The magical number seven plus or minus two appears in every introductory text. It also appears in a lot of design advice about menu lengths, which makes me suspect it has been stretched well past what the original paper said.
Is the number still accepted, and what was actually measured?
Accepted Answer
Miller's 1956 paper measured immediate recall of unrelated items, and the seven applies to chunks rather than to raw items. A chunk is whatever your existing knowledge lets you treat as one unit.
That distinction is the interesting half. The string 1 9 4 5 is four digits to someone who reads it as digits and one chunk to someone who reads it as a year. Capacity has not changed, the encoding has.
The current estimate is lower, around four chunks, following Cowan's work in 2001. Seven survives in textbooks partly by inertia and partly because it is a better slogan.
The design advice is where it goes properly wrong, and it is worth being explicit since you raised it.
Miller measured immediate recall of items held in mind with no external support. A menu is on the screen. You are reading it, not remembering it, so the capacity limit does not apply.
Miller himself was irritated by the reach of the number. Applying a recall limit to a visible list is using a finding in a situation it never described.
Also note the paper's own framing. Miller called the recurrence of seven across unrelated experiments possibly coincidental, and said so in the title's joke about being persecuted by an integer.
The discussion on each answer is open to members.
Join free to read the rest