Re:Look Ma, No Buy

sunil gupta (sgupta@umich.edu)
Sun, 12 Feb 1995 09:48:06 -0800

In response to my earlier post "LOOK MA, NO BUY", Jill Ellsworth cautions that
"this is research that needs a pretty hard look". We agree. And, to
facilitate
this process, and encourage suggestions for improvement, we have made details
about our methods and results fully available to the Web community. Caveats
and cautions have been discussed in an earlier posting to this list (dated
12/20/94, "Some comments regarding the WWW survey"), and at our Web sites
(http://www.umich.edu/~sgupta/conres.htm#intro, and
http://www.cc.gatech.edu/gvu/user_surveys/User_Survey_Home.html#confounds ).
And, we are committed to taking specific steps (some based on critiques of
respondents, and members of this list) to make improvements the next time
around.

However, it is also important to note that the general nature of our
results agree
with those from other technology/computer/online surveys (providing some
convergent validity, albeit limited), that the respondents who filled out only
the initial questionnaire(s) in our survey were not significantly different
from
those who filled out all of the surveys (providing an approximate assesment of
non-response bias), and our results are not at great variance with what the
experts might expect (providing face validity, e.g., as many have mentioned on
this list, our data also show that browsing is more common than buying).

Consequently, I think that rejecting our results and their implications, out of
hand, as Dr. Ellsworth proceeds to do, is extreme. Regarding the topic of my
posting, the discussion thus far, has featured the opinions of knowledgeable
individuals on this list, and anecdotes assembled by reporters for magazines
(Forbes and Computerworld). To this I added conclusions based on responses from
hundreds and thousands of Web users. The results agree with the basic
premises of
past discussions, but provide additional detail regarding information
search which
allows for a richer conclusion. And, in the entire discussion I took care to
repeatedly use the phrase "for our respondents ...". Obviously, savvy
decision
makers and inet-marketers should take the entire discussion into account.

Those who would like additional details about how the original survey
announcements
were made, and our views on issues of reliability, validity and biases, should
send me e-mail (sgupta@umich.edu). And from all, we solicit specific
suggestions
for how to improve the survey the next time around. The better we can make it,
the better the results and data available to the entire Web community.

Sunil Gupta
HERMES Project
Michigan Business School
Ann Arbor, MI 48109-1234
(313-764-6355)