[Sigia-l] A matter of reliability

[Trenouth, John]

> Do we stop accepting projects that do not allow statistically
significant
> user interviews?

Henry Ford said "If I had asked my customers what they wanted, they
would have asked for a faster horse."  So studying statistically more
significant numbers of customers would have simply revealed more people
who wanted faster horses.  

Statistical significance is for the design world like Emerson's
hobgoblin.

Just for a quick and familiar example, over a decade ago Virzi, Sokolov
and Karis from Verizon Labs (then GTE Labs) conducted a study that
determined a sample of about 8 users is a sufficiently large sample.
Beyond that the cost no long justifies the effort due to diminishing
returns.  

If you have an ACM membership you can find the paper at:
http://portal.acm.org/citation.cfm?id=238516