Hick's Law

011 min

Hick's Law at a glance

• What it is: Choice reaction time grows with the log of the option count, not linearly.
• Where it comes from: Hick (1952) and Hyman (1953) measured reaction time to lit signals in the lab.
• Apply it by: Group and prioritize options; deleting them outright is not the only fix.
• It backfires when: Splitting one choice into many small screens can add time, not save it.

022 min

What Hick's Law actually says

Hick's Law says that the time a person takes to choose among a set of options grows with the logarithm of the number of options, once those options are presented together, are equally likely, and are already known to the chooser. Written as a formula: T = a + b × log₂(n + 1). T is the choice reaction time, n is the number of options, and a and b are constants fitted to the person and the task: a is the fixed baseline time to notice a signal and start responding at all, and b is how much each doubling of the options adds.

The logarithm is the part most people get wrong on first read. Going from 2 options to 4 adds a fixed amount of time, and going from 4 to 8 adds roughly that same amount again — not double it. Wikipedia states the practical result plainly: "The plain language implication of the finding is that increasing the number of choices does not directly increase the time to choose. In other words, twice as many choices does not result in twice as long to choose." The brain does not check each option one at a time; the formula behaves as if it eliminates half of the remaining options at each step, the way a binary search narrows a sorted list.

Hick's 1952 paper covered options that were all equally likely. Hyman's 1953 paper extended the formula to options that are not equally likely, replacing the plain count of options with a measure from information theory called entropy. A signal that appears nine times out of ten adds almost no extra choice time, because it is already expected. A signal that appears once in ten, unpredictably, adds far more time, even inside a set of the same size. This is why the finding carries both names: the Hick-Hyman Law is Hick's original relationship, generalized by Hyman to account for how predictable each option is, not only how many there are.

031 min

Where Hick's Law comes from

William Edmund Hick published the founding paper in 1952, in the Quarterly Journal of Experimental Psychology. Ten lamps were arranged in a circle around a seated participant, each lamp paired with a telegraph key operated by a different finger. A pre-punched tape lit a random lamp every five seconds, and the gap between the light coming on and the matching key being pressed was recorded on a moving paper strip. Wikipedia describes the setup directly: "In his first experiment, 10 lamps were arranged in a circle around the subject, each paired with a Morse key operated by a different finger." As the number of lamps in play increased, reaction time rose — by a shrinking amount each time, not a steady one.

Ray Hyman published a companion paper the following year, in the Journal of Experimental Psychology. His experiment varied not just how many signals were possible, but how equally likely each one was and how predictable it was given what had come immediately before. The same logarithmic relationship held once the plain count of options was replaced with a measure of how much genuine uncertainty each choice actually carried.

042 min

How Hick's Law shows up in real interfaces

Human-computer interaction research on menu depth backs this up. A University of Maryland study of web menu hierarchies reported: "reducing the depth of hierarchies improves performance in terms of speed and search efficiency" (Zaphiris, Shneiderman & Norman, 1999). That study built on an earlier finding that "short-term memory is a limitation of the increased depth of the hierarchy" (Miller, 1981). Cutting the number of choices on each screen, at the cost of adding more screens, can make the whole task slower — the opposite of what a designer citing Hick's Law usually intends.

051 min

A second case: the same menu, two different users

Hick's Law is usually paired with Fitts's Law to model how long it takes to use a desktop application's menu, and a 2007 menu-performance study by Andy Cockburn, Carl Gutwin, and Saul Greenberg shows why the pairing only works once expertise is accounted for. Their model splits menu-selection time into two different processes, and which one applies depends on whether the person already knows where the item sits.

A new user, who has not yet learned the menu's layout, has to visually scan the list to find the item. The researchers describe this directly: "there is consistent empirical evidence that novices' search time is linear with menu length." A frequent user, who has built a spatial memory of where things are, skips the scan entirely and moves straight to a remembered position. At that point, the same study reports, "for expert users there is little visual search involved in menu selection, reducing the task time to Hick-Hyman decision time plus Fitts' pointing time."

The variable that changes between these two cases is not the menu. It is whether the chooser already knows what the choice contains. Hick's Law predicts the expert's time well. It says almost nothing useful about the novice's, because the novice is not yet choosing among known alternatives — they are still searching for one.

062 min

Applying Hick's Law

Hick's Law is easiest to apply well by grouping and prioritizing, not by deleting options outright:

• Group related options under a small number of labeled categories, so a person scans a short category list instead of the full option list.
• Put the option most people actually pick first, or make it the visible default, so choosing it costs no separate decision.
• Reveal advanced or rarely used options only once someone asks for them, instead of removing them for good.
• Order a list by how often each item actually gets used, not alphabetically, so the common case sits where the scan starts.
• Where speed genuinely matters and every option is equally valid — an emergency stop control, a single confirmation step — keep the count itself down to two or three, since that is where each added option costs the most time.

A long menu does not have to violate Hick's Law by itself: as Nielsen Norman Group puts it, "combining Hick's Law with other design techniques can make long menus easy to use" — grouping, search, and recency all lower the effective choice count without deleting anything.

It backfires when a team reads "fewer choices, faster decisions" as permission to hide functionality a returning user actually needs. Cognitive Load Theory covers the related cost of holding several options in mind at once. Working memory explains why hiding an option is not free either — a person who already holds it in mind pays almost nothing extra for it staying visible, while hiding it behind a menu costs them a fresh search.

072 min

When Hick's Law does not apply

Hick's Law is a claim about one specific kind of moment: choosing among a known, fixed, simultaneously visible set of options that are already understood. It is not a general claim that more options are always worse, and treating it as one is the most common misreading of the law in interface design.

Searching is not choosing. Wikipedia's own entry states the limit directly: "Hick's law is sometimes cited to justify menu design decisions. For example, to find a given word (e.g. the name of a command) in a randomly ordered word list (e.g. a menu), scanning of each word in the list is required, consuming linear time, so Hick's law does not apply." A person searching for an unfamiliar item in an unfamiliar list is doing a linear search, not a logarithmic choice, no matter how few items sit on any one screen.

Splitting one decision into several smaller ones is not the same as reducing the decision. Hick's Law prices a single simultaneous set of options; it says nothing about the extra baseline time each additional screen adds on top. And the law says nothing about whether an option is easy to understand once seen — a menu of three confusingly worded options can cost more time than a well-labeled menu of eight, because Hick's Law measures choosing among options, not understanding them. That failure runs in the other direction from Choice Overload, which is about too many options at once rather than too few per screen spread across too many screens.

081 min

Hick's Law vs. nearby concepts

Hick's Law is often confused with the Paradox of Choice, since both describe a cost from having more options — but they measure different things. Hick's Law comes from controlled experiments on reaction time: how long it takes to choose, in fractions of a second. The Paradox of Choice is about satisfaction: whether a person feels worse about the option they picked once more alternatives existed. A choice can be fast under Hick's Law and still leave the person who made it regretful.

Hick's Law is also paired with, and sometimes confused with, Fitts's Law. The two are usually taught together because most interface actions combine them: Hick's Law prices deciding which target to go for, and Fitts's Law prices physically reaching it. A menu with few items but a tiny, hard-to-hit target can score well under Hick's Law and poorly under Fitts's Law at once.

091 min

Where the evidence is contested

Hick's Law replicates well for the exact task it was measured on: choosing quickly among a known, equally likely set of options shown at once. Later studies found real exceptions, not just an occasional bad fit.

Wikipedia's summary of that research states: "Exceptions to Hick's law have been identified in studies of verbal response to familiar stimuli, where there is no relationship or only a subtle increase in the reaction time associated with an increased number of elements." In these cases, adding more options barely slows the response at all, which the logarithmic formula does not predict well.

A sharper exception involves saccades, the fast, automatic eye movements people make while scanning a scene. Some saccade studies found the opposite of Hick's Law's prediction: reaction time stayed flat, or even dropped slightly, as the number of possible targets grew.

There is also disagreement about the shape of the relationship itself. Some researchers argue a sigmoid, S-shaped curve fits certain reaction-time data better than a straight logarithmic line, particularly where a response is highly predictable from what came right before it.

None of this undoes the core finding inside its original domain — a person choosing quickly among a small, known, equally likely set of options. It does mean the "more options always means slower" reading, applied more broadly, claims more than the evidence supports.

101 min

How Hick's Law changed since 1953

Hick and Hyman's 1952 and 1953 papers gave the field a formula. What happened afterward was mostly about testing how far that formula's predictions extend.

In 1964, researcher E. Roth used the relationship to study differences between individuals: how quickly a person's reaction time grew as choices increased correlated with their measured IQ, with faster processors showing a shallower slope in the curve. Wikipedia summarizes the finding: "E. Roth (1964) demonstrated a correlation between IQ and information processing speed, which is the reciprocal of the slope of the function." That changed the constant b from a fixed number into a value that varies meaningfully between people.

Decades later, human-computer interaction researchers returned to the law for a narrower reason: predicting how fast people would use software menus. Steven Seow's 2005 comparison of the Hick-Hyman Law against Fitts's Law, and the Cockburn, Gutwin, and Greenberg menu-performance model discussed earlier, both treated the original 1952 formula as one ingredient inside a larger model of menu use, not a complete account on its own.

The core relationship Hick and Hyman measured is still the accepted default for its original domain. What changed is how narrowly researchers now define the conditions under which it applies.

112 min

Frequently asked questions about Hick's Law

?7 questions

Questions people ask

§8 sources

Sources