[Sigia-l] research: when is it enough?

[Ziya Oz]

"Fiorito, David" wrote:

> The question "is our corporate intranet useful to our employees" will require
> a much more substantial sample size.

Or a better formulation of the question if you want to get anywhere: "is our
corporate intranet useful to our employees" *as perceived by*

the guy who signs the check for it
the employees
the IT people who develop and support it
the outside consulting firm auditing it
our shareholders
etc.

> There are specific formulas for arriving at a sample size that
> would be considered fairly accurate and statistically significant.

As there are notions of "what time it is" that are "considered fairly
accurate and statistically significant"...and dead wrong according to, say,
USNO Master Clock time. So what?
 
> If we approach our craft in such an unscientific fashion then we will be
> written off as walking opinion generators.  Anyone can do that.  Who needs a
> highly skilled IA when the pointy haired manager can spout opinions too.

Look, if 'methodologies' and 'sampling sizes' were all you need, then all TV
sitcom pilots (that go through endless 'research') would be instant hits,
polls would render elections superfluous, every newly introduced Nabisco
product would be a market leader, etc. In real life, some people consider
such 'research', give it the amount of attention it deserves and move on.
 
> If we are to be taken seriously then we need to get serious about what we do
> and how we come up with the answers to our employers' questions.

And one way to do that would be to avoid using 'research' as snake oil and a
substitute for thinking and common sense.

Best,

Ziya