Re: The Septapartist Myth

["Nagai, Paul" <pnagai -at- VISA -dot- COM>]

First,

Thanks Mark. I've posted the URL to Miller's article here before, but
getting input from the man himself, nicely done.

Chris says;
> 1) Usability Tests reveal that many users will try and digest a
> complete
> list (in my case a procedure viewed via Winhelp, e.g. Printing an
> <item>) before attempting to execute the task in the interface.
> Therefore, elements of numbered list that exceeds 7 points are likely
> to
> be forgotten, resulting in the user returning the Help again.  Miller
> himself states that his work related to 'immedaite recall' and in this
> case, it applies whole heartedly.
>
First: I'm quite willing to bet that usability tests do not, with any
scientific rigor, define the user's experience and knowledge related to
the procedure in question. Second: The seven list elements are not
(taking into consideration each user's experience and knowledge)
controlled such that they represent seven "chunks." Third: Miller's
immediate recall is in a laboratory environment where there are no
auditory or visual distractions, no computer interface to interact with,
etc. ALL of which interfere with the type of memory Miller wrote about
... short term memory ... which is fairly narrowly defined in
perception/cognition research world.

It is very, very, very difficult to know (or even study) whether or not
any given user or test-subject has been presented with 7+/-2
items/chunks, using the type of content we're in the business of
creating, and thus to conclude what they will remember for a very brief
period of time, without distraction or interference.

If the research described above is correct about user behavior
(memorizing the entire list ... and I have no reason to doubt that part
of the research), then Miller's research would suggest that writers
should consider limiting lists to well below 7 items given the
complexity of the real-world computing environment.

I have no argument with anyone, with or without carefully constructed
empirical tests, who claims the number of steps or points in printed (or
online) documentation needs to be carefully controlled and when
possible, fewer such steps or points will lead to better comprehension,
storage, and, eventually, recall.

However as Miller himself has now spoken on the subject ... get over the
"magic" number 7 ... at least as it relates to technical writing.

------
Paul Nagai


 From ??? -at- ??? Sun Jan 00 00:00:00 0000=