Hick's Law

The law

The time it takes to choose between options grows with the number of options, roughly with the logarithm of that number rather than in step with it.

Doubling the options does not double the decision time; it adds a roughly constant amount. More options are still slower, but each new one costs less than the last.

Where it comes from

William Edmund Hick reported the finding in On the rate of gain of information, published in the Quarterly Journal of Experimental Psychology in 1952. Participants faced a row of lamps, each paired with a key. When a lamp lit, they pressed its key as fast as they could. Hick varied how many lamps were in play and found that reaction time rose with the logarithm of the number of alternatives.

Ray Hyman's 1953 study, Stimulus information as a determinant of reaction time, extended the result. When some alternatives were more likely than others, reaction time followed the amount of information in the choice, not just the count of options. The usual statement of the law, reaction time rising linearly with information measured in bits, is often called the Hick–Hyman law.

Both experiments measured choice reaction: each stimulus had one known, practiced response. That condition matters for how the law applies to interfaces.

What it means for an interface

• Familiar choices scale well. When people already know the options, such as a keyboard shortcut set or a well-learned menu, adding items slows them only a little.
• Likely options should be cheap. Hyman's result means a choice is faster when one option is much more likely. A sensible default, or the most common action made prominent, reduces the work of deciding.
• Unfamiliar choices cost more. When people must read and understand each option, they are no longer in Hick's experiment. They are searching, and searching an unfamiliar list takes time that grows with its length.
• Split decisions that are genuinely separate. A form that asks for one decision at a time, with only the options relevant to that decision, keeps each choice small.

Examples

Plan selection. A pricing page with three clearly different plans is quicker to decide on than one with seven overlapping tiers, because each option has to be read and compared. Marking the plan most people choose turns one option into the likely one.

Settings. A settings screen that shows every option at once makes people scan the whole page to find one switch. Grouping settings by task and hiding rarely used ones behind an "Advanced" section keeps each view short.

Command menus. Experienced users of an editor pick from long command menus quickly because they know where things are. That is the case Hick measured, and it is why long menus can work for experts.

Where it is misapplied

• "Fewer options is always better." The law describes the time to choose, not the value of the result. Removing an option people need trades a small speed gain for a failed task.
• Applying it to search. A person scanning an unfamiliar list reads items one by one. That time grows roughly with the list's length, and the logarithmic curve does not apply.
• Hiding options to shorten a list. Moving choices into nested menus adds steps and makes each option harder to find. A longer, well-organized list is often faster than a deep tree.
• Citing the formula as a prediction. The constants in the formula were fitted to lamps and keys. They do not predict how many milliseconds your menu will take.

Checklist

• The most common choice is the default or the most prominent option.
• Each step asks for one decision, with only the options relevant to it.
• Rarely used options are grouped out of the way, not removed.
• Lists people must read are ordered or grouped so they can be scanned.
• Option labels are distinct enough that people do not have to compare them closely.

Related

• Miller's Law: how many options people can hold in mind while comparing.
• Fitts's Law: once chosen, how long the option takes to reach.
• Jakob's Law: familiar options are the fast case Hick measured.
• Patterns: Accordion.