Resumen

GEMPP FUENTEALBA, Rene  y  CUESTA SAFNRIO, Sergio. Conditional Standards of the Error of Measurement for the Chilean Norms of the EPQ-R: Applying a Binomial Model to a Personality Test. Ter Psicol [online]. 2007, vol.25, n.1, pp.51-62. ISSN 0718-4808.  http://dx.doi.org/10.4067/S0718-48082007000100004.

The Standard Error of Measurement (SEM) is an index of precision of an examinee's test score. The SEM, however, is not constant throughout the full range of scale scores. Both theoretical and empirical studies indicate that the SEM is smaller for scores near the extremes of the scoring scale and larger near the middle of the scale. A value of the SEM that applies to a specific score level is referred to as Conditional Standard Error of Measurement (SEM_COND). This paper uses a Beta Binomial Model (Lord, 1964, 1965) to estimate the SEM_COND for the Chilean norms of the EPQ-R (N= 1666), developed by Kaplan and Liberman (1992). It is concluded that the SEM_COND provides more psychometric insight about the EPQ-R scales than the traditional SEM. In addition, the results show the validity of the Beta Binomial Model as a psychometric framework for analyzing a personality test. Recommendations are given for the practical use of SEM_COND

Palabras clave : Conditional Standard Error of Measurement; Standard Error of Measurement; Beta Binomial Model; EPQ-R; Personality.