BoatBanter.com - View Single Post

(BW) typed:

Maybe my last THRILL posting appeared a bit like SPAM. Just to allay
any questions - I'm a researcher at the Royal College of Art, England.
I've got an arts grant to research this subject of thrill, and develop
ideas for how thrill could be engineered.

I'd recommend you ask yer tutor about course offerings in research
methodology. The so-called "survey" you present offers nothing in the
way of scientific rigor. Or even rigour, for that matter. Have you
ever wondered why so many surveys you have seen include question after
question where you are asked to "pick the one (and only one) answer
that most closely matches tour feelings: agree strongly, agree
slightly, disagree slightly, disagree strongly", or "Pick the one (and
only one) answer that most closely describes youe experience: always,
often, sometimes, rarely, never"? Notice that these answers are fixed
and finite, and describe a continuum from one extreme to the other.

Questions couched as yers (known as "open-ended questions" amongst
statistical researchers in the social sciences) evoke a collection of
anecdotal replies which cannot be normalized or subjected to rigorous
analysis. You need to design yer survey to ask questions that can be
answered as in the examples I've shown, such that a reply of "agree
strongly" can be assigned a value of 4 and "disagree strongly" can be
assigned a value of 1. You can then statistically determine that (for
example) "males aged 21-25 tended to agree slightly (the mean was
2.8), but urban white males disproportionately agreed fairly strongly
(3.6)".

'Course, that is all irrelevant, since you are trying to "administer"
this "survey" in a totally uncontrolled fashion. Remember those
classes I typed of in research methodology? They will also train you
in how to select a qualified sample, how to control for bull****
responses, how to induce candidate respondents to co-operate
honestly...

If you cannot quantify it in a way amenable to statistical analysis,
it ain't research. If you cannot show that yer sample was rigorously
selected, it ain't research. If you do not have a reliable control
(comparison) sample, equally rigorously selected, it ain't research
Yer Grants Committee oughta have its collective head examined for
giving research funding to someone with no training in survey
methodology.

-Richard, His Kanubic Travesty
--
================================================== ====================
Richard Hopley, Winston-Salem, NC, USA
rhopley[at]earthlink[dot]net 1-301-775-0471
Nothing really matters except Boats, Sex, and Rock'n'Roll.
rhopley[at]wfubmc[dot]edu 1-336-713-5077
OK, OK; computer programming for scientific research also matters.
================================================== ====================