Liste des Groupes | Revenir à rb tech |
On 9/8/2024 9:24 PM, Jeff Liebermann wrote:On Mon, 09 Sep 2024 09:01:29 +0700, John B. <slocombjb@gmail.com>>
wrote:
As I've posted before, a good friend did financial studies - Is it
practical to sell refrigerators on a time purchase pay plan and Is it
financially logical to establish a copper processing plant in
Thailand, was two I remember. He once pointed out that a "survey" can
be constructed to produce any result that you want to produce.
>
Since then I've been very leery about surveys :-)
Good friend is quite correct. Survey can and often are written to
favor a desirable result. I know, I've done it. Some example:
"Different Types of Survey Bias and How to Avoid Them"
<https://www.surveylegend.com/survey-bias/survey-bias/>
"Avoiding biased questions: 7 examples of bad survey questions"
<https://delighted.com/blog/biased-questions-examples-bad-survey-questions>
"How to Create a Bad Survey Instrument"
<https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5559231/>
Incidentally, every time I visit a doctors office or hospital as a
patient, I receive a survey form via email or text message asking
questions about the visit. The questions are usually fair, the
choices clear, but there is still a big problem. They don't ask
questions that reflect my concerns about the visit or medical
practices in general or might reflect badly on the doctor or hospital.
In other words, the omission of important questions and concerns is
just as important as the inclusion of biased questions.
I just don't. Ever.
>
The typical customer survey has questions offering a choice
of three or four irrelevant, inapplicable answers. pfffft.
Les messages affichés proviennent d'usenet.