Description
Models decision time as a logarithmic function of the number and probability of choices, to estimate how option sets affect selection speed.
The Hick-Hyman law states that choice reaction time increases with the information content of the choice: RT = a + b·log2(n + 1) for n equally likely options, or more generally with the entropy of the choice set when probabilities differ. It applies to decisions where users already know the options and must select one, not to visual search through unfamiliar lists, where time grows roughly linearly. Its rigorous use in UX means checking whether that precondition holds, estimating choice probabilities from usage data, and fitting parameters for the specific context. This avoids the common misapplication of citing Hick's law to justify removing options. Properly applied, it shows that ordering by frequency, grouping and reducing entropy can speed selection, and that splitting choices across many levels can cost more than it saves.
Attributes
- Discipline
- Task Analysis & Cognitive Modelling
- Loss type
- Active
- Stage
- Design
- Data
- Quantitative
- Effort
- Days
Steps
- Confirm the task is a choice among known options rather than search through unfamiliar items.
- Estimate option probabilities from usage logs and compute the entropy of the choice set.
- Run a timed selection task with 10-20 participants across set sizes or layouts.
- Fit RT against log2(n + 1) or entropy and check the fit.
- Use the model to compare flat, grouped and frequency-ordered designs, including the added cost of extra navigation levels.
- Choose the structure minimising total predicted time, then validate with usability testing.
Sources
- Human-Computer Interaction
Also known as
- Hick's Law
- Hick-Hyman Law
- Choice reaction time modelling
Categories
- User Research
- Task Analysis & Cognitive Modelling