift patterns. The fati-5-point scale (0 = very often to 4 = never; i.e. the higher the score,the better the individuals awareness of the work environment).Items were customised for the offshore drilling environment. Theyincluded 5 positively worded statements such as I take note of ob-jects/events on the rig even if they are not directly related to mywork, I think ahead of my work to plan for different possible out-comes and 15 negatively worded (reverse scored) statements Iam easily distracted by my thoughts and feelings. A nal questionasked at what time of the shift/rotation respondents felt leastaware.

6.3.2. Cognitive failuresThe CFQ (Broadbent et al., 1982) was included for validation6.3. Measures

The questionnaire consisted of ve sections: WSA, cognitivefailures, fatigue and sleep disruption, safety behaviour, and acci-dent history.

6.3.1. Work SA (WSA)This scale was developed to measure specic awareness of the

work environment on the drilling rig. It was adapted from the Cog-nitive Failures Questionnaire (Broadbent et al., 1982) and drew onour previous work on situation awareness in drilling, which in-volved interviews with experienced drillers (Sneddon et al.,tion. Respondents were given envelopes in which to return thecompleted questionnaires to the research team. No signicant dif-ferences were found on any measure between the samples sur-veyed using the two methods.

A. Sneddon et al. / Safetthese items were listed on a 5-point scale (ranging from 1 = neverto 5 = always). For both these scales, a higher score indicated great-er sleep disruption and fatigue.6.3.5. StressStandard occupational stress scales were rejected due to their

length, or the unsuitability of the items. Instead the measure ofstress was derived from offshore stress scales (Parkes, 1998; Suth-erland, 1994; Sutherland and Cooper, 1996) into a list of 32 itemscustomised for the drilling industry. The stressors included workoverload, threat of job loss, demands of work on private life, andmaking mistakes. Respondents rated how much stress they per-ceived from each of the items on a 6-point scale (ranging from0 = no stress to 5 = extreme stress).

6.3.6. Unsafe behaviourThis measure was the safety behaviour scale of the Offshore

Safety Questionnaire (OSQ) (Mearns et al., 1997), which assessesto what extent respondents participate in short-cuts and violatingbehaviours. It has been used in a number of offshore oil industrystudies (e.g. Mearns et al., 2003, 2010) and has acceptable internalreliability. The Cronbachs alpha of the scale in the current studywas 0.88. The 11 items were rated on a 5-point scale (1 = neverto 5 = always), modied from the three-point scale used in theOSQ, as this was felt to limit respondents answers. Higher scoresrepresented more unsafe behaviour.

6.3.7. History of accidents and near-missesThe nal section recorded respondents accident history while

working offshore. They were asked if they had been involved inan accident at any time in their offshore career; if they had expe-rienced an accident on board the rig within the last 12 months thatrequired a trip to the medic; and if they had experienced a near-miss on board the rig in the last 12 months. Four nal items askedfor: (a) very brief details of the accident or near-miss; and (b) thetime of day/period of trip the accident or near-miss. We acknowl-edge that the respondents may have been unwilling to report suchincidents due to the possibility of reprisal but by assuring con-dentiality of the results we hoped that this unwillingness wouldbe counteracted to some extent.

6.4. Analyses

The structure of the WSA was examined with a principal com-ponents analysis. To test relationships between stress, fatigueand WSA levels, correlations (Pearsons Product Moment) andregression analyses were used. The Sobel test was run to test formediation effects of WSA between stress and safety non-compliance.

7. Results

7.1. Principal components analysis

Principal components analysis (using the Varimax technique,which results in a rotated, orthogonal solution for the matrix) ofthe WSA scale was conducted, in order to test the factor structure.A loading level of 0.4 was set, as suggested by Field (2005). A four-factor structure (see Table 1) emerged, accounting for 53.6% of thevariance. They were labelled as follows: concentration; attention;anticipation and distraction.

7.2. Hypotheses results

The total WSA scale was found to have a Cronbachs alpha of0.86, and was highly correlated with the total CFQ (r = .70,

ence 56 (2013) 8088 83p < 0.01), conrming Hypothesis 1 that the scales are measuringsimilar underlying constructs. The four sub-scales also had alphasranging from 0.65 to 0.83. It is acknowledged that an alpha of 0.65

d itimeorin

upt

vemwithif I a

y Sci ce 56 (2013) 8088Table 1Principal components analysis of the WSA scale.

Factor one Concentration

Loading .72 I am not able to keep my mind focused on work an.69 I nd it difcult to concentrate for long periods of t.69 I tune out during routine work, or when work is b.62 My work area is often cluttered or disorganised.60 I have trouble getting back into work after an interr.59 I become bored with my work quickly.55 I am easily distracted by visual stimulation (i.e. mo.52 I often nd I have carried out work on auto-pilot,.47 I nd it difcult to pay attention to someone, even

Cronbachs alpha .83

84 A. Sneddon et al. / Safetis rather low (Nunnaly, 1978) but given that we were developing anew scale, we believed it was acceptable to retain this scale for cur-rent purposes.

Table 2 displays the bivariate correlations between the four ex-tracted factors and the other test variables, while Table 3 showsthe accident history of the group.

It was found that higher levels of stress had a negative relation-ship with WSA (Hypothesis 2b), as did higher levels of sleep dis-ruption (Hypothesis 2c) and fatigue (Hypothesis 2d).

Consistent with expectations, lower levels of WSA were signif-icantly related to increased unsafe behaviour in the workplace(Hypothesis 3a), (r = .51). This could be because the people whoadmit to having lower WSA are also the people who are willingto admit to non-compliance. Table 3 shows the self-reported fre-

Table 2Means, standard deviations and pearson correlation coefcients between the measured va

Variable Mean SD 1 2

1. Overall WSA 53.42 7.95 2. WSA concentration 22.98 4.79 .89** 3. WSA anticipation 15.05 2.57 .53** .18*

4. WSA attention 8.26 1.44 .61** .47**

5. WSA distraction 7.30 1.93 .75** .63**

6. Sleep disruption 30.66 7.11 .32** .28**7. Fatigue 31.03 7.49 .40** .36**8. Stress 57.23 26.24 .44** .40**9. Unsafe behaviour 17.45 5.32 .51** .51**

* p < 0.05.** p < 0.01.

Factor two Anticipation

Loading .72 I ensure I know most rig activities that are ongoing so I.72 I think ahead of my work to plan for different possible.69 I take note of objects/events on the rig even if they are.64 I nd it easy to keep track of everything that is going o.54 I nd it easy to remember work instructions

Cronbachs alpha .69

Factor three Attention

Loading .82 I often speak or act without thinking.62 I often have difculty paying close attention to details,.60 When I nish reading or being told instructions, I often

Cronbachs alpha .65

Factor four Distraction

Loading .66 I am easily distracted by background noise.63 I often daydream during work.50 I am easily distracted by my thoughts or feelings

Cronbachs alpha .73

Table 3Self-reported accident and near-miss involvement.

Yes % No % n

Accident history at any time in career 92 51 90 49 182Accident in past 12 months 3 2 112 98 115Near-miss in past 12 months 24 14 149 86 173has a tendency to wander

g

ion

ent)out being aware of itm being spoken to directly

can keep an eye on thingsoutcomesnot immediately related to my workn around meenquency of involvement in an accident or near-miss. Hypothesis3b was supported: those who had previously experienced an acci-dent were found to have signicantly lower WSA than those whohad never experienced an accident (t(166) = 2.33, p < 0.05). Incontrast, no support was found for the hypothesis that individualswith lower levels of self-report WSA are more likely to have beeninvolved in a near-miss than those with higher levels, (t(158) =1.44, p = ns). This could be due to the fact that the participantsdid not want to report near misses, although reporting near missescould be seen as less threatening than reporting accidents. Individ-uals who had experienced an accident had signicantly higher un-safe behaviour scores than individuals who had not (t(171) = 2.07,p < 0.05).2 Those who had experienced a near-miss in the last 12-months reported that they engaged in signicantly more unsafebehaviour than those who had not (t(163) = 3.76, p < 0.01).

In addition to the above analyses, a linear multiple regressionanalysis (Field, 2005) was carried out to determine what combina-tion of workplace variables predicted global WSA (see Table 4).

riables.

3 4 5 7 8 9 10

.23** .19* .36**

.08 .22** .33** .14 .26** .36** .67** .24** .28** .32** .49** .67** .21** .37** .35** .17* .33** .39**

which often results in careless errorshave to re-read them or ask for them to be repeated as I dont remember them

2 This calculation used the accident statistics from the item Accident history atany time in career rather than Accident in past 12 months as only threerespondents had actually been involved in an accident in the last 12 months.

y SciPredictor variables b t SE

Stress .34 3.74** .03Fatigue .14 1.30 .11Sleep disruption .06 0.64 .10F(3,173) = 12.77, p < 0.001, adjR2 = 0.22

** p < 0.01.

Table 5Multiple regression models predicting factors concentration, anticipation, atten-tion and distraction from the WSA scale.

b t SE

Concentration predictor variablesStress .31 3.34** .17Fatigue .13 1.17 .07Sleep disruption .04 .41 .06F(3,173) = 13.83, p < 0.001, adjR2 = 0.17

Anticipation predictor variablesStress .29 2.92** .01Sleep disruption .06 .64 .04Fatigue .01 .07 .04F(3,172) = 4.12, p < 0.01, adjR2 = 0.05

Attention predictor variablesStress .19 1.98* .01Sleep disruption .07 .73 .02Fatigue .08 .71 .02F(3,176) = 5.76, p < 0.001, adjR2 = 0.08Table 4Multiple regression model predicting global WSA.

A. Sneddon et al. / SafetStress was the only factor to make a signicant contribution tothe model, indicating that those who report higher stress levelsalso report poorer WSA. The model explains 22% of the variancein the global WSA scores.

Regression analyses for each of the four WSA factors were alsoconducted (see Table 5).

Similar to the regression conducted for global WSA, the onlysignicant predictor was stress, explaining 17%, 5% and 8% of thevariance in the WSA factors concentration, anticipation andattention, respectively (those who report higher stress also reportpoorer concentration, projection and attention levels). When thevariables were entered to predict distraction, sleep disruptionwas the only signicant predictor, explaining 17% of variance.

Earlier analyses indicated that stress was a signicant predictorof unsafe behaviour at work, and so the next step was to testwhether WSA was a mediator of the relationship between stressand unsafe behaviour. Baron and Kenny (1986) recommend usingthe Sobel test (Sobel, 1982) to identify the percentage of the totaleffect that is mediated, and the ratio of the indirect to the direct ef-fect (see Preacher and Hayes, 2004). A syntax program was down-loaded from the internet (http://www.ats.ucla.edu/STAT/spss/faq/mediation.htm). Tables 6 and 7 display the results.

The results show that WSA mediates 47.82% of the relationshipbetween stress and unsafe behaviour, therefore accounting for al-most half of the effect.

8. Discussion

Previous studies have demonstrated that cognitive failure is re-lated to workplace safety behaviour and accident occurrence, and

Distraction predictor variablesSleep disruption .15 1.65** .03Stress .15 1.60 .01Fatigue .16 1.36 .03F(3,174) = 10.51, p < 0.001, adjR2 = 0.14

* p < 0.05.** p < 0.01.this study aimed to discover how different occupational factorsinherent to the offshore drilling industry can affect attentiveness,and subsequent accident risk.

Regarding Hypothesis 1, a signicant negative association wasfound between the Cognitive Failures Questionnaire and WSA,indicating that the more mishaps and lapses of attention an indi-vidual reported, the less workplace awareness he/she reported.This suggests that the skills required to maintain attention ineveryday life are similar to the abilities that control attentivenessin the offshore drilling environment, as it is only the context thatchanges, and not the mental activities. Reason (1988) suggestedthat some people are more likely to experience cognitive failuresdue to their more rigid style of cognitive management and atten-tional focus. This could be an area for further investigation, partic-ularly in high hazard domains where more exible styles ofcognitive management and attentional focus may be necessary tokeep the appropriate perspective on ongoing operations. Suchstyles could either be selected for or trained for, but only if theWSA scale can be shown to have predictive validity for perfor-mance in the drilling sector. The development of more realisticsimulators since this study was conducted, provides the opportu-nity to test the measure in such controlled environments.

Individuals reporting higher levels of stress were found to havepoorer WSA. The literature on indicates that stress has a tendencyto cause individuals to narrow their eld of attention (Endsley,1995) and can impair cognitive resources by undermining workingmemory (Hockey, 1986). Higher levels of sleep disruption and fati-gue also correlated with decreasedWSA. This corroborates Wallaceet al.s (2003) nding that individuals who scored higher on day-time sleepiness also experienced more cognitive failures. Sleep dis-

Table 6Summary of regression analysis results for WSA between stress and unsafe behaviour.

Variables Regression coefcient SE

Stress and unsafe behaviour 0.08** 0.14Stress and WSA 0.14** 0.02WSA and stress on unsafe behaviour 0.26** 0.05

** p < 0.01.

Table 7Sobel test results for WSA between stress and unsafe behaviour.

Sobel test Percentage of the totaleffect that is mediated

Ratio of the indirect todirect effect

3.76** 47.82 0.92

** p < 0.01.

ence 56 (2013) 8088 85ruption is part of working offshore and these ndings suggest thatthis is detrimental to employees by decreasing their WSA levels.Companies may wish to consider altering the shift patterns thatare in place to make them more stable, for example, allow workersto always work a day or night shift rather than switch shift pat-terns in the middle (split/swing shift), or installing extra soundproong in cabins to allow personnel to enjoy more undisruptedsleep.

Stress and fatigue were correlated but the multiple regressionanalysis showed that stress was the only signicant predictor ofWSA. When the component factors of WSA were examined in moredetail, stress was also found to be the only signicant predictor forconcentration, projection and attention. Although the effect of fati-gue appears to have been diminished in these analyses, sleep dis-ruption was still a predictor of distraction. Wallace et al. (2003)found that scores on the daytime sleepiness scale correlated signif-icantly with their distractions factor in samples of undergraduatestudents and military personnel, thus corroborating the resultsfound in this study.

data, so the usual caveats should be applied when consideringthe implications of the ndings. Clearly longitudinal or experimen-

y SciThe hypothesis that individuals who report poorer WSA willparticipate more frequently in unsafe work behaviours was sup-ported. In a study on production workers, Wallace and Vodanovich(2003a) also found that cognitive failure scores were positively re-lated to safety non-compliance. Wickens et al. found that drivingerrors, lapses and violations were predicted by cognitive failureand the loss of SA through use of mobile phones has been foundto be associated with trafc violations, e.g. speeding (Kass et al.,2007). These ndings suggest that unsafe work behaviours suchas taking short-cuts and breaking rules could be due to lapses ofattention or awareness, rather than deliberate violations. On theother hand, perhaps those who violate simply do not bother topay attention. It may be that this is an issue with the projectionstage of WSA, in that an individual is more likely to take a short-cut as he/she cannot (or is poor at) predicting the possible negativeoutcomes of this action. Conversely, those with better quality WSAmay be more able to accurately predict what may happen and aremore aware of the risks, and so consequently are less likely to carryout the unsafe behaviour.

Support was found for the hypothesis that individuals who hadbeen involved in an accident at some point during their offshore ca-reer would have signicantly poorer WSA scores. Likewise, Wallaceand Vodanovich (2003b) reported that in military electrical work-ers, cognitive failures were positively correlated with both automo-bile accidents and workplace accidents. Larson et al. (1997) alsofound that accidents and scores on the cognitive failures scale wereassociated. These results conrm the importance of maintaininggood quality SA in an attempt to successfully prevent accidents.There was no support for the hypothesis that individuals withpoorer WSA would be more likely to have been involved in anear-miss within the last 12 months. However, this is most likelydue to the fact that only 14% of the sample was in this category,and therefore the calculation had very little power. It was foundthat individuals who had been involved in an accident at any pointduring their offshore careerwere signicantlymore likely to engagein non-compliance at work than their colleagues who had neverexperienced an accident, as were those who had experienced anear-miss in the last 12-months (compared to those who had not).

8.1. Stress, WSA and safety

The results of a mediation test showed that 48% of the relation-ship between stress and unsafe behaviour was mediated by WSA,suggesting that WSA is an important construct when investigatingworkplace safety. High hazard/high reliability organisations shouldnote the results of this and similar studies and either use furthervalidated WSA scales in their selection processes or incorporateSA into their training programmes. It also important to note theimpact stress has on human performance and measures to reducethe impact of stress in the workplace should be implemented by allorganizations.

8.2. The WSA scale

The principal components analysis of the WSA scale produced afour-factor model (concentration, attention, anticipation and dis-traction). The CFQ is characterised by three components percep-tion, memory and motor function (Broadbent et al., 1982) andWallace and Chen (2005) also identied a three-factor model(memory, attention and action) as the best t for their Work Cog-nitive Failures Scale using conrmatory factor analysis. One possi-ble reason for the four factors emerging in the current study is thatthey are an artefact of the hazardous and fast moving drilling envi-

86 A. Sneddon et al. / Safetronment. Furthermore, distraction can affect all three stages of per-ception, comprehension, or anticipation, which could explain whythese distractive items clustered into a single factor.tal studies would be required to further test the impact of workfactors on SA in personnel in the drilling sector and other high riskindustries. Conducting research in the offshore environment bringswith it a specic set of challenges including accessing personneland following up on individuals who are often working as contrac-tors and therefore subject to regular changes in employer and theinstallation they are working on. Offshore tracking systems (e.g.the VANTAGE system, which records how long and the locationwhere personnel have been working offshore) could facilitate lon-gitudinal studies on the effects of work stress and fatigue on per-formance in the future.

Other factors to be considered include the work situation at thetime of the survey. When the questionnaires were being com-pleted, the operating company to which the drilling companieswere contracted had just put their drilling contract out to bid,and a new drilling company had won the tender. Subsequently,there was much unease in the workforce many staff had alreadybeen made redundant, and the remaining personnel did not knowwhether they were going to be transferred to the new drilling com-pany when they took over the contract in a few months time, or bemade redundant. While the researcher was on board, severalrespondents commented that the current conditions were causingmany to feel under more pressure than usual, and therefore thestress levels reported in the questionnaire may not be indicativeof the true stress levels of offshore workers in normal conditions.It would be also benecial to try to gain access to industry recordsof accident occurrence in order to gain a more objective report onaccident frequency, rather than relying solely on self-report meth-ods that may provide an underestimation of the actual number ofincidents and accidents.

While the response rate of the study was reasonable for this re-mote population (49%), the length of the questionnaire may haveinuenced response rates. While the initial research required thequestionnaire to be this length, future work could use a shortenedversion. As mentioned earlier, Wallace and Chens (2005) WorkCognitive Failures Scale (WCFS), assessed cognitive failures in theworkplace, while also investigating the associations betweenWCF and occupational safety. The WCFS was based upon Broad-bent et al.s (1982) Cognitive Failures Questionnaire, and containeditems based upon the constructs of memory, attention and action;when factor analysed, this factor structure was found to hold. Sim-ilar to the ndings reported above regarding WSA and non-compli-ance, Wallace and Chen (2005) found a signicant negativecorrelation between WCF scores and safety behaviour in produc-tion, construction and plant operating personnel (the more WCFreported, the less safety compliant respondents reported them-selves to be). Given their evolution from the CFQ, it is not surpris-ing that there was some degree of overlap between the itemscontained in both the WCFS and the WSA scale, for example, itemson daydreaming or distraction. The WCFS appears to be success-fully measuring cognitive failures in the workplace, and with fur-ther renement, the WSA scale can measure general workplaceattentiveness comparison of these scales would now be useful.

10. Conclusion

This study has developed a measure of situation awareness9. Limitations of the study and directions for future research

All results reported in this study are based on self-reported

ence 56 (2013) 8088(WSA) specically for use with offshore drilling crews. The WSAscale shows evidence of content, construct and concurrent validityhowever more work is required to establish discriminant and pre-

dictive validity. This study has shown that higher levels of stressand fatigue are linked to lower levels of WSA, which in turn areindicative of increased participation in unsafe work behaviours,and higher accident risk. These results need to be replicated bytesting in drilling simulators or even in longitudinal studies. Socialdesirability and presentation bias could have affected the results,despite the assurances of anonymity and condentiality and thesebiases should be controlled for, for example by using the MarloweCrown Social Desirability Scale (Crowne and Marlowe, 1960). Situ-ation awareness training is not generally used in the oil industryalthough it is provided in other high risk sectors (aviation, mari-time, nuclear) Crew Resource Management (CRM) (Kanki et al.,2010) and in many of these domains, non-technical skills, suchas SA, are regularly checked as part of licence revalidation (Flinet al., 2008). Following the Deepwater Horizon accident, the off-shore oil and gas industry is beginning to develop CRM syllabifor drill crew and the above results suggest that SA, as well as fa-tigue and stress management will need to be key components.

Given these ndings, and with recent accidents highlightingthat offshore drilling crews are employed in one of the most haz-ardous maritime occupations, it seems justiable to suggest thatoffshore companies may also benet from reviewing their workingpatterns and conditions to ensure that their impact on cognitiveskills is fully understood as part of their risk mitigation strategy.

References

Banbury, S., Dudeld, H., Hoermann, H., Soll, H., 2007. FASA: development andvalidation of a novel measure to assess the effectiveness of commercial airlinepilot situation awareness training. International Journal of Aviation Psychology17, 131152.

Baron, R.M., Kenny, D.A., 1986. The moderatormediator variable distinction insocial psychological research: conceptual, strategic and statisticalconsiderations. Journal of Personality and Social Psychology 51, 11731182.

Broadbent, D., Cooper, P., Fitzgerald, P., Parkes, K., 1982. The cognitive failuresquestionnaire (CFQ) and its correlates. British Journal of Clinical Psychology 21,116.

Brown, K., Ryan, R., 2003. The benets of being present: the role of mindfulness inpsychological well-being. Journal of Personality and Social Psychology 84, 822848.

Crowne, D., Marlowe, D., 1960. A new scale of social desirability independent ofpsychopathology. Journal of Consulting Psychology 24 (4), 349354.

Dawson, D., Reid, K., 1997. Fatigue, alcohol and performance impairment. Nature388, 235.

Endsley, M., 1988. Situation Awareness Global Assessment Technique (SAGAT). In:Proceedings of the National Aerospace and Electronics Conference (NAECON),IEEE, New York, pp. 789-795.

Endsley, M., 1995. Toward a theory of situation awareness in dynamic systems.Human Factors 37, 3264.

Endsley, M., 1999. Situation awareness in aviation systems. In: Garland, D., Wise, J.,Hopkin, V. (Eds.), Handbook of Aviation Human Factors. Lawrence ErlbaumAssociates, Mahwah, NJ.

Field, A., 2005. Discovering Statistics Using SPSS, second ed. Sage, London.Fletcher, A., Dawson, D., 2001. Field-based validations of a work-related fatigue

model based on hours of work. Transport Research Part F 4, 7588.Flin, R., Slaven, G. (Eds.), 1996. Managing the Offshore Installation Workforce.

Penwell Books, Oklahoma.Flin, R., OConnor, P., Crichton, M., 2008. Safety at the Sharp End. A Guide to Non-

Technical Skills. Ashgate, Aldershot.

anspo

awedia

me

den-re

ult

A. Sneddon et al. / Safety Science 56 (2013) 8088 87Acknowledgements

This research was sponsored by a studentship from Shell Explo-ration & Production (European ONEgas Asset). The views presentedhere are those of the authors and should not be taken to representthe position or policy of the funding body. We would like to thankall those drilling staff who gave their time to completequestionnaires.

Appendix A. The work situation awareness (WSA) scale

(Circle one number on each line)

I often daydream during workI am easily distracted by background noiseI nd it easy to remember work instructionsMy work area is often cluttered or disorganisedI am easily distracted by my thoughts or feelingsI think ahead of my work to plan for different possible outcomesI ensure I know most rig activities that are ongoing so I can keepI nd it difcult to pay attention to someone, even if I am beingI become bored with my work quicklyI nd it difcult to concentrate for long periods of timeI am easily distracted by visual stimulation (e.g., movement)I often nd I have carried out work on auto-pilot, without beingI take note of objects/events on the rig even if they are not imm

my workI nd it easy to keep track of everything that is going on aroundI have trouble getting back into work after an interruptionI am not able to keep my mind focused on work and it has a tenWhen I nish reading or being told instructions, I often have to re

for them to be repeated as I dont remember themI often speak or act without thinkingI have difculty paying close attention to details, which often reserrorsI tune out during routine work, or when work is boringGibbs, M., Hampton, S., Morgan, L., Arendt, J., 2005. Effects of Shift Schedule OnOffshore Shiftworkers Circadian Rhythms and Health (Research Report 318).HSE Books, Suffolk.

Gugerty, L., Tirre, W., 2000. Individual differences in situation awareness. In:Endsley, M., Garland, D. (Eds.), Situation Awareness. Analysis and Measurement.Lawrence Erlbaum Associates, Mahwah, NJ, pp. 249276.

Hancock, P.A., Szalma, J.L., 2008. Performance under Stress. Ashgate Publishing,Aldershot, England.

Hayes, J., 2012. Operator competence and capability lessons from the Montarablowout. Safety Science 50, 563574.

Helmreich, R., Musson, D., Sexton, J., 2004. Human factors and safety in surgery. In:Nora, P.F., Manuel, B. (Eds.), Surgical Patient Safety: Essential Information forSurgeons in Todays Environment. American College of Surgeons, Chicago, IL.

Veryoften

Quiteoften

Occasionally Veryrarely

Never

0 1 2 3 40 1 2 3 40 1 2 3 40 1 2 3 40 1 2 3 40 1 2 3 4

eye on things 0 1 2 3 4ken to directly 0 1 2 3 4

0 1 2 3 40 1 2 3 40 1 2 3 4

are of it 0 1 2 3 4tely related to 0 1 2 3 4

0 1 2 3 40 1 2 3 4

cy to wander 0 1 2 3 4ad them or ask 0 1 2 3 4

0 1 2 3 4s in careless 0 1 2 3 40 1 2 3 4

Hockey, G., 1986. Changes in operator efciency as a function of environmentalstress, fatigue and circadian rhythms. In: Boff, K., Kaufman, L., Thomas, J. (Eds.),Handbook of Perception and Human Performance. Wiley, New York, pp. 44/4144/49.

HSE, 2006. The Development of a Fatigue/Risk Index for Shiftworkers, RR446. HSEBooks, Sudbury.

Jones, D., Endsley, M., 2000. Can real-time probes provide a valid measure ofsituation awareness. In: Proceedings of the Human Performance, SituationAwareness and Automation: User-Centered Design for the New Millennium, SATechnologies, Savannah, GA.

Kanki, B., Helmreich, R.L., Anca, J., 2010. Crew Resource Management, second ed.Academic Press, San Diego.

Kass, S., Cole, K., Stanny, C., 2007. Effects of distraction and experience on situationawareness and simulated driving. Transportation Research Part F: TrafcPsychology and Behaviour 10 (4), 321329.

Larson, G., Alderton, D., Neideffer, M., Underhill, E., 1997. Further evidence on thedimensionality and correlates of the cognitive failures questionnaire. BritishJournal of Psychology 88, 2938.

Martin, M., 1983. Cognitive failure: everyday and laboratory performance. Bulletinof Psychonomic Society 21, 97100.

Mearns, K., Hope, L., 2005. Health and Well-Being in the Offshore Environment: The

models for individuals and teams. Theoretical Issues in Ergonomics Science 9(4), 297323.

Sexton, J.B., Thomes, E.J., Helmreich, R.L., 2000. Error, stress and teamwork inmedicine and aviation: cross-sectional surveys. British Medical Journal 320,745749.

Skogdalen, J., Utne, I., Vinnem, J., 2011. Development of safety indicators forpreventing offshore oil and gas deepwater drilling blowouts. Safety Science 49,11871199.

Smith, A., 2001. Offshore fatigue: a study of ships in the offshore oil industry. In:Proceedings of the SIRCs Second Symposium, Cardiff University, 29th June.

Sneddon, A., Mearns, K., Flin, R., 2006. Safety and situation awareness in offshorecrews. Cognition, Technology and Work 8, 255267.

Sobel, M.E., 1982. Asymptotic condence intervals for indirect effects in structuralequation models. In: Leinhardt, S. (Ed.), Sociological Methodology 1982.American Sociological Association, Washington, DC, pp. 290312.

Stanton, N., Chambers, P., Piggott, J., 2001. Situational awareness and safety. SafetyScience 39, 189204.

Sutherland, K.M., 1994. Psychosocial Factors: An Investigation of the Offshore OilIndustry. Unpublished PhD Thesis, The Robert Gordon University.

Sutherland, V.J., Cooper, C.L., 1986. Man and Accidents Offshore: The Costs of Stress

88 A. Sneddon et al. / Safety Science 56 (2013) 8088Management of Personal Health. RR 305. HSE Books, Sudbury.Mearns, K., Flin, R., Fleming, M., Gordon, R., 1997. Human and Organisational

Factors in Offshore Safety (OTH 543). HSE Books, Sudbury.Mearns, K., Whitaker, S., Flin, R., 2003. Safety climate, safety management

practice and safety performance in offshore environments. Safety Science 41,641680.

Mearns, K., Hope, L., Reader, T., 2006. Health and Well-Being in the OffshoreEnvironment, RR376. HSE Books, Sudbury.

Mearns, K., Hope, L., Ford, M., Tetrick, L., 2010. Investment in workforce health:exploring the implications for workforce safety climate and commitment.Accident Analysis and Prevention 42, 14451454.

Nunnaly, J., 1978. Psychometric Theory. McGraw-Hill, New York.ODea, A., Flin, R., 2001. Site managers and safety leadership in the oil and gas

industry. Safety Science 37 (1), 3957.Ognianova, V., Dalbokova, D.L., Stanchev, V., 1998. Stress states, alertness and

individual differences under 12-hour shiftwork. International Journal ofIndustrial Ergonomics 21, 283291.

Parker, A., Balanda, K., Briggs, L., Hubinger, L., 1998. A Survey of the Work and SleepPatterns of Great Barrier Reef Pilots. Australian Maritime Safety Authority. ISBN:0-642-47618-7.

Parkes, K., 1998. Psychosocial aspects of stress, health and safety on North Seainstallations. Scandinavian Journal of Work, Environment and Health 24, 321333.

Preacher, K.J., Hayes, A.F., 2004. SPSS and SAS procedures for estimating indirecteffects in simple mediation models. Behavior Research Methods 36 (4), 717731.

Reason, J., 1988. Stress and cognitive failure. In: Fisher, S., Reason, J. (Eds.),Handbook of Life Stress, Cognition and Health. Wiley, Chichester, pp. 405421.

Reason, J., Lucas, D., 1984. Absent-mindedness in shops: its correlates andconsequences. British Journal of Clinical Psychology 23, 121131.

Report to the President, 2011. National Commission on the Deepwater Horizon OilSpill and Offshore Drilling.

Rosekind, M., Gander, P., Miller, D., Gregory, K., Smith, R., Weldon, K., Co, E., McNally,K., Lebacqe, J., 1994. Fatigue in operational settings: examples from the aviationenvironment. Human Factors 36, 327338.

Salmon, P.M., Stanton, N.A., Walker, G.H., Green, D., 2006. Situation awarenessmeasurement: a review of applicability for C4i environments. AppliedErgonomics 37 (2), 225238.

Salmon, P.M., Stanton, N.A., Walker, G.H., Baber, C., Jenkins, D.P., McMaster, R.,Young, M.S., 2008. What really is going on? Review of situation awarenessamong Workers on Oil and Gas Rigs. Lloyds List/Dietsmann, London.Sutherland, V.J., Cooper, C.L., 1996. Stress in the offshore oil and gas exploration and

production industries: an organizational approach to stress control. StressMedicine 12, 2734.

Sutherland, K.M., Flin, R.H., 1989. Stress at sea: a review of working conditions inthe offshore oil and shing industries. Work and Stress 3 (3), 269285.

Taylor, R.M., 1990. Situation Awareness Rating Technique (SART): The Developmentof a Tool for Aircrew Systems Design (AGARD-CP-478). Advisory Group forAerospace Research and Development, Neuilly-sur-Seine, France.

Tucker, P., Brown, M., Dahlgren, A., Davies, G., Ebdne, P., Folkard, S., Hutchings, H.,Akerstedt, T., 2010. The impact of junior doctors worktime arrangements ontheir fatigue and well-being. Scandinavian Journal of Work Environment andHealth 36 (6), 485651.

Tversky, A., Kahneman, D., 1974. Judgment under uncertainty: heuristics and biases.Science 185, 11241131.

VANTAGE Card. .

Vom Hofe, A., Mainemarre, G., Vannier, L., 1998. Sensitivity to everyday failures andcognitive inhibition: are they related? European Review of Applied Psychology48, 4955.

Wadsworth, E.J.K., Simpson, S.A., Moss, S.C., Smith, A.P., 2003. The Bristol stress andhealth study: accidents, minor injuries and cognitive failures at work.Occupational Medicine 53, 392397.

Wadsworth, E.J.K., Allen, P.H., McNamara, R.L., Smith, A.P., 2008. Fatigue and healthin a seafaring population. Occupational Medicine 58, 198204.

Wallace, J.C., 2004. Conrmatory factor analysis of the cognitive failuresquestionnaire: evidence for dimensionality and construct validity. Personalityand Individual Differences 37 (2), 307324.

Wallace, J.C., Chen, G., 2005. Development and validation of a work-specicmeasure of cognitive failure: implications for occupational safety. Journal ofOccupational and Organizational Psychology 78, 615632.

Wallace, J.C., Vodanovich, S.J., 2003a. Workplace safety performance:conscientiousness, cognitive failure, and their interaction. Journal ofOccupational Health Psychology 8 (4), 316327.

Wallace, J.C., Vodanovich, S.J., 2003b. Can accidents and industrial mishaps bepredicted? Further investigation into the relationship between cognitive failureand reports of accidents. Journal of Business and Psychology 17 (4), 503514.

Wallace, J.C., Vodanovich, S.J., Restino, B.M., 2003. Predicting cognitive failures fromboredom proneness and daytime sleepiness scores: an investigation withinmilitary and undergraduate samples. Personality and Individual Differences 34,635644.

Stress, fatigue, situation awareness and safety in offshore drilling crews1 Introduction2 Measuring SA3 Factors affecting SA and attentiveness3.1 Stress and SA3.2 Fatigue, sleep disruption and SA

4 The working environment in the offshore drilling industry5 Relationships between performance shaping factors, work SA and accidents5.1 Hypotheses

6 Method6.1 Sample6.2 Procedure6.3 Measures6.3.1 Work SA (WSA)6.3.2 Cognitive failures6.3.3 Fatigue6.3.4 Sleep disruption6.3.5 Stress6.3.6 Unsafe behaviour6.3.7 History of accidents and near-misses

6.4 Analyses

7 Results7.1 Principal components analysis7.2 Hypotheses results

8 Discussion8.1 Stress, WSA and safety8.2 The WSA scale

9 Limitations of the study and directions for future research10 ConclusionAcknowledgementsAppendix A The work situation awareness (WSA) scaleReferences