Difficulty-skill balance does not affect engagement and enjoyment: a pre-registered study using artificial...

Area of Science:

• Psychology
• Human-Computer Interaction
• Artificial Intelligence

Background:

• Intrinsic motivation and flow theories suggest optimal engagement occurs when task difficulty matches performer skill.
• Previous research in this area is limited by measurement issues and the lack of scalable methods to control objective difficulty-skill ratios.
• Understanding this relationship is crucial for designing engaging experiences.

Purpose of the Study:

• To investigate the impact of precisely manipulated difficulty-skill ratios on task enjoyment and engagement.
• To develop and validate a novel AI-controlled game environment for studying the difficulty-skill balance.
• To address limitations in previous research by employing objective difficulty manipulation and rigorous methodology.

Main Methods:

• Developed a two-player tactical game test suite featuring an AI opponent.
• Utilized a variant of the Monte Carlo Tree Search algorithm to precisely control difficulty-skill ratios.
• Conducted a pre-registered study with 311 participants, assessing enjoyment and engagement levels across different ratios.

Main Results:

• The AI successfully produced targeted difficulty-skill ratios without participants detecting the manipulation.
• No significant differences in enjoyment or engagement were observed across the varied difficulty-skill ratios.
• Participant awareness of the manipulated difficulty-skill ratios was minimal.

Conclusions:

• The 'goldilocks' hypothesis of optimal difficulty-skill balance may not universally apply to enjoyment and engagement.
• AI-controlled difficulty presents a viable and scalable paradigm for future research on task engagement.
• Further research is needed to explore other factors influencing engagement beyond simple difficulty-skill matching.