The need for competence: towards a more precise definition

Self-determination theory ( SDT) posits that in addition to the need for autonomy and relatedness, people also have a fundamental need for competence. In attempting to use this need for competence in educational software, researchers have found its definition to be confusing and incomplete. Two recent papers address this by clarifying and formalizing the definition.

The Self-Determination Theory

Self-determination theory (SDT) is currently the most influential motivational theory in psychology. The theory states that people have three basic psychological needs: autonomy, relatedness, and competence. These needs form the basis for optimal motivation and well-being.

The need for competence

The need for competence is usually defined as the need to be effective in our interactions with the environment and to develop our capabilities. However, recent research has revealed a problem: the definition of the need for competence is not as clear-cut as previously thought. This became particularly clear when researchers and developers wanted to apply the theory to educational software and other digital applications. Computers require clear, unambiguous definitions – and these were lacking.

Step 1: Identifying four facets

In their article “Why Self-Determination Theory Needs Formal Modelling”, Deterding et al. analyzed the ZDT literature in a systematic way. They found that what we call 'competence' actually consists of four different facets:

1. Effectance: Perceiving that your actions have an effect on the environment
2. Skill use: Being able to apply your abilities
3. Task performance: Performing well on intended tasks
4. Capacity Growth: Developing new or stronger skills

This discovery was important because these facets do not always go together. For example, you can have an effect on your environment without performing as intended, or you can perform a task well without developing new skills.

Step 2: Computational Models

The follow-up study by Lintunen et al. (2025) built on this using computational models. These are mathematical models that can be converted into computer programs. They describe exactly how certain processes work. For example, consider how a thermostat regulates the temperature: it uses a model that describes when the heating should be on or off.

The researchers found that useful models already existed in artificial intelligence for each facet of competence :

• For effectance: A model (RIDE - Rewarding Impact-Driven Exploration) that tracks exactly how much effect actions have on the environment
• For skill use: A model (VIC - Variational Intrinsic Control) that learns when which skills can best be used
• For task performance: A model (RIG – Reinforcement Learning with Imagined Goals) that measures how well predetermined goals are achieved
• For capacity growth: A model (IMRL – Intrinsically Motivated Reinforcement Learning) that tracks how skills develop

These models make it possible to define much more precisely what each facet of competence entails and how different situations affect it. These improved definitions make it possible to develop sharper measurements, create better educational software and test the theory more rigorously.

Theory and practice reinforce each other

These developments show how practical application attempts can lead to theoretical improvements. The need to implement ZDT in computer systems forced researchers to be more precise than ever before. This has not only led to better definitions, but also to new insights into how competence works. This shows how theory development and practical application can reinforce each other. The pursuit of practical applicability forces precision, which in turn leads to theoretical refinement. This development illustrates how psychological theories can grow through confrontation with the demands of the digital world.