In assessing a phenomenon, it is essential to “filter” the available evidence. A key filtering procedure is represented by the definition of the phenomenon. In this respect, Condon’s definition, which has already been quoted, suffers from the defect that it allows a great deal of “noise” to accompany whatever “signal” there may be in the data. Most students of the UFO phenomenon would adopt a more restrictive definition such as that adopted by Hynek (1972), who recommends that a “UFO report” be defined as “a statement by a person or persons judged responsible and psychologically normal by commonly accepted standards, describing a personal, visual, or instrumentally aided perception of an object or light in the sky or on the ground and/or its assumed physical effects, that does not specify any known physical event, object, or process or any psychological event or process.” However, the definition of the phenomenon is only one filtering procedure. In discussing a complex phenomenon such as the UFO phenomenon, it should be followed by further “filters” that may comprise restrictions on allowable evidence, classification schemes, etc. The staff summaries, indeed, provide a breakdown of evidence into categories, but this is only a rudimentary scheme of analysis.

Another important point of scientific methodology is that, if one is evaluating a hypothesis (such as ETH), it is beneficial to regard this hypothesis as one member of a complete and mutually exclusive set of hypotheses. This point also seems to have been clearly recognized by Thayer (Condon & Gillmor, 1968, p. 116), but it was apparently ignored by Condon and by other members of the project staff.

Finally, in evaluating a hypothesis, one must avoid procedures of data reduction that depend upon the truth or falseness of that hypothesis. Put another way, one must avoid “theory-dependent” arguments. This requirement, above all, makes the appraisal of the UFO phenomenon very difficult: if we entertain the hypothesis that the phenomenon may be due to an extremely advanced civilization, we must face the possibility that many ideas that we accept as simple truths may, in a wider and more sophisticated context, not be as simple and may not even be truths.

As a specific example, one may draw attention to the argument (Condon & Gillmor, 1968, p. 143) that a supersonic UFO should produce a sonic boom. This is certainly true of every supersonic object that man has constructed. But we should not assume that a more advanced civilization could not find some way of traveling at supersonic speeds without producing a sonic boom. Petit (1986) has paid special attention to this aspect of UFO reports and has proposed a procedure involving magnetohydrodynamic processes whereby the shock wave of a supersonic object would be suppressed.

Although it is simple to state this requirement concerning data reduction, it is by no means simple to put it into effect. It may, indeed, be necessary to proceed by trial and error: whenever one runs into an impasse, a situation in which it is impossible to reconcile the established data with any explicitly considered hypothesis (including that of ETH), one may need to review the