P02-141 - Examining Multidimensionality of the Referential Thinking Scale

Abstract

Previously, we showed the usefulness of the REF scale to assess referential thinking (Rodríguez-Testal et al., 2001; 2009) although it isn’t specific for patients with psychotic disorders (Rodríguez-Testal et al., 2008).

Objectives

This instrumental work aims to replicate the exploratory factor analysis about the Referential Thinking Scale (REF scale) already developed by Lenzenweger et al. (1997) to examine its multidimensionality.

Methods

Participants: The analyzed sample consisted of 193 participants (67.36% women, mean 28.36 years old, SD = 10.35), of whom 131 were patients.

Design, materials and procedure: We used the REF-scale (Lenzenweger et al., 1997) adapted to Spanish language. This questionnaire consists of 34 items that assess the frequency of referential thinking on a dichotomic scale (true/false). We used SPSS 15.0 to conduct a principal-components factor analysis with a varimax and oblimin rotation.

Results

The principal-components factor analysis method led to 5 factors that explain 37.35% of variance for the rotated solution. Because of inter-factors correlations are small, we considered these factors as being independent. The five factors were labeled as: Laughter, Commentaries (it accounted for 8.92% of variance); Guilt (it accounted for 8.77% of variance); Causal Explanations (it accounted for 7.17% of variance); Songs, Newspapers, Books (it accounted for 6.44% of variance); and Attention, Appearance (it accounted for 6.04% of variance).

Conclusions

It's obtained the five factors isolated in previous studies (Lenzenweger et al., 1997; Rodríguez-Testal et al., 2001). However, the multidimensionality of the REF scale must be viewed with caution because of a small percentage of explained variance.