Offshore Safety Questionnaire (OSQ)

Abstract

The Offshore Safety Questionnaire (OSQ) is an established psychometric assessment developed by Kathryn Mearns, Rhona Flin, and colleagues at the University of Aberdeen to evaluate organizational safety climate and self-reported compliance among personnel working on offshore oil and gas installations. Originating from research commissioned following the 1988 Piper Alpha disaster and the subsequent Cullen Report (1990), the OSQ provides an empirical mechanism for benchmarking human factors and organizational safety performance in high-reliability and high-hazard environments. The instrument comprises 30 items divided into two interrelated core sections: a 19-item safety climate attitude section and an 11-item self-reported unsafe behaviour inventory. The safety climate attitude section measures two primary latent factors: Perceived Management Commitment to Safety and Perceived Supervisor Competence and Commitment, evaluated on a 5-point Likert agreement scale. The unsafe behaviour section captures two behavioural dimensions: Unsafe Behaviour Under Incentives/Pressure and General Unsafe Behaviour, rated on a 5-point frequency continuum ranging from “Never” to “Constantly.” Psychometric evaluations across multiple North Sea installations demonstrate robust internal consistency (Cronbach’s alphas typically between .78 and .89), sound construct validity confirmed via exploratory and confirmatory factor analyses, and significant predictive validity regarding offshore accident, near-miss, and lost-time injury (LTI) rates. The OSQ remains a foundational tool in occupational health psychology, industrial safety research, and safety management benchmarking.

Keywords

Offshore Safety Questionnaire, OSQ, safety climate, safety culture, management commitment to safety, supervisor competence, unsafe behaviour, high-hazard environments, occupational safety, industrial psychometrics, North Sea oil platforms, Piper Alpha inquiry, accident prevention.

Authors

The Offshore Safety Questionnaire was designed and validated by an interdisciplinary team of applied psychologists and safety scientists at the Industrial Psychology Research Centre (IPRC), School of Psychology, University of Aberdeen, Scotland, UK:

• Kathryn Mearns, Ph.D.: Former Reader in Organisational Psychology at the University of Aberdeen and internationally recognized expert on safety culture, risk perception, and organizational benchmarking in hazardous industries.
• Rhona Flin, Ph.D., FBPsS, FRSE: Emeritus Professor of Applied Psychology at the University of Aberdeen and Professor of Industrial Psychology at Aberdeen Business School, Robert Gordon University. Professor Flin is a global authority on human factors, non-technical skills, crew resource management (CRM), and leadership in high-risk operational environments (e.g., offshore energy, aviation, surgery).
• S. Whitaker, M.Sc.: Research Fellow in the Department of Psychology at the University of Aberdeen, contributing to large-scale data collection, industrial benchmarking, and statistical analysis for UK Health and Safety Executive (HSE) research contracts.
• R. Gordon, Ph.D.: Applied occupational psychologist specializing in workforce safety attitudes, human error taxonomy, and root-cause safety investigations.
• Paul O’Connor, Ph.D.: Human factors researcher and applied psychologist, later affiliated with the National University of Ireland (NUI) Galway and the US Naval Postgraduate School, focusing on high-reliability team training and safety assessment.

Purpose

The primary purpose of the Offshore Safety Questionnaire is to measure, diagnose, and benchmark workforce safety attitudes, perceptions of safety climate, and the prevalence of non-compliant or unsafe operational behaviours among personnel deployed to offshore hydrocarbon installations. Operating an offshore oil and gas installation involves complex, safety-critical sociotechnical systems where physical isolation, harsh meteorological conditions, heavy machinery, high-pressure hydrocarbons, and rotational shift patterns converge to create extreme operational risks. Traditional safety metrics, such as Lost Time Injury Frequency (LTIF) or Total Recordable Incident Rate (TRIR), represent “lagging indicators” that register failures after harm has already occurred. The OSQ was developed to provide “leading indicators”—evaluating organizational climate, supervisory dynamics, and latent behavioural patterns capable of predicting hazardous events before they materialize.

From an applied industry perspective, the OSQ serves several distinct functions:

• Cross-Installation Benchmarking: Enabling operating companies and regulatory bodies (such as the UK Health and Safety Executive) to quantitatively compare safety perceptions across diverse installations, operating companies, drilling contractors, and third-party service providers.
• Diagnostic Safety Audits: Pinpointing specific organizational deficiencies, such as disconnects between workforce risk awareness and executive commitment, perceived production pressures, or deficits in frontline supervisory safety leadership.
• Targeted Intervention Design: Facilitating evidence-based human factors interventions, such as supervisory coaching, procedural redesign, and safety reporting system overhauls.
• Longitudinal Evaluation: Tracking shifts in organizational safety climate over time, particularly following major structural changes, regulatory adjustments, corporate mergers, or safety improvement campaigns.

In academic research, the OSQ bridges theoretical models of organizational climate with empirical worker behaviour. It enables researchers to examine the causal pathways linking upstream organizational values, supervisory enforcement, operational pressures, self-reported compliance, and catastrophic systems failures.

Psychological Construct

The Offshore Safety Questionnaire measures safety climate—defined as the shared perceptions among employees concerning the priority of safety relative to production and operational demands within their work environment (Zohar, 1980). Rather than attempting to evaluate safety culture in its entirety—which encompasses deep-seated assumptions, historical values, and cultural artifacts—safety climate assesses the contemporary, measurable “snapshot” of those values in action. The OSQ operationalizes this domain through four major latent dimensions across its two sections:

1. Perceived Management Commitment to Safety (Attitudinal Subscale)

This dimension assesses the workforce’s belief that executive and senior operational leadership genuinely values human life and physical well-being above commercial throughput, operational speed, and corporate profit. When management demonstrates genuine commitment, employees perceive that safety systems are authentic rather than performative compliance exercises. Key operational expressions of this construct include:

• Equitable parity between safety integrity and production throughput under operational stress (e.g., Item 1).
• Willingness to invest capital and resources to swiftly eliminate safety hazards (e.g., Item 2).
• Proactive safety engagement rather than reactive interventions solely triggered by accidents or external audits (e.g., Items 5 and 6).
• Openness to workforce safety feedback and two-way risk communication (e.g., Items 7 and 19).

2. Perceived Supervisor Competence and Commitment (Attitudinal Subscale)

Frontline supervisors represent the operational interface between high-level management policies and shop-floor execution. This dimension captures the direct workforce appraisal of the supervisor’s daily safety leadership, technical mastery, and fairness in procedural enforcement. A supervisor may understand safety rules conceptually, but if they lack technical operational competence or capitulate to scheduling pressures, frontline safety is compromised. Indicators include:

• Consistent enforcement of safety standards when tasks fall behind schedule (e.g., Item 8).
• Technical capability to direct operations safely and recognize unforeseen hazards (e.g., Item 10).
• Active encouragement of psychological safety, enabling subordinates to stop the job or highlight hazards without fear of reprisal (e.g., Items 9 and 16).
• Continuous, everyday safety communication rather than periodic or superficial surveillance (e.g., Items 12 and 14).

3. Unsafe Behaviour Under Incentives/Pressure (Behavioural Subscale)

This subscale quantifies the frequency with which personnel deliberately circumvent established safety barriers, rules, or permits due to external constraints, structural bottlenecks, or perceived personal/team rewards. In offshore environments, production deadlines, drill delays, shift-change departures, and peer expectations can incentivize risk-taking. Specific manifestations include:

• Taking procedural shortcuts to ensure timely job completion (e.g., Item 20).
• Circumventing formal protocols due to equipment shortages or procurement delays (e.g., Item 21).
• Submitting to perceived informal pressures from crew members or line superiors to expedite tasks (e.g., Item 22).
• Bypassing mandatory safety steps to catch crew-change helicopters or transition off-shift (e.g., Item 23).

4. General Unsafe Behaviour (Behavioural Subscale)

This construct captures non-compliance, cognitive failure, or risk acceptance that occurs independently of direct scheduling pressure. It reflects normalized operational drift, lapses in situational awareness, and underestimation of routine operational hazards. Manifestations include:

• Executing tasks without completing mandatory pre-job safety checks or permit-to-work stipulations (e.g., Items 24 and 25).
• Failure to utilize mandatory personal protective equipment (PPE) (e.g., Item 27).
• Failing to document or report “minor” unsafe conditions, reinforcing systemic blind spots (e.g., Item 29).
• Continuing critical technical work while impaired by physical fatigue or cognitive distraction (e.g., Item 30).

Theoretical Framework

The OSQ is rooted in contemporary industrial/organizational psychology, human factors engineering, and socio-technical systems theory. Its theoretical lineage combines four foundational models:

1. Zohar’s Multilevel Safety Climate Theory

Dov Zohar’s foundational work (1980, 2000) conceptualizes safety climate as an emergent property of organizations, formed through employee sensemaking regarding the policies, procedures, and practices (POPs) enforced in their environment. While formal corporate policy may declare “Safety is our Top Priority,” employees deduce true organizational priorities by observing the consequences of operational trade-offs (e.g., what happens when safety slows down production). The OSQ explicitly operationalizes this duality by separating senior management commitment (declarative priorities and systemic governance) from supervisory commitment and competence (direct behavioural reinforcement and execution).

2. James and Jones’ Climate Differentiation

The OSQ respects the psychometric distinction delineated by James and Jones (1974) between psychological climate (an individual-level cognitive appraisal of the work environment) and organizational climate (shared, aggregated perceptions across a collective unit). By surveying individuals on standard Likert scales and calculating within-group consensus metrics ($r_{wg}$, Intraclass Correlation Coefficients), researchers use the OSQ to aggregate individual scores to the installation or departmental level, capturing an empirical macro-measure of safety climate.

3. Reason’s Swiss Cheese and Latent Human Error Model

James Reason’s (1990, 1997) model of system safety emphasizes that major industrial disasters rarely stem solely from isolated active failures (operator slips, lapses, or rule violations). Instead, catastrophic breakdowns emerge when active failures penetrate multiple defective organizational defenses (latent conditions). In the OSQ framework, poor management commitment, production-safety conflict, and ineffective supervision represent systemic latent conditions. These latent conditions cultivate a climate in which active non-compliance (the 11 unsafe behaviour items) becomes normalized, increasing the likelihood that latent vulnerabilities align into an accident trajectory.

4. Social Exchange Theory and the Psychological Contract

Under Blau’s (1964) Social Exchange Theory, workplace behaviours reflect reciprocal social obligations. When offshore personnel perceive that senior leaders and supervisors invest in robust protections, listen to workforce suggestions, and maintain safe working environments without prioritizing financial throughput, employees reciprocate with high safety citizenship, voluntary compliance, and vigilance. Conversely, when workers perceive management as hypocritical or indifferent, the psychological contract deteriorates, normalizing rule violations and procedural drift.

Validity

The psychometric validity of the Offshore Safety Questionnaire has been substantiated through empirical investigations across hundreds of installations operating in the UK Continental Shelf (UKCS), the Norwegian Continental Shelf, and international maritime sectors.

Construct Validity

Construct validity was established through exploratory and confirmatory factor analyses. In the primary validation studies led by Mearns et al. (2001, 2003), principal axis factoring confirmed distinct construct validity between workforce perceptions of organizational environment (attitudes) and self-reported operational practices (behaviours). In the 19-item climate section, items consistently converge onto two distinct latent dimensions—Management Commitment and Supervisor Competence—explaining between 45% and 58% of the total variance across multi-year data waves. The 11-item behavioural section cleanly bifurcates into pressure-induced violations versus general non-compliance, demonstrating that worker non-compliance is multidimensional.

Convergent and Discriminant Validity

Convergent validity is supported by strong correlations between the OSQ dimensions and related organizational constructs, including job satisfaction, organizational trust, and procedural justice. Mearns and Flin demonstrated that offshore workers who scored high on Perceived Management Commitment to Safety reported significantly higher levels of trust in management ($r = .52, p < .001$) and perceived procedural fairness ($r = .48, p < .001$). Discriminant validity has been demonstrated through the empirical separation of supervisory technical competence from general interpersonal liking, confirming that respondents evaluate safety-specific leadership independently of broad personal affinity.

Predictive and Criterion Validity

A critical benchmark for any occupational safety metric is its criterion-related predictive validity against objective safety outcomes. The OSQ exhibits strong predictive validity:

• Lost Time Injuries (LTIs) and First-Aid Cases: In a benchmark study of 13 offshore installations over a 12-month prospective period, Mearns et al. (2003) revealed that aggregate installation scores on Perceived Management Commitment and Supervisor Competence were negatively correlated with official installation incident rates ($r = -.41$ and $r = -.47, p < .05$, respectively).
• Reporting of Near Misses: Higher scores on workforce safety voice (Item 16, Item 9) predicted significantly higher volumes of reported near-misses and hazard observations. In high-reliability paradigms, an elevated volume of proactive near-miss reports signifies a healthy reporting culture rather than a dangerous worksite.
• Self-Reported Incident Involvement: Scores on the Unsafe Behaviour Under Incentives/Pressure subscale directly predicted self-reported accident and minor injury involvement ($eta = .34, p < .001$), accounting for substantial variance beyond demographic controls, age, and offshore experience.

Reliability

The OSQ demonstrates solid internal consistency and temporal stability across diverse offshore samples:

Internal Consistency

Across validation cohorts spanning several thousand offshore employees, the OSQ’s subscales have demonstrated robust Cronbach’s alpha ($lpha$) coefficients:

• Perceived Management Commitment to Safety: $\alpha$ coefficients consistently range from .84 to .89.
• Perceived Supervisor Competence and Commitment: $\alpha$ coefficients range from .80 to .86.
• Unsafe Behaviour Under Incentives/Pressure: $\alpha$ coefficients range from .77 to .83.
• General Unsafe Behaviour: $\alpha$ coefficients range from .73 to .79.
• Composite 19-Item Attitude Scale: Overall instrument reliability regularly exceeds .90.

Average inter-item correlations within subscales fall securely within the recommended .30 to .55 range, confirming that while items share a common underlying construct, they avoid redundant collinearity.

Test-Retest Stability and Aggregation Metrics

In stability studies over 3-to-6-month intervals among continuous installation crews, test-retest reliability coefficients ranged from $r = .71$ to $r = .78$, indicating that the scale captures stable organizational climate traits while remaining sensitive to tangible operational changes (e.g., leadership turnover or serious safety incidents).

Because safety climate is frequently analyzed as an installation-level property, Mearns et al. examined group-level aggregation metrics:

• Within-Group Agreement ($r_{wg}$): Median values exceeded .82 across installations, well above the .70 threshold required to justify data aggregation.
• Intraclass Correlation Coefficients: ICC(1) values ranged from .12 to .18, and ICC(2) values ranged from .68 to .81, confirming that meaningful variance exists between different offshore platforms and that platform-level mean scores provide reliable assessments of group climate.

Factor Analysis

The latent structural integrity of the OSQ was rigorously validated through consecutive phases of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA):

Exploratory Factor Analysis (EFA)

Initial validation studies by Mearns, Whitaker, and Flin (2001, 2003) subjected the 19 attitudinal statements and 11 behavioural statements to separate factor analyses to account for differences in response formats and conceptual focus:

• Safety Climate Attitudes (19 Items): Principal Axis Factoring (PAF) and Principal Components Analysis (PCA) with both Varimax (orthogonal) and Oblimin (oblique) rotations revealed an unequivocal two-factor solution based on Kaiser’s criterion (eigenvalues > 1.0) and scree plot inspections. Factor 1, Perceived Management Commitment to Safety, accounted for the largest share of common variance (34.2%), with high loadings from items addressing resource allocation, production-versus-safety trade-offs, and structural communication. Factor 2, Perceived Supervisor Competence and Commitment, accounted for 11.6% of the variance, driven by items assessing direct supervisory conduct, technical competence, and daily rule enforcement. All items exhibited primary loadings > .45, with cross-loadings generally below .25.
• Unsafe Behaviours (11 Items): Factor analysis of the behavioural inventory yielded two clean factors. Factor 1, Unsafe Behaviour Under Incentives/Pressure (eigenvalue = 3.65, explaining 33.2% of variance), gathered items describing shortcuts driven by time pressure, tool unavailability, shift handovers, and supervisory coercion. Factor 2, General Unsafe Behaviour (eigenvalue = 1.28, explaining 11.6% of variance), clustered items related to fatigue, absent risk appraisals, neglect of PPE, and failure to report minor hazards.

Confirmatory Factor Analysis (CFA)

Subsequent structural equation modeling across validation datasets confirmed the superiority of this hypothesized two-tier, two-factor structure over competing one-factor unidimensional models. Model fit evaluations using standard structural metrics yielded the following parameters:

• Chi-Square to Degrees of Freedom Ratio ($\chi^2/df$): Values consistently fell between 1.85 and 2.40, well within the conventional threshold (< 3.0) for good model fit.
• Comparative Fit Index (CFI): Ranged from .93 to .96, indicating strong fit to empirical data.
• Tucker-Lewis Index (TLI): Ranged from .92 to .95.
• Root Mean Square Error of Approximation (RMSEA): Values ranged from .042 to .056 (with 90% confidence intervals between .038 and .062), demonstrating low residual error.
• Standardized Root Mean Square Residual (SRMR): Observed at .044.

Alternative single-factor models (forcing all 19 climate items onto a solitary general safety factor) displayed significantly deteriorated fit ($\Delta \chi^2(1) > 450, p < .0001; CFI = .78; RMSEA = .104$), establishing that management-level governance and supervisory-level execution represent distinct operational constructs in the minds of offshore personnel.

Instrument / Measurement Tool

The Offshore Safety Questionnaire is formatted for paper-and-pencil or secure digital self-administration. It contains 30 core operational items split into two distinct structural modules:

Structural Specifications

• Instrument Type: Standardized self-report diagnostic organizational survey.
• Total Number of Items: 30 items (19 safety climate attitude items + 11 unsafe behaviour items).
• Target Population: Offshore installation managers, production engineers, drilling crew, maintenance technicians, catering personnel, and subsea contractor staff.
• Administration Time: Approximately 10 to 15 minutes.
• Administration Mode: Anonymous paper booklet distribution during offshore safety meetings or secure digital entry via installation computer terminals. Anonymity is critical to minimize social desirability bias when reporting unsafe behaviours.

Response Scales

• Items 1–19 (Safety Climate Attitudes): Evaluated using a 5-point Likert agreement scale:
  • 1 = Strongly Disagree
  • 2 = Disagree
  • 3 = Neither Agree nor Disagree
  • 4 = Agree
  • 5 = Strongly Agree
• Items 20–30 (Unsafe Behaviour / Non-compliance): Evaluated using a 5-point behavioural frequency scale:
  • 1 = Never
  • 2 = Rarely
  • 3 = Sometimes
  • 4 = Often
  • 5 = Constantly

Scoring Guidelines and Reverse Scoring

• Reverse-Scored Items: To prevent acquiescence response bias, four items in the Safety Climate Attitudes section are negatively worded and must be reversed prior to score computation ($1
  ightarrow 5, 2
  ightarrow 4, 3
  ightarrow 3, 4
  ightarrow 2, 5
  ightarrow 1$):
  • Item 3: “Management turns a blind eye to safety issues when production targets are at risk.”
  • Item 5: “Management only shows interest in safety after an accident occurs.”
  • Item 12: “My supervisor pays attention to safety only when inspectors or senior managers are around.”
  • Item 13: “My supervisor values production more than personal safety.”
• Subscale Score Calculation:
  • Perceived Management Commitment to Safety: Calculated as the mean of Items 1, 2, 3 (reversed), 4, 5 (reversed), 6, 7, 16, 17, 18, and 19. Higher scores reflect stronger perceived organizational safety prioritization.
  • Perceived Supervisor Competence and Commitment: Calculated as the mean of Items 8, 9, 10, 11, 12 (reversed), 13 (reversed), 14, and 15. Higher scores denote effective frontline supervisory safety leadership.
  • Unsafe Behaviour Under Incentives/Pressure: Calculated as the mean of Items 20, 21, 22, and 23. Higher scores indicate frequent violation of safety barriers under operational constraints.
  • General Unsafe Behaviour: Calculated as the mean of Items 24, 25, 26, 27, 28, 29, and 30. Higher scores reflect frequent procedural drift and normalized risk acceptance.

Permissions & Fee and Test Year

The Offshore Safety Questionnaire was finalized and published in its standardized form in 2003 under contract to the UK Health and Safety Executive (HSE) and published in peer-reviewed scientific journals (e.g., Mearns, Whitaker, & Flin, 2003). As part of research funded by the UK Government via the HSE (Research Report Series / HMSO), the core items are generally available for academic research, educational inquiry, and public safety benchmarking under fair use and Crown copyright research provisions.

Commercial deployment, cross-platform industrial benchmarking databases, or tailored proprietary safety climate audits may require consultation with the copyright holders or authors via the University of Aberdeen’s Industrial Psychology Research Centre (IPRC). Researchers wishing to deploy the tool in academic studies should cite the original development studies (Mearns et al., 2003) and adhere to psychometric reporting standards.

References

• Blau, P. M. (1964). Exchange and power in social life. John Wiley & Sons.
• Cullen, W. D. (1990). The public inquiry into the Piper Alpha disaster (Volumes 1 & 2, Cm 1310). Her Majesty’s Stationery Office (HMSO).
• Flin, R., Mearns, K., O’Connor, P., & Bryden, R. (2000). Measuring safety climate: Identifying the common features. Safety Science, 34(1–3), 177–192. https://doi.org/10.1016/S0925-7535(00)00012-6
• James, L. R., & Jones, A. P. (1974). Organizational climate: A review of theory and research. Psychological Bulletin, 81(12), 1096–1112. https://doi.org/10.1037/h0037511
• Mearns, K., Flin, R., Gordon, R., & Fleming, M. (1998). Measuring safety climate on offshore installations (Work Research Report OTH 538). Health and Safety Executive (HSE Books).
• Mearns, K., Whitaker, S., & Flin, R. (2001). Benchmarking safety climate in hazardous environments: A longitudinal study (Offshore Technology Report OTO 2001/047). Health and Safety Executive.
• Mearns, K., Whitaker, S., & Flin, R. (2003). Safety climate, safety management practice and safety performance in offshore environments. Safety Science, 41(8), 641–680. https://doi.org/10.1016/S0925-7535(02)00011-5
• Mearns, K., Whitaker, S., Flin, R., Gordon, R., & O’Connor, P. (2003). Factoring the human into safety translating research into practice. Volume 1: Benchmarking human and organisational factors in offshore safety (Research Report 059). Health and Safety Executive (HSE Books / HMSO). https://www.hse.gov.uk/research/rrhtm/rr059.htm
• Reason, J. (1990). Human error. Cambridge University Press. https://doi.org/10.1017/CBO9781139062366
• Reason, J. (1997). Managing the risks of organizational accidents. Ashgate Publishing.
• Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65(1), 96–102. https://doi.org/10.1037/0021-9010.65.1.96
• Zohar, D. (2000). A group-level model of safety climate: Testing the effect of supervisory leadership on group safety climate and performance. Journal of Applied Psychology, 85(4), 587–596. https://doi.org/10.1037/0021-9010.85.4.587

Items of the Scale