Statement of Commissioner Orson Swindle
Concurring in Part and Dissenting in Part
to Prepared Statement of the Federal Trade Commission
on "Online Profiling: Benefits and Concerns"
before the Committee on Commerce, Science, and Transportation
United States Senate

File No. P994809

I concur in the issuance of the Prepared Statement of the Federal Trade Commission on "Online Profiling: Benefits and Concerns" before the Committee on Commerce, Science, and Transportation, United States Senate (June 13, 2000) ("Commission Statement"), but I dissent from how certain consumer opinion surveys are used in the Commission Statement.

First, consumer opinion surveys like the ones used in the Commission Statement often are not reliable predictors of consumer behavior. For several reasons, and as the Commission Statement acknowledges in footnote 8, survey results should be examined with scrupulous care. Surveys are one-time snapshots of consumer opinion, are easily biased by design, and must be examined for methodological integrity.

Ideally, consumer opinion surveys should complement, but not be a substitute for, empirical evidence of consumer behavior relating to privacy. They should not serve as the substantive basis for policy.⁽¹⁾

Second, when the Commission reports to or testifies before Congress, it owes the Congress a certain degree of thoroughness. A statistic included in a Commission report likely will be given credibility beyond what might attach to the use of that same number in a brief news story or an advertisement. Because of the added degree of credibility attached to a Commission report, the Commission should not uncritically repeat estimates, projections, or other statistics unless it knows how the numbers were derived, including the assumptions on which they may have been based. This requires going directly to the source of a number. If that standard of analysis cannot be met, then the Commission either should not use the number or should explicitly qualify its use of the number by the uncertainties attached to it.

For example, both the Online Profiling Report and this testimony contain an estimate of future advertising revenue drawn from an overview of a July 1999 report by a management consulting firm. (see "Online Profiling: A Report to Congress" at 2, n.7; Commission Statement at n.22). The Commission has no basis for assessing what assumptions went into that projection, nor does the Report or the testimony highlight that the July 1999 date of the projection alone likely means it is less accurate in light of the tremendous growth in online commerce since then. In my dissent from the Commission's 2000 Privacy Report, I criticized the Commission's use of a lost sales projection by the same management consulting firm based on the repetition of that projection in a news article and the information available from an online overview of the study. An examination of the full study revealed that the lost sales projection was based on assumptions that completely invalidated the Privacy Report's reliance on that lost sales projection. See 2000 Privacy Report, Dissenting Statement of Commissioner Orson Swindle at 13-14.

Another example of relying on numbers without assessing their validity is the testimony's reference to an Odyssey study in which 92% of respondents from online households stated that they do not trust online companies to keep their personal information confidential. (Commission Statement at 5-6 n.10). This figure comes from the same Odyssey Study cited by the majority in the Privacy Report and appears to be subject to the same flaws that I discussed in my dissent from the Privacy Report. Unfortunately, the Odyssey Study does not reveal the specific questions used to derive the 92% that either agree or strongly agree with the proposition repeated in the Commission Statement. If the Odyssey Study uses the same methodology as for other questions, it likely biases the responses to "agree" categories by not allowing a choice to "somewhat disagree." (See 2000 Privacy Report, Dissenting Statement of Commissioner Orson Swindle at 11.)

I respectfully ask that Congress keep these limitations in the data in mind as it considers the Commission's Online Profiling Report and the Commission Statement.

1. A portion of my dissent from the Commission's 2000 Privacy Report addressed the Commission's dubious reliance on consumer opinion surveys. See Dissenting Statement of Commissioner Orson Swindle, Federal Trade Commission, "Privacy Online: Fair Information Practices in the Electronic Marketplace: A Report to Congress" (May 22, 2000) at 12-16.