Ingenieur at the screen

How To Evaluate Expertise: the Dreyfus Model

Time to Read 4 min

Since some time we use the Dreyfus model for the assessment of the professional development and the skills of our engineers. I am repeatedly asked about this model. Here the story on how we got to the model and about the version we use:

How to Asses the Abilities of a Specialist?

This is the actual question I asked myself when the need for an assessment of our staff cropped up. Not to generate a ranking at the end of the year, but above all to staff our projects with the right mix of skills. The engineer should not be challenged too much and whenever possible also not be challenged too little.

My wife works in nursing and she taught the Benner model. It is a model describing the professional development of nursing specialists, Funnily enough, this model is based on a model for technical specialists, which was created by the brothers Dreyfus in 1980 for the US Air Force and which they later refined. It defines five competency levels which every specialist runs through in her/ his personal development, not everybody to expert:
Novice - Beginner- Competent- Proficient- Expert.

This was exactly what I needed. The scientific papers however were a little bit too abstract for practical use... Until after some searching on the web I found a version of the model by S. Lester.

We adapted it for our use, now the model looks like this:

Level Knowledge Work Output and Quality Autonomy Coping with Complexity Networked Thinking
Intuitive (before rationale) decision making
Authoritative knowledge of technical field and deep tacit understanding across area of practice Excellent results for open tasks are achieved with relative ease in the time estimated Able to take responsibility for going beyond existing standards and creating own interpretations Holistic grasp of complex situations, moves between intuitive and analytical approaches with ease, can structure open problems Sees overall picture and alternative approaches, has a vision of what may be possible
Holistical (before analytical) assessment of context
Deep understanding of technical field and area of practice Immaculate standard is achieved for open tasks routinely and in the time estimated Able to take full responsibility for own work (and that of others if applicable) Deals with complex situations holistically, certain decision-making Sees overall picture and how individual actions fit within it
Recognition (before no) of relevance
Good working and background knowledge of practice Useful results are reached also for open tasks in the estimated time, though may lack refinement Able to achieve most tasks using own judgement Copes with complex situations through deliberate analysis and planning Identifies actions at least partly in terms of longer-term interrelations
Knowledge in (before without reference to) context
Working knowledge of key aspects of practice Straightforward tasks likely to be completed to an acceptable standard Able to achieve some steps using own judgement, but supervision needed for overall task Appreciates complex situations but only able to achieve partial resolution Sees actions as a series of steps


Minimal or textbook knowledge without connecting it to practice Unlikely that satisfactory performance is attained unless closely supervised Needs close supervision or instruction Little or no conception of dealing with complexity Tends to see actions in isolation

The beauty of the table lies in its description of observable behavior. In addition the model has a clear focus on the skills in practice, where e.g. textbook knowledge is necessary but far from sufficient for good performance.

The Expert Can Not Explain It

Who has not yet had the experience that an expert can immediately name the right solution for a problem, but cannot explain to a beginner why this is the right answer. The model also responds to this: the expert finds her/ his solution intuitively, based on the accumulated experience. Whoever stands on a different level needs a rational justification. The expert might be able to deliver this justification only after prolonged thinking.

Use of the Dreyfus Model at Solcept

For all engineers we successfully use the above table for self assessment and assessment by others of their professional roles. The consensus of those assessments is then used to plan training and the assignment of jobs, i.e. the kind of tasks for the employee.

Personally I would see optimization potential with the last two columns of the table. Those could be merged because they are partially redundant.

Do You know Other Models? How do You Use Those?

All models are wrong, but some are useful... What do you think about this?

Andreas Stucki

Do you have additional questions? Do you have a different opinion? If so, email me or comment your thoughts below!



No Comments

What is Your Opinion?

* These fields are required

Let us discuss your idea/ your project