In an area as complex as privacy, it is unlikely that surveys can accurately
reflect consumers' true preferences and concerns.
Survey results can vary quite widely based on what questions are asked and
not asked. Polls that lead off with provocative questions, for example, are
likely to get responses that address the provocation. Other questions are
phrased to guarantee a certain response. A survey asking
"How concerned are you about strangers getting personal information about you
and your family?" is guaranteed to get a response that claims high levels of
privacy concern.
Many privacy surveys distinguish poorly among a variety of issues. Thus,
credit fraud and identity fraud may be referred to as privacy problems,
even though these things are crime problems unrelated to legitimate uses of
information. Spam and junk mail may be referred to as privacy
problems even though they are better characterized as matters of annoyance and
inconvenience. "Privacy" may also be used almost interchangeably with "security."
Many surveys are of little or no use in solving real problems because they do not
identify what version of privacy the respondent is talking about.
Many privacy surveys also mislead as to the strength of consumers' interest
in privacy. For example, when consumers are given the option of stating only
whether they are "concerned" or not, they inevitably say that they are — and
indeed they should be. But this does not place privacy on a scale with other
concerns. How does privacy compare to the opportunity to learn about a child
safety seat recall? How does protecting privacy compare to the opportunity to
go see a movie? If given a choice of how to spend five minutes, would a consumer
have a hot dog, opt out of the data sharing his or her bank does, look up what
a "cookie" is online, or watch the sun set? Privacy surveys do not reveal this,
but choices like these are the best indicators of true concern.
The least manipulative and probably most accurate form of consumer survey is an
unprompted survey, in which people are asked to list the issues of concern to them
without being prompted or given a list of possible responses. This type of survey
is often used to identify election issues because candidates need to know what will
motivate votes, not just what voters will say to a pollster. In recent such surveys,
privacy does not poll strongly in relation to other concerns such as medical care,
education, crime, or Social Security.
Links:
Comments? comments@privacilla.org
(Subject: PrivacySurveys)
[updated 07/19/01]
