Re: CALL FOR PARTICIPATION: Demographics Workshop

Terje Norderhaug (Norderhaug.CHI@xerox.com)
Sat, 2 Dec 1995 12:16:43 -0800


At 5:54 PM 12/1/95, Paul Burchard wrote:
>Rohit Khare <khare@pest.w3.org> writes:
>>
>> WORKSHOP ON INTERNET SURVEY METHODOLOGY AND WEB DEMOGRAPHICS
>
>> 5. Are ethical issues raised by this activity, and what might be
>> needed to protect individual rights?
>
>I hope this workshop will also include informed discussion of the
>international _legal_ issues surrounding privacy.

The ethical discussion should also go beyond merely privacy issues,
particularly when providing the data requires some effort on the respondent
(rather than being automatically collected).

Most of the current surveys done by email or on the web are a vaste of
people's time. It is all to easy to make a "survey" and poll a high number
of people without considerations for scientific methods such as sampling to
reduce the number of people that are bothered. The collective time vasted
by many of the surveys are considerable, in many cases without requiring
any substantially increased time spent by the researcher.

The methods used by most - if not all - surveys on the web are so flawed
that they in essence are worthless, as the results can NOT be used to draw
any conclusions about the general population of web users outside those
that completed the survey. This includes surveys that might appear
"scientific" and sound, e.g. the "Web Survey" (Hi Jim ;-) where the
sampling strategy is so flawed that you might have got the same result by
inviting people on the entrance of any WWW conference to complete the
questionaire, thus far from providing any information about what
characterizes the average WWW user. It is an ethical consideration when
surveys pretend to be "academic" as many people without knowledge of
scientific methods will believe that the results are usable. The ethical
responsibility also lies with the reviewers that let junk research pass to
publication.

Some of the surveys don't make their results public. Vasting people's time
and resources without giving something back should be on the list of
ethical considerations. People should refuse to answer any survey that does
not clearly state in which way the results will be made available or
contribute to our shared knowledge.

There seem to be a misunderstanding among many that creates surveys on the
Internet that getting a higher number of people to respond to a survey make
the results more valid. Anybody with the slightest knowledge of scientific
methods will know that this is only to a very limited extent the case.
Appropriate sampling strategies may make it possible to draw better
conclusions, while at the same time reducing the number of people that has
to participate to create useful knowledge. If you decide to follow a link
from a page that encourages you to fill out a survey, you should know that
this probably do not increases the accumulative knowledge about the web,
and rather is a vaste of your time.

-- Terje <Norderhaug.CHI@Xerox.com>
<URL:http://www.ifi.uio.no/~terjen/>