Human factors, as perceived by the maintenance workforce, were used as the measure for comparing work areas within a petroleum company. These factors were then compared to an objective measure of reliability (Mean Time Between Failures) in order to determine which factors would be most predictive of plant reliability and process safety. Maintenance personnel were surveyed using scales based on Problem-solving, Vigilance, Design and maintenance, Job-related feedback and Information about change. Analysis of Variance was used to assess the strength of these variables in relation to Reliability Level. Significant differences were observed between different reliability levels based on workforce perceptions of problem-solving requirements and the design and maintainability of plant. Conclusions were that perceptions of human factors in the workplace can be predictive of group-level performance, and that if issues relating to design and maintainability are not addressed at the design stage, greater problem-solving abilities will be required from maintenance personnel. Practitioner summary: Workforce perceptions of plant performance could provide a statistically valid measure of current and future reliability. A survey of perceptions of human factors was conducted with maintenance personnel in a petroleum company. Results indicated significant relationships between reliability and requirements for Problem-solving, as well as Design and Maintenance of equipment. Abbreviations: HFIT: human factors investigation tool, FPSO: floating production, storage and offtake, MTBF: mean time between failures, CPS: cognitive problem- solving, WDS: work design questionnaire, SPSS: statistical package for the social sciences, PAF: principal axis factoring, ANOVA: analysis of variance, ANCOVA: analysis of co-variance, M: mean, SD: standard deviation.