Re[2]: Mac vs. PC survey

[Arlen -dot- P -dot- Walker -at- JCI -dot- COM]

My last comment on this topic:

      Then I'm sure you'd find it equally "interesting" if someone used
      their personal web pages to find out if people who dropped by
      preferred mice to "smart keys" in their use in word processors.  I
      wouldn't.  Anecdotal data is *always* statistically suspect.

Sheesh!
1) I had assumed the original commentator had to be joking, because I
assumed anyone on this list was capable of reading and comprehending. This
assumption is now called into question because of the continued repetition
of a statement that clearly contradicts what was stated in the original
post. The survey was not *taken* on web page, the results are *posted*
there. There is a difference between "taking a survey" and "posting the
results of a survey." Only one requires a web page. Perhaps a friend can
explain it offline.

2) Data, whether anecdotal *or* statistical, is suspect. That's why it must
be provided in full for analysis before use. One can use statistically
sound methods to produce garbage just as easily as anecdotal evidence. The
methods say nothing about the pre-suppositions of the pollster, and it is
those presuppositions which are the controlling factor for the result. A
good example of this is the Microsoft "usability study" which "proved"
Windows 95 was easier to use by setting out tasks which were carefully
constructed with the limitations of W95 in mind. (To break it down into
plainer english, polls/studies support the the point of view of the person
who pays for them. I'm sorry I failed to get that completely into words of
one syllable, but perhaps it's still understandable.)

3) Majority preference of one tool over another says absolutely zip, zilch,
zero, nada about the fitness of a particular tool for use by a particular
user in accomplishing a particular purpose. The user determines the fitness
of a tool on an individual basis. Just ask lefthanders how well they get
along with all those wonderful tools designed for the right-handed
majority.

As for the childish little snipe quoted above:
If I manufactured a Word Processor, and I had such a query on my personal
web page and told my user base about it, I certainly *would* be tempted to
use the results as input for the developers. After all, it's a measurement
of the preferences of my customer base, at least of those who cared enough
to respond. Any method for gathering those preferences is self-selecting,
as not everyone returns warranty card surveys, participates in mail
surveys, or in any other kind of non-intrusive survey that's reasonable to
attempt.

      There *are* some statistically defensible surveys on user
      preferences, but they're done by people who know something about
      statistics.

And generally know nothing about users. ;{>} There's an interesting
characteristic of users. They quite often don't know what it is they prefer
until after they've tried it for a while. They might put a particular
characteristic/feature at the top of their "wanted" list, only to find out
after it was used for a while that they really don't need it as much as
they thought, while something which they originally thought was fluff can
become indispensible after a few weeks use. (I'm sure we can *all* point to
features of our tools which fit both categories.)

None of the rags mentioned as nebulous possible sources (Look out -- I
think someone else just committed the crime of anecdotal evidence) for the
studies are noted for their rigorous statistical accuracy.

In this context, I'm reminded of my wife's favorite accounting joke:

Ask a Mathematician what the sum of 2 and 2 is, and he'll ask which number
system you want to use.
Ask an Engineer, and he'll ask how many digits of precision you want in
your answer.
Ask an Accountant, and she'll reply: "What do you *want* the answer to be?"

Like Accounting, Statistics can be used in support of anything a clever
person wants to use them to prove. In the end, all that matters is which
tool fits your hand best. And on that matter I, for one, am "shocked and
appalled" that *anyone* would let *any* kind of a popularity poll dictate
to them which tools best fit their hand. The only valid way to determine
that is to put the tool in your hand and *use* it.

Speaking of which, I think it's time we all do just that.


Have fun,
Arlen
Chief Managing Director In Charge, Department of Redundancy Department
DNRC 224

Arlen -dot- P -dot- Walker -at- JCI -dot- Com
----------------------------------------------
In God we trust; all others must provide data.
----------------------------------------------
Opinions expressed are mine and mine alone.
If JCI had an opinion on this, they'd hire someone else to deliver it.

                  TECHWR-L List Information
To send a message about technical communication to 2500+ list readers,
E-mail to TECHWR-L -at- LISTSERV -dot- OKSTATE -dot- EDU -dot-  Send administrative commands
         ALL other questions or problems concerning the list
       should go to the listowner, Eric Ray, at ejray -at- ionet -dot- net -dot-