Human error, or problematic system design creating an error trap that is, to quote this paper, “loaded and ready to strike”? A brief paper I summarised (post in the coming weeks) re-analysed the investigation report from a “labor auditor”. The incident involved the amputation of an operator’s arm, during operation of a brake-clutch type mechanical… Continue reading Human error or an error trap ready to strike? The case of a mechanical press arm amputation

This study explored bias in incident investigations, using simulated investigation interviews with 34 experienced investigators in the construction industry. I believe this is the dataset where the “New Blame” concept and paper was also derived from (link in comments to that summary). Providing background: ·        Humans use mental shortcuts to make predictions or solve problems quickly—called… Continue reading Exploring bias in incident investigations: An empirical examination using construction case studies

I found this excerpt interesting as I read about the normalisation of deviance (NoD) in mega projects. NoD comes from Diane Vaughan’s analysis of the Challenger disaster, representing a series of “missteps, flawed assumptions, and a culture of risk taking”. Further,  “Social normalization of deviance means that people within the organization become so much accustomed… Continue reading Normalisation of deviance in project management

This evaluated the associations between a contractor safety prequalification assessment program on leading/lagging indicators. The assessment included Safety Management System [SMS] scores, Safety Programs (eg falls, hearing protection], Special elements (drug testing, return to work], history of OHSA citations, and various injury metrics. Safety assessments from 2198 construction contractors were included. Results Increased SMS scores… Continue reading Testing the associations between leading and lagging indicators in a contractor safety pre-qualification database

This paper analysed the use of the Take 5 (Last-Minute Risk Assessment, STOP, SLAM etc.) based on interview and field observations at a major infrastructure construction project. Data collection occurred during alternating periods of enforced Take 5 usage, optional Take 5 usage, and banned Take 5 usage. There’s so much in this paper that I… Continue reading Should We Cut the Cards? Assessing the Influence of “Take 5” Pre-Task Risk Assessments on Safety

Given my recent bias posts, I wanted to balance this with more posts on the effective use of heuristics; most of the time, in most situations, our heuristics (and biases, since bias isn’t a dirty word) help us navigate the world. Gigerenzer provides several examples of heuristics (image 2), their conditions, and how they perform… Continue reading Heuristics that make up the cognitive ‘adaptive toolbox’

This study explored safety performance and network resilience among construction teams. Data came from three mega-projects in the Middle East using Social Network Analysis. Resilience (** capacity for resilience) was observed via actual safety performance data and overall network resilience, simulation through agent-based modelling. Providing background: Results Overall: The findings suggest that “Better resilience is… Continue reading Exploring associations between resilience and construction safety performance in safety networks

Does investigator experience help mitigate bias in accident investigations? A study to be posted in the coming weeks suggests not – finding that more experienced professional investigators were more biased than laypeople when it came to implicating worker action as causal in the accident. That is, experienced accident investigators were more likely to blame workers… Continue reading Professional accident investigators attribute more blame to workers than laypeople, exhibiting a “human error bias”

Take 5, STOP, SLAM and other Last Minute Risk assessments seem like a good idea in principle, but how do they work function in practice? Not so well, according to one study. I finally summarised Havinga, Shire & Rae’s 2022 study which empirically explored the functioning of Take 5 during a trial at a major… Continue reading Do Take 5, Last Minute Risk Assessments etc. work? One study suggests not so well.

This explored the perspectives of 44 healthcare workers regarding the effectiveness of RCA-based interventions/controls, using a Resilience Engineering framing to understand the realities of their everyday work. The study argues that a shortcoming of traditional event-analysis investigation is that they focus on incidents or adverse events – themselves only representing a small subset of hazardous… Continue reading A Bottom-Up Approach to Understanding the Efficacy of Event-Analysis in Healthcare: Paradigm Shift from Safety to Resilience Engineering