M-Mode / I-Mode Observation

Description

Code each moment of a session by whether the person is applying rules mechanically or stepping back to reason about the system, and study what makes them switch.

In the MU-puzzle Hofstadter separates working inside a system, applying its rules step by step (the Mechanical mode), from thinking about the system, noticing patterns and asking whether the goal is reachable at all (the Intelligent mode). People using software move between the same two modes. Most of the time they follow learned steps without thought; when something unexpected happens they shift to reasoning about how the product works. Those shifts are expensive and revealing: they mark where the interface stopped carrying people and where their understanding gets tested. This method adds a mode code to ordinary usability or field observation. Every segment is coded M (following steps, little hesitation, no talk about the system) or I (pausing, theorising aloud, trying something to see what happens), and the analysis focuses on the triggers for each switch into I-mode and on whether people come back to M-mode with a correct theory or a wrong one.

Attributes

Loss type
Latent
Effort
Days

Steps

  1. Run normal think-aloud sessions or field observation, recorded so moments can be timestamped.
  2. Agree on markers for each mode before coding, such as hesitation over 3 seconds, statements about how the system works, or exploratory clicks for I-mode.
  3. Code the timeline of each session into M and I segments; have two people code a sample and resolve differences.
  4. List every switch from M to I with what happened just before it.
  5. Group the triggers (unexpected state, missing feedback, broken pattern, error message) and note the theory each person formed.
  6. Prioritise triggers that are frequent, long-lasting or lead to wrong theories.

Sources

  • Hofstadter, D., Gödel, Escher, Bach: An Eternal Golden Braid (1979)

Also known as

  • Mechanical versus intelligent mode observation
  • Rule-following versus reasoning coding

Categories

  • User Research
  • Usability & Evaluation