[DSRP Evidence](https://dsrpevidence.org/)

# Risk management in a dynamic society: a modelling problem

## Details

**Authors** Rasmussen, Jens

**Year** 1997

**Publisher** Safety Science

**Kind of work** article

**Discipline** Engineering

**Applied** false

[Read it at the publisher](https://doi.org/10.1016/s0925-7535(97)00052-0) 
10.1016/s0925-7535(97)00052-0

## In authors' words

### What they set out to do (purpose)

To ask whether existing models of accident causation fit a fast-changing, competitive society, and to propose a way of modeling risk management.

### Who or what was studied (sample)

A conceptual and review paper drawing on the author's research on industrial risk management; no single empirical sample.

### How they did it (methods)

A conceptual analysis of the socio-technical system involved in risk management, from legislators through managers and work planners to system operators, treated as a control structure. It includes a review of paradigms in decision theory, management research and safety research.

### What they found (results)

The author argues that risk management must be modeled through cross-disciplinary studies that treat it as a control problem, representing the control structure that spans all levels of society for each hazard category. He argues that this requires a system-oriented approach based on functional abstraction rather than structural decomposition. He also proposes replacing task analysis focused on action sequences and human errors with a model of the mechanisms that shape behavior, in terms of work-system constraints and boundaries of acceptable performance. He reports a convergence of research paradigms in the human sciences, guided by cognitive science concepts, that supports this approach.

## Commentary

### In short

An argument that industrial safety depends on control running through a chain from legislators to managers, planners and operators, and that accidents should be modeled across that whole chain rather than as individual operator error.

### Response

The paper describes a layered socio-technical system (systems) in which each level constrains and is constrained by the others through control relationships (relationships). The fit with systems is partial in one respect. The author explicitly prefers functional abstraction to structural decomposition, so the levels are a control hierarchy rather than a simple nesting of parts in a whole. Perspectives is arguable and is not claimed: the paper notes that each level is studied by a different discipline, but its argument is about integrating those studies rather than about viewpoint as such. The work is a modeling argument supported by review, not a test.

**Patterns it shows** S, R

**Added** 2026-11-17

**How to cite this** Rasmussen, Jens (1997). Risk management in a dynamic society: a modelling problem. Safety Science.
