The factor structure of the Paranormal Belief Scale: More evidence in support of the Oblique Five

Author(s): DeCicco, P.

Journal/Book: J Parapsychol. 1997; 61: PO Box 6847, College Station, Durham, NC 27708. Parapsychology Press. 243-251.

Abstract: Using confirmatory factor analysis we sought to confirm the findings from a recent test of two competing models of Tobacyk's Paranormal Belief Scale (PBS) (Tobacyk & Milford, 1983; Tobacyk, 1988): the Big Orthogonal Seven and the Oblique Five (Lawrence, Roe, & Williams, 1997). 225 students at Coventry University were administered Tobacyk's (1988) Revised Paranormal Belief Scale. Questionnaire responses were analyzed using the specialized statistical package EQS (Bentler, 1989). The first test of Tobacyk and Milford's Big Seven model showed that once again the model provided a very poor fit to our sample data. The second test, an oblique five-factor model based upon statistical criticisms outlined by Lawrence (1995a), also showed that the Oblique Five provided a far better fit to our sample correlation matrix, although that model still did not provide an acceptable fit to the sample data. A test of the difference in fit between the two models replicated Lawrence et al.'s (1997) previous finding that the Oblique Five were a significantly better fit than the Orthogonal Seven (P <.00001). Once again, a common theme for post hoc model improvements suggested by EQS was the freeing up of the correlations between factors. We conclude that the evidence for the relative superiority of the ''oblique five'' factor structure is overwhelming, and suggest that future use of the scale be guided by our five-factor conception of the PBS, rather than the incumbent seven-factor orthogonal structure.

Note: Article Lawrence TR, Coventry Polytech, Sch Hlth & Social Sci, Priory St, Coventry CV1 5FB, W Midlands, ENGLAND

Zurück | Weiter