The Revised Junior Eysenck Personality Questionnaire (JEPQ-R): Dutch replications of the full-length, short, and abbreviated forms

https://doi.org/10.1016/S0191-8869(00)00166-5Get rights and content

Abstract

This study examines the full-length, short and abbreviated forms of the Revised Junior Eysenck Personality Questionnaire (JEPQ-R) in a Dutch sample of 215 boys and 207 girls, aged 12–14. The reliability and concurrent validity of the scales of the full-length form (JEPQ-R, 81 items), short form (JEPQR-S, 48 items) and abbreviated form (JEPQR-A, 24 items) are satisfactory. Results indicate that the short and abbreviated forms can be recommended as functional equivalents to the 89-item full-length JEPQ-R when the research context does not permit the application of the longer instrument. The need for further examination and development of the short and abbreviated Psychoticism scale is emphasized and other recommendations for future research are given.

Introduction

The Junior version of the Eysenck Personality Questionnaire (JEPQ; Eysenck & Eysenck, 1975) has frequently been used to measure the basic dimensions of personality, as proposed by Eysenck, 1947, Eysenck, 1952. This instrument contains P, E, and N scales that measure Psychoticism, Extraversion, and Neuroticism, respectively, as well as a Lie-scale (L), thought to reflect dissimulation. Because the P scale of the JEPQ showed psychometric weaknesses, Corulla (1990) revised the scale in order to improve the low reliability and to increase the low range of scoring. Indeed, in the revised JEPQ (JEPQ-R) the reliabilities for P were satisfactory, ranging from 0.75 to 0.82 for boys and from 0.68 to 0.77 for girls, aged 11–15 years. The range of scoring also increased considerably, with means of 7.68 for boys and 5.17 for girls instead of around 3.0 and 2.0 as reported by Eysenck and Eysenck.

In a Dutch replication study, De Bruyn, Delsing and Welten (1995) found factor patterns of E, N, and L that were similar to the factor patterns reported by Corulla (1990). Differences were found for P, however, indicated by a Tucker's φ congruence coefficient of only 0.66. The difference between P in both studies was reflected by the fact that in Corulla’s study, 75% of the P items had their highest loading on P, whereas this was true only for less than half of the P items reported by De Bruyn et al. Not surprisingly, the reliability of the P scale in the latter study was only moderate (i.e. 0.57).

The JEPQ-R is a relatively long test, containing 89 items. One of the practical disadvantages of a long test is that because of time constraint, research situations may permit the use of only a small number of items. In recognition of this fact, Corulla (1990) developed a short form of the JEPQ-R (JEPQR-S), consisting of a total of 48 items, 12 items for each of the four scales. Importantly, the reliabilities of the short scales were acceptable and were comparable to those for the full-length scales. In addition, because of the high correlations between the short scales and their corresponding full-length scales, the short form seemed to tap much of the same information as the full-length form. Since Corulla was the first to examine the JEPQR-S and not many replications have been made until now, more research seems warranted to investigate the validity of this short form.

Because even a 48-item instrument may be too long in specific research situations, Francis (1996) developed an abbreviated form of the short form JEPQ-R, which was labelled JEPQR-A. Using JEPQR-S data, obtained in a sample of 1597 13–15 year olds as a starting point, he decreased the number of items for each scale from 12 to 6, resulting in a 24-item instrument. Item selection was based on two criteria. First, items were selected on the basis of satisfactory item — rest correlations. Second, items had to represent the domain of items included within the JEPQR-S parent scale. The resulting four scales had reliabilities that, although being lower than the corresponding JEPQR-S scales, were quite reasonable considering the fact that only six items were included in each scale. The only exception was the L-scale, which had a reliability of 0.57. Francis concluded that the reliability and validity of the JEPQR-A would permit its use, but that further examination among other age groups and in other cultures was recommended.

Since De Bruyn et al. (1995) studied the factor structure of the JEPQR on a sample of 260 children, their sample did not allow for an examination of the factor scales for boys and girls separately. The aim of the present study is first to replicate and extend De Bruyn et al.’s study of the full-length JEPQ-R in The Netherlands. We will examine the full-length JEPQ-R for the total sample under study as well as for boys and girls separately, and compare the results with De Bruyn et al. and Corulla (1990) when possible. Second, we will replicate Corulla's analyses concerning the JEPQR-S, and third, we will examine the validity of the JEPQR-A in The Netherlands and compare the results with the findings reported by Francis (1996).

Section snippets

Participants

Participants were 422 Dutch adolescents (215 male, 207 female), ranging in age from 12 to 14 years (M=12.8 years), attending secondary school classes. They completed a Dutch version of the 89-item JEPQ-R (De Bruyn et al., 1995). Assessment took place during regular school hours and participation was voluntary.

Measures

The Dutch version of the JEPQ-R (Corulla, 1990, De Bruyn et al., 1995) is a questionnaire that consists of 89 dichotomous items referring to habits or behavioural tendencies. The

Full-length JEPQ-R

Principal component analysis revealed a great number of factors with Eigenvalues greater than 1. As can be seen in Fig. 1, the scree plot (Cattell, 1966) indicated that four factors might be retained for rotation. Separate analyses for boys and girls revealed that the factor patterns for E, N, and L were basically the same for these two subsamples. Table 1 shows that there were high correspondences between boys and girls for E, N, and L, with Tucker’s φ of 0.80, 0.86, and 0.83, respectively.

Discussion

The aim of the present study was to evaluate the Dutch versions of the three forms of the JEPQ-R. These forms were the full-length JEPQ-R, the JEPQR-S (Corulla, 1990), and the JEPQR-A (Francis, 1996).

References (12)

There are more references available in the full text version of this article.

Cited by (0)

View full text