Page 207 - Privacy in a Cyber Age Policy and Practice

NOTES 195

16. Institute for Statistics Education, “Glossary of Statistical Terms Test-Retest
Reliability,” available at http://www.statistics.com.
17. See, for example, Oscar H. Gandy Jr., “Public Opinion Surveys and the
Formation of Privacy Policy,” Journal of Social Issues 59 (2003): 283–99
(the difficulty of framing neutral questions is “especially problematic in
the realm of privacy policy”); Susan Freiwald, “A First Principles Approach
to Communications’ Privacy,” Stanford Technology Law Review (2007): 3
(questions on privacy might be “too complicated and too easily skewed” to
give accurate results).
18. Ruut Veenhoven, “Why Social Policy Needs Subjective Indicators,” Econstor,
available at http://www.econstor.eu/handle/10419/50182.
19. Pew Research Center for the People and the Press, “Methodology: Question
Wording, ” available at http://www.people-press.org/; examples of such bias
include “social desirability bias” (“inaccurate answers to questions that deal
with sensitive subjects” like drug use or church attendance, especially in face-
to-face interviews), “acquiescence bias” (In a poll asking whether military
strength was the best way to secure peace, 55 percent were in favor when it
was phrased as a “yes or no” question, but only 33 percent were in favor when
“diplomacy” was offered as an alternative), and question order effects. (A 2008
poll found that an additional 10 percent of respondents expressed dissatis-
faction with current affairs if they were previously, rather than subsequently,
asked if they approved of the president’s performance). See also Andrew
Binder, “Measuring Risk/Benefit Perceptions of Emerging Technologies,” Pub-
lic Understanding of Science (2011) accessed at http://pus.sagepub.com/. (Short
opinion polls may yield different results than longer academic surveys.)
20. B. J. McNeil et al., “On the Elicitation of Preferences for Alternative Therapies,”
New England Journal of Medicine 306 (1982): 1259.
21. The archetypal example of survey error leading to false results was the famous
“Literary Digest Poll” that falsely predicted Roosevelt’s loss of the 1936 presi-
dential election results by relying solely on telephone and car owners, a dispro-
portionately Republican group. See also “The War Over Love Heats Up Again,”
Los Angeles Times, October 29, 1987. (A 1987 mail-in survey on love found
that 98 percent of women were unhappy in their relationships, while a tele-
phone poll found that 93 percent were happy—possibly because the unhappy
had more motivation to mail in their response); Russell D. Renka, “The Good,
the Bad, and the Ugly of Public Opinion Polls,” Southeast Missouri State Uni-
versity. (Internet polls, which tend to attract an unrepresentative sample of the
population and to lack safeguards against multiple voting, can be particularly
susceptible to this type of errors.)
22. See, for example, Floyd J. Fowler, Jr. Survey Research Methods, 4th ed.
(Thousand Oaks, CA: SAGE Publications, Inc., 2009). SAGE Research
Methods. Web.
23. Jason Zengerle, ‘‘The. Polls. Have. Stopped. Making. Any. Sense,” New York
Magazine, (September 30, 2012); Thomas Fitzgerald, “Rethinking Public
Opinion,” The New Atlantis, 21 (Summer 2008): 45–62.