REFERENCE MANUAL FOR EYSENCK'S PERSONALITY QUESTIONNAIRE-R (EPQ-R) by By HJ. Eysenck Scanned with CamScannerEYSENCK'S PERSONALITY QUESTIONNAIRE ' The Eysenck Personality Questionnaire is the result of many years of developmental work. It was designed to give. rough and ready measure of three important personality dimensions: Paychoticism; Extraversion, and Neuroticism. Each of these three traits are measured by means of 100 questions, carefully selected after lengthy item analysis and factor analysis, The earlier history of the development of inventories for the purpose of measuring these three traits have been reviewed in The Structure of Human Personality, and the detailed description of the actual derivation of the scale here presented has also been Published. The theoretical background and the experimental validation of the concepts of Psychoticism, neuroticism, and extraversion have been given in a series of books. It must suffice here to say that Psychoticism is an independent dimension which describes the personality as solitary, troublesome, cruel, lacking in feeling and empathy, hostile to others, sensation seeking, and liking odd and unusual things. Neurotielsm refers to the general emotional liability of a person, his emotional over-responsiveness and his Hiability.to neurotic breakdown under stress. Extraversion as opposed to introversion, tefers to the out-going, uninhibited, sociable proclivities of a person. These three dimensions are conceived of as being quite independent; thus all the theoretically possible combinations of scores may in fact be observed, : The final version of the EPQ is described in considerable detail in the book titled {Psychoticism: A Dimension of Personality” (Eysenck and Eysenck, 1975); most of-the data here given are taken from that book. However, accumulation of data has continued beyond the point where the book was written, and consequently we have on occasion, given means, * Standard deviations and other statistics based on larger samples than those referred to in the book. Our collection of samples has been much simplified by our finding (to be discussed Presently) that social class has relatively little influence on scores on the scale used, age and Sex are clearly much more important, as well as being much easier to ascerta The relative lack of portance of social class makes the construction of reasonable sample much easier, and if our claim be accepted that class is of marginal influence only, then our samples may be regarded as reasonably representative. Even so, we have always d to include members of all social class. We do not claim to have used truly random or representative samples, but we have avoided the frequent error of test makers of using students as representatives of non-psychiatric population; the great majority of oureybjects are in fact not students, but wage earners, housewives, and other ‘normal’ groups, The final version of adult EPQ may prove interesting to the r loadings for each Item used in previous studies against the factor loadings of the same item obtained in previous analyses; this is « more convincing argument for factor immutability than statistical estimates of standard errors, or other r manipulations of data from a single application. Of equal interest, of course, is the comparison of data from the male and femalé samples; we have looked for consistency in this comparison before accepting items into our final questionnaire. fj der to check our fettor Table | shows the loadings of all the items on four factors extracted and.rotated into oblique simpje structure. Also given are the letters indicating the scales in which the items appear in scoring key. The population on which the analysis was carried out consisted of 1,796 males and 2,565 females; this is an unusually large number, but we believe that many factor analyses reach inconclusive results because of insufficient aumbers of subjects Comparison with previous analyses, on smaller number of subjects, and using somewhat different sets of items, indicated that by and large items have retained their loadings with considerable accuracy. aw Scanned with CamScanner‘Table 11 Loading of items on four 1861078 (a , 2,565 mn me J as +t “AD 1 Ez 10 152 a 06 04 * 59 08 . 203.215 prvezumemeacgemceac sway CSL OS J sme gum umzm eZee Ze Scanned with CamScannerSSE: n= 1,796 n=2,565 li E N L P E N L 08 59 -.03 01 “lS 23 05 00 “12 34 08 -.08 313 24 -.40 +36 L E P N L E Pp N L E P N L E P N P N L E P N L P N P N L P N L E ~ By N L E P N L Pp be seen that these are Table 2 shows the inter-correlation between the factors; it rather low, with the exception (not unexpected) of L, which correlates negatively with P and ‘These correlations are sufficiently low to indicate that the respondents fil questionnaires with a minimum of dissimulation. There is a slight correlation (nej between E and N, too slight to matter substantively, and an even slighter one between WM (positive). All in all, the factors are-as near orthogonal as could have been hoped. Scanned with CamScannerNORMS le or2,312 male and 3,262 female subjects, and it MAY be of interest to consider the distribution ofthe scores. The reason for this interest simply thet these provide an aad aculos demonstration of the observed form of distribution; it would be So eetaderotine the shape of the observed distributions as being necessarily indicative of the "Whe intron ofthe underlying traits. The observed distribution of scores depends on such factors ie jer-coreltion, and many other such factors which might not be replicated in another scale aaa ve different items. Ideally, of course, one would select items in such a way that normality of distribution could be obtained, but this is less important than the other criteria we have had in mind for i je selection of items, and consequently we have not used score distribution in considering items for inclusion, Seer a erty we ably those for P, deviate markedly from normality, being positively or negatively skewed. There would be little point in providing a more accurate statistical description of the shape of the curves in terms of higher moments, and we have not done s0- Scored forms are In many research applications the examiner will have no need to convert raw Sere obtained with the scoring key stencil to sten score. However, commonly the test user ‘would like to know how aie shal stands in relation toa defined population. For this purpose the raw scores should be Converted into sten scores. Use Table 9a and 9 to convert raw scores to sten scores. The term sien caoverfrom “standard ten, The sten scores are distributed over ten equal intervals of standard scores point from 1 to 10. ‘Table 9a: Conversion of raw scores into sten scores (N=3262) (Females) Sten P E 1 - 1-2 1-2 - 2 : 3.5 3-4 1 3 - 6-7, 3-7 23 4 1 8-10 8-10 45 s 2 11-12, 112 6-7 6 3 13-15 13-15 8-9 ~ 7. 4 16-17, 16-17 10-11 ™“ 8 5-6 18:19 18-20 12-13, 9 en?) 20-21 21-23 14-16 19 came ae : : 17-21 ‘Mean 2.63 12.60 12,74 173 S.D. 2.36 4.83 AD) nen 888s SS For fer ‘or females, if the raw score on Lie seale exceeds 8, reject the scores. 10) Scanned with CamScanner9b Conversion of raw scorer Into aten scores (N=2312) 1 | 1 5 1 Ps 5 2 . 5-6 1-2 - 3 : 18 3-4 4-2 . s 12 9-1 $7 364 8 3 1213 89 56 \ 6 4-5 14-16 10-12 7-8 . u 6-7 17-18 13-14 9-10 . \ 8 19.20 18-17 1-13 9 9-10 a 18-19 14-15 10 11-25, . 20-23 16-21 Mean 3.78 13.19 9.83 6.80 S.D 3.09 491 SAB Ala | For males, If raw acore on Lie acale exceeds 7, reject The score of 5-6 denote avert express gradually the gr decrease of strength. rage strength of the factor, scores above 6, i.e, from 7-10, strength of the factor and score below 5, l.e. from 1-4, indicate gradual In Table 10 conversion of stons to centile is presented, Please note that the ordinary use of this table will be for whole number stens only. There are marked in bolt prift. The stens to a decimal place are given only for researchers who may want to work with sten values which arg th@-means of the’ groups. For individuals, however, only whole nlimber stens are used. ; FE eee Sten Centite Sten Centile Sten Centite arr 40 22.7 1.0 14 4d 24.2 mW 16 42 25.8 12 18 43 214 13 2.0 44 29.1 74 ! 23 45 30.8 18 84.1 2.6 46 32.6 16 85.3 2.9 4.7 34.8 17 6.4 32 48 36.3 18 87.5 3.6 49 38.2 19 88.5 ; 4.0 50 404 8.0 89.4 ; 4s 5.1 ant 90.3 5.0 5.2 44.0 2 Scanned with CamScannerSten 23 24 25 26 27 28 29 3.0 31 32 33 34 3.5 3.6 37 38 3.9 Centile 5.5 6.0 67 14 81 88 97 10.6 1s 12.5 13.6 14.7 15.9 WA 18.4 19.8 21.2 Sten $3 34 5.8 5.6 $7 58 5.9 6.0 6.1 6.2 63 6.4 6.5 6.6 67 68 69 Centile 46.0 48.0 50.0 52.0 54.0 56.0 $7.9 59.9 61.8 63.7 65.5 67.4 69.2 70.9 12.6 14.2 75.8 Sten Contin 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9.0 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 10.0 Scanned with CamScanner 919 92.6 93,3 93.9 94.5 95.0 95.5 96.0 96.4 96.8 97.1 974 977 98.0 98.2 98.4 98.6 983(L) Scoring Key of EP Scale Mode of Response Serial Number of Item Score Psychoticism 2,6,9,11,18,53,57,61,71,90 22,26,30,33,43,46,50,65, 67,74,76,79,83,87 Neuroticism NIL 3,7,12,15,19,23,27,31,34, 37,38,41,47,54,58,62, 66,68,72,75,77,80,84,88 Extraversion LieScale 21,29,42 1,5,10,14,17,25,32,36,40, 45,49,52,56,60,64,70,82,86 4,8,16,24,28,39,44, 48,51,59,63,69,73, 81,85 13,20,35,55,78,89 Scanned with CamScanner