Mixed-methods design, IRB exempt (UF Protocol ET00049273). Eleven participants aged 18–26 completed one session each over five days in March 2026, recruited via Discord and personal networks and distributed across four self-reported gaming-frequency bands.

Each session ran ~20 minutes: a short pre-play survey, a 5–7 minute gameplay session with no instruction beyond “pick the mode that feels right,” and a 5–7 minute semi-structured interview. Quantitative gameplay data was paired with qualitative interview data analyzed via Braun and Clarke’s six-phase thematic analysis.

Gaming frequency did not reliably predict tutorial preference. But eight themes surfaced across interviews, and two were unusually strong: all 11 participants affirmed that being able to select their own guidance level was a positive experience.

Stated preference diverged from behavioral usage — several participants who said they preferred Explicit actually spent the majority of their session in Minimal.

The prototype was designed to be the artifact. Jump was intentionally mapped to “B” rather than space bar to neutralize prior muscle memory — a choice that surfaced in interviews as a design decision in its own right. Placeholder art was used throughout so visual polish wouldn’t introduce unintended differences between tutorial states.

The project won “Outstanding Achievement in Experience Design” at Convergence 2026. The thesis now sits behind my senior capstone, alongside a full process gallery on Behance where you are invited to play it yourself!

The broader takeaway: adaptive tutorial systems built only on performance data — or only on stated preference — miss part of the picture. The most reliable signal across 11 participants wasn’t which mode they picked; it was that being allowed to pick at all mattered to every one of them.