Skill Acquisition Models

5-Stage Model (Dreyfus & Dreyfus, 1980)

Which Dreyfus’ Model is correct? You may have come across an alternative Dreyfus’ Model that lists Novice, Advanced Beginner, Competent, Proficient, and Expert. That model (summarized by Michael Eraut in 1994) is inconsistent with Dreyfus’ original paper (see explanation).
Chess Master

Chess Master

In education, training, and operation research, the prevalent explanation for how learners acquire skills through formal instruction and practice is offered by Dreyfus & Dreyfus, (1980; link). This model was originally created for the military research and described four binary qualities of mental activities in skill acquisition:

  • Recollection (non-situational or situational),
  • Recognition (decomposed or holistic),
  • Decision (analytical or intuitive), and
  • Awareness (monitoring or absorbed).

Dreyfus and Dreyfus distributed the four qualities over five consecutive stages of expertise: Novice, Competent, Proficient, Expert, and Master (see Figure below).

At the stage of Novice, individuals often (blindly) follow rules as given, without context, with no sense of responsibility beyond following the rules exactly. At the stage of Competence, individuals learn to organize principles to quickly access the particular rules that are relevant to the specific task at hand. This stage is characterized by active decision-making in choosing a certain course of action. The stage of Proficiency is reached when individuals begin to develop intuition to guide their decisions and devise their own rules to formulate plans.

The next two stages of Expert and Master must come with practice over a much longer period of time when the procedure to be learned is ingrained into long-term memory through pattern-recognition. The demonstration of skills will eventually become implicit, automated, and seemingly effortless – the sign of a true master. The overall progression is, thus, from rigid adherence to rules (novice) to an increasingly intuitive mode of reasoning based on tacit knowledge.

Model of Skill Acquisition (Dreyfus & Dreyfus, 1980)

Figure 1: Model of Skill Acquisition (Dreyfus & Dreyfus, 1980)

3-Phase Model (Fitts & Posner, 1967)

Fitts & Posner’s model (1967) proposed that (physical) learning can be divided into 3 phases:

  1. Cognitive phase: In this phase, learners need to break down the desired skill into smaller different parts and understand how these parts come together as a whole for the correct performance of the task. The formation of schemas (the way an individual organizes the different parts into a whole) is important in directing the acquisition process.
  2. Associative phase:  In this phase, learners repeatedly practice the tasks until patterns of responding emerge. This is the stage where learners figure out what works and what doesn’t. Effective actions are kept and ineffective actions are dropped. The overall actions of the skills are learned or automated as a learner’s sensory system acquires the accurate spatial and symbolic data required for the completion of the skills. The amount of important stimuli associated with a task will determine how long it takes to complete this phase.
  3. Autonomous phase (or Procedural phase): this phase involves perfecting the skill acquisition. The ability to discriminate important from unimportant stimuli is made quicker and less cognitive process is required because the skill has become automated. Important to this phase of the model is experience and factual knowledge stored for the observed skill.

4-stage “Predictive Cycle” Model (Tadlock & Stone, 2005)

The model proposed by Tadlock and Stone (2005) differs significantly from Fitts and Posner’s in that it does not require conscious understanding of a skill’s components. Rather, the learner is only required to be consciously aware of a concept of the desired end result. The “Predictive Cycle” model includes the following stages:

  1. Attempt,
  2. Fail,
  3. Implicitly analyze the result, and
  4. Implicitly decide how to change the next attempt so that success is achieved.

The stages are repeated over and over until the learner builds or remodels the neural network to guide an activity appropriately and accurately without conscious thought. As individuals repeat this cycle, they will gradually increase their performance by weeding out unnecessary steps, keeping only the effective ones. This model has been successfully applied to reading remediation (Scott, Nelsestuen, Autio, Deussen, & Hanita, 2010).

3-stage Shu-Ha-Ri 【守破离】 Model

The Masters of many Japanese art forms (e.g., tea ceremony, martial art, opera) adhere to a 3-stage skill acquisition model called しゅはり守破离】 (or, Shu-Ha-Ri). These stages have a close connection with the procedural instructions associated with 【道】, which can be transliterated as the Path (as in the ‘right path’), or the Way (as in the ‘right way’ of doing things).

Martial arts, for example, are conducted in ‘schools,’ (which are more appropriately, training grounds) called dōjō 【道场】. Students of the martial arts would train under the tutelage of a Sensei (or Master Teacher) for a period of time, eventually graduate (leave) to open their own dōjō. In this context, the three stages of Shu-Ha-Ri can be perceived as the learning experience of a student whose goal is to master ‘the Way.’

  1. Shu (Keeping with the Path) 【守道】 – Students must follow strictly the instructions of the Sensei.
  2. Ha (Breaking from the Path) 【破道】 – Students who become proficient in the instructions are now allowed to explore new Ways to improve self/skill.
  3. Ri (Leaving the Path) 【离道】 – Students who achieve Expert or Master levels are ready to leave (and branch out) to start their own dōjō.

In comparison to the “Predictive Cycle” model, Shu incorporates the first two steps (1 & 2) of Attempt and Fail, while Ha incorporates the last two steps (3 & 4) of Analyzing the result and Deciding how to change for improvement. In comparison to the Fitts and Posner’s model, Shu would be the same as the Cognitive Phase, while Ha is equivalent to the Associative Phase.


Combining the Models: Competency-based Training

From a training and improvement perspective, trainers are seldom required to take an individual from Novice all the way to the Expert or Master level. It would be expecting too much as the time available for training (courses) is often limited to a few months or weeks (if not days). Hence, it would be more realistic to expect to teach the Novices the basic Ways (or ‘moves’) and the thinking processes involved in how to go about learning new things (i.e., meta-cognition skills). Ultimately, the purpose of such training is to learn the basic skills, enough to perform a job. Accuracy and proficiency should naturally comes from real- world (workplace) practices.

In this sense, a short-term training can be deemed successful when an individual (1a) is capable of performing the required task (1b) within a reasonable time frame, (2) being able to judge whether the task performed is up-to-par, (3) having some idea about how to improve the accuracy and/or proficiency, so as to prevent (financial) losses, or endangering themselves and others. For instance, a medic is expected to perform a suture correctly and within a certain time frame if she is to be deployed, and a qualified pilot should be able to take off and land a plane safely in any weather condition within a certain time frame. Virtual environments and serious games are excellent options for this purpose because the learners will not accidentally harm themselves or others in the process.

Since the training of Experts and Masters requires a very long time, short-term training only addresses the skill acquisition from Novice to Proficient. Combining the various models, we have:

  1. Novice stage — learners should become familiarized with the basic instructions, practice them (mentally and physically, where appropriate), and keep trying (despite failures) until they have a good command of the basic moves. As newbies, they are expected to make mistakes and be unsure of the correct ways to do things. The training goal for Novices in this stage is for them to connect basic moves into a rudiment “course of action” – recognizable as a complete course with appropriate “start” and “end” points. This stage is consistent with Stage 1 of Fitts and Posner, Stage 1 and 2 of Tadlock and Stone, and Stage 1 of ShuHaRi. [keywords: expect to fail, repetition, correction, basic instruction, practice, connecting tasks to procedure]
  2. Competent stage — learners begin to explore new ways of performing the procedures, with determination. The goal is to practice deliberately with a purpose to improve the execution of procedures with accuracy and proficiency. Learners should be capable of self-analysis to mentally determine if the changes are signs of improvement or not. Learners should be exposed to varying difficulties, and changing conditions of the procedures, to expose them to the most commonly encountered scenarios at workplace. Actions that lead to improvement should be retained whereas those that are not, discarded or modified. This particular stage is consistent throughout the various models and easily recognizable as: (a) Stage 2 of Fitts & Posner, (b) Stage 3 & 4 of Tadlock and Stone, and (c) Stage 2 of ShuHaRi. [keywords: persistence, motivation, meta-cognition, self-analysis, improvement, formative evaluation]
  3. Proficient stage — learners achieved basic proficiency in the skills and know some improvement tactics. To determine if the learners are competent enough to join the workforce – equivalent to competency-based training (see next section), some kind of test or assessment maybe needed to evaluate if baselines of accuracy and proficiency have been attained. Are they sufficiently better than an untrained person? Is the training program effective? Have they reached the bare minimum to join the workforce and (hopefully) be productive? Graduates may further improve themselves by seeking new lessons, new teachers, or take part in a tournament to determine their ranking in the community. However, this remain largely a personal choice (motivation). [keywords: assessment, competency-based training, likely-expert, summative evaluation]

Determined learners can continue the learning journey to eventually become Experts and Masters capable of teaching others, but that remains a largely personal decision.

Interestingly, medical education has started changing into a competency-based model. This is a marked departure from the time-based Flexner-style approach (i.e., 6 years + 1 year residency) from before. Perhaps the medical schools finally realize the need to ensure medical students who are entering the profession are indeed competent in their job? The Proficient Stage mentioned above is exactly what is needed in competency-based training.

As expected, the push towards competency-based education will require new assessments and new approaches to analyzing assessment results. When we apply competency-based training to the virtual environment, the same is true: We will need new assessment metrics and methods to analyze the obtained data for assessment.



  • Dreyfus & Dreyfus (1980). A five-stage model of the mental activities involved in directed skill acquisition. [PDF]