Some Problems to Evaluate Deja Experience

The research evaluating the deja vu experience has been plagued with a number of problems. First, there is often a failure to report some of the basic details of the research, such as the actual survey question used (Buck, 1970; Buck & Geers, 1967; M. A. Harper, 1969; Myers & Grant, 1972) or the incidence rate found through this query (J. C. Dixon, 1963; Osborn, 1884).

For example, for one of the more thorough deja vu questionnaires, Sno et al. (1994) provided considerable detail on the test–retest and interitem reliability, as well as on face and construct validity, but failed to publish the actual questionnaire or any descriptive statistics from four groups of individuals (people with schizophrenia, people with depression, those prone to complex seizures, and controls).

Unrepresentative samples are frequently used in deja vu research. In many instances, a sample of convenience is used. Leeds (1944) queried individuals in a shop where he worked, Zuger (1966) surveyed patients undergoing therapy from him, and Kohr (1980) polled members of the Association for Research and Enlightenment, which studies paranormal experiences and of which he was also a member.

Osborn (1884) distributed his survey “at Princeton and elsewhere” (p. 478) without further clarification, and several studies used hospital patients as representative controls to compare with psychiatric subgroups (Chapman & Mensh, 1951; Richardson & Winokur, 1967). M. A. Harper (1969) sampled workers in a public health department, but to his credit he stated that “no attempt was made to make this a representative sample of

the normal population” (p. 69).

Finally, Gaynard (1992) surveyed students attending a university where he taught a popular course called “Aspects of the Paranormal,” with many students aware of his positive bias toward this subject.