Cognitive Work Analysis

Cognitive Work Analysis (CWA) is a framework developed by Jens Rasmussen for analyzing the constraints that shape human work in complex systems. Rather than starting with the task as given, CWA begins with the system's goals, the physical and organizational constraints, and the strategies that competent operators develop. It treats expertise not as a fixed property of individuals but as an emergent property of the interaction between the operator, the tools, and the work domain.

The framework's central tool is the abstraction hierarchy — a multi-level representation that maps from physical form (what the system is made of) through functional purpose (what the system does) to normative constraints (why it must do it). This hierarchy makes visible the deep structure of the work domain and the degrees of freedom available to operators. It is not a task analysis but a constraint analysis: it asks not what operators are supposed to do, but what they could do, and what constraints limit their options.

The Abstraction Hierarchy

Rasmussen's abstraction hierarchy organizes a work domain into five levels, each answering a different question:

Functional purpose. Why does the system exist? What are its goals in the broader sociotechnical context? For a nuclear power plant, this is the production of electricity and the protection of public safety.

Abstract function. What are the system's functional constraints and priorities? This includes mass and energy balances, safety margins, and economic constraints.

Generalized function. What are the core processes that realize the abstract functions? Heat transfer, power conversion, control loops, information processing.

Physical function. What are the physical components and their causal connections? Pumps, valves, sensors, controllers.

Physical form. What are the physical materials, geometries, and concrete properties? The steel of the reactor vessel, the concrete of the containment structure.

The hierarchy is a navigation map for cognitive activity. Operators move between levels depending on the problem they face. A routine fault is handled at the physical function level; a novel emergency requires reasoning at the abstract function or functional purpose level. Experts navigate this hierarchy fluidly; novices get stuck at lower levels because they lack the mental model that connects the physical to the purposeful.

The Three Phases of CWA

CWA proceeds through three analytical phases:

Work domain analysis. Identifies the goals, constraints, and resources of the domain using the abstraction hierarchy. This phase answers the question: what is the structure of the work environment, independent of any particular task or operator?

Control task analysis. Identifies what needs to be done to satisfy the domain's goals and constraints. This phase answers the question: what tasks must be performed, what decisions must be made, and what information is required?

Strategies analysis. Identifies how the tasks can be performed. This phase answers the question: what strategies do operators actually use, and what strategies could they use? It recognizes that there are typically multiple valid ways to achieve a goal, and that expert performance involves selecting among these strategies based on context.

CWA and Design

The design implication of CWA is that systems should be designed to support the full range of strategies that competent operators might employ. This contrasts with traditional task analysis, which identifies the correct way to perform a task and designs the system to enforce that way. CWA recognizes that the correct way cannot be fully specified in advance because the work domain contains too much uncertainty, and that operators must have the flexibility to adapt.

This insight connects CWA directly to resilience engineering: both approaches treat human operators as adaptive resources rather than error sources, and both emphasize the importance of designing systems that maintain performance under uncertainty rather than optimizing for nominal conditions. The difference is that CWA provides analytical tools for understanding the domain, while resilience engineering provides organizational principles for maintaining adaptive capacity.

Critique and Extensions

CWA has been criticized for being labor-intensive and requiring deep domain expertise to apply correctly. The abstraction hierarchy is not a template that can be filled in mechanically; it requires judgment about what counts as a functional purpose, what counts as a physical form, and how the levels connect. This has limited the framework's adoption in industry, where faster and cheaper methods are preferred.

Extensions of CWA include Cognitive Work Analysis for Teams, which extends the framework to multi-operator environments, and Ecological Interface Design, which translates the abstraction hierarchy into interface requirements. These extensions recognize that individual cognition is embedded in social and technological contexts, and that the analysis must include the team's distributed cognition and the tools' representational affordances.

Cognitive Work Analysis is not a method for controlling operators. It is a method for understanding the space of possible actions that a competent operator might take, so that the system can be designed to support that space rather than constrain it artificially. The goal is not to specify behavior but to make the right behavior obvious.