Se espe 2 on clin ye MANUAL FOR &M avngiry ae ®P ursonauiry TIwvenrory By. HLJ.EYSENCK Professor of Psychology, University of London Psychologist Maudsley and Bethlem Royal Hospitals Copyright 1959 by H. J. Eysenck, Reprinted in India, 1993aA MAUDSLEY PERSONALITY iSVENTORY. "The Maudsley Personality Inventory (M PI.) 1s ne result of many years of developmental work. it wes Je- signed to give a rough-and-ready measure o* two impor- tant personality dimensions: Neuroticism, or emotionali- ty, and Extraversion. Each of these two traits are mea- sured by means of 48 questions, carefully selected after lengthy item analyses and factor analyses) The earlier history of the development of inventories’ for the pur- pose of measuring these two traits has been reviewed in THe STRUCTURE OF HUMAN PERSONALITY and the detailed description of the actual derivation of the Scales here presented has also been published earlier The theoretical background and the experimental valida- tion of the concepts of neuroticism and xtraversion have been given in a series of books (mentioned in the bibliggraphy as S.No. 4, 5, 8). It must suffice here to say that neuroticism refers to the general emotional stability fa person, his emotional over responsiveness, and his Stability fo neurotic breakdown under stress. Extraver- sion. as opposed to introversion refers to thi: out-going, uninhibited, sociable procilivities of a persc*). The two dimensions are conceived of as being qu’'> indepen- Gent. thus all the theoretically possible com inations of scores may in fact be observed. The M.P.L. is much shorter than most other ques- tionnaires; nevertheless, for certain purposes even 48 questions may be too many to ask. Consequently, a short form of the M.P.I. has been prepared; this consists of two scales of six iteris each, taken from long form of the scales. On the printed form of the Questionnaire, these 12 questions making up the short ‘orm of the M.P.l. have been given on the first page. immediately after the instructions; thus if only the short form is to be given, only the first page will have to be filled in by the./ a subject. 1 the,{ull scale is to be administered, ‘both fides of the sheet will of course have to be filled in by the subject The M.P.1. has been administered to large num- ber of subjects, both normal and neurotic. The original sample of 200 men and women on whom item analyses and factor analyses were carried out has since been supplemented by groups’ of students; Industrial appren- tices, nurses, and a quota sample of the whole popula- tion; also available are data on various neurotic groups, on prisoners, and on sufferers from psychosomatic dis- orders. A detailed survey of many of these results, mest of them unpublished, has been given by A. Jensen. As standardization data, only on the normal quota sample has been used, and the various abnormal groups set out in Table |. In addition, large numbers of American stu- dents have been tested by A.W. Bending, of the Univer- sity of Pittsburgh, and by various other American Inves- tigators; details of these researches will in due course be published in American Journals, but would not.be included here, except insofar as cited in Table |. | am rdebted to many colleagues for contributing data; some of their names will be found elsewhere. \ 4 _ RELIABILITY OF SCALES Both split-half and Kuder-Richardson reliability coefficients have been calculated on many samples. For the Nevroticism scale, these values nearly all lie be- tween .65, and .90; for the Extraversion scale, they ie between .75 and .85 with the majority above .80. Retest reliabilities are available only on,about 100 cases; they are .83 and .81 respectively. Ag regards the short scaie, split-haif reliabilities on a quota, sample of 2,000 man and women were .80 and .72. ‘ 2 . oe 7 podastes [ “PScoRRELATIONS OF THE SCALES WITH OTHER INVENTORIES The correlations between the long and the short M.P.1, scales are .86 and .87 respectively for N and E The Heron neuroticism and sociability scales correlate 64 and .80'with the corresponding M.P.l. scales; the Cattell's neuroticism and introversion scales. correlate 34 and .53 (forms A and B), and .66 and .67 (forms C and D) with the corresponding M.P.I, scales. (The corre- lation were derived from 134 neurotics). The. Taylor Manifest Anxiety Scale correlates .77 with N and -.35 with E on 254 American students. Correlation with the Guilford’ and R scales were .92 and .79 respectively in the original item analysis group of 400. Correlations with the Minnesota T.S.E. scales on 87 students and student nurses were as follows: N with T 04; NwithS = (33; 5 NwithE’ = 17; a Ewith T = -.05; EwithS = .81; EwithE = -.08. In summary: N correlates with the Heron end Cattell neuroticism scales, the Guilford C scale, and the Taylor scale. E correlates with the leron and Catieli extraversion scales, the Guilford R scale, the Minnesota ility scale, and negatively with the Taylor scale“RCORRELATION BETWEEN. NEUROTICISM AND 4 EXTRAVERSION Theoretically, the two dimedsions are indepen- gent, and consequently zero correlations would be ex- pected between them. On some 20 different normal samples in both England and America the actual correla- tion has been in the neighbourhood of -.15; occasionally some samples give positive correlations and others give higher negative correlations. There appears to be no doubt that while the scales are almost entirely indepen- dent, yet there Is a very slight negative correlation in most samples. For the short scale, a correlation of -.05 was found in a quota sample of 1,600 men and women. In neurotic, psychosomatic and prisoner groups howev- er, the correlation rises to between -.30 and -.40; similar correlations can be found in normal groups when only subjects with high neuroticism are included. The cause of this phenomenon lies in the non-linearity of the re- gression lines at high introversion scores: this study based on a ample of 1,200 normal subjects. The theo- retical explanation of this phenomenon lies probably in a positive feed-back effect of the high conditioning poten- tial believed to underlie introversion, and the high auto- nomic reactivity believed to underlie neuroticism. What- ever the cause, care should be taken in experiments concerned with the correlates of extraversion-introver- sion so that the extreme groups having very high and very low scores respectively on E are matched on neu- roticism scores; unless careful matching is carried out in the group with high E. The scores are likely to have lower N scores than the low E group. STANDARDIZATION DATA Table 1 gives the most relevant standardization data. The groups employed have already been dis- cussed briefly; it should be added that the hospital groups of neurotics were diagnosed by experienced psy- chiatrists or else had their case-papers carefully scruti- nized by three experienced clinical psychologists who artived at a unanimous diagnosis independently, The number of cases in sub-groups are not large, but suc- cessive small samples from different hospitais showed great stability in means and variances TABLE - 4 STANDARDIZATION GROUPS Description of Group Size Nmecn Nsd Emean sd Normals (English) Quota Sample 1,800 19.09 1102 2491 9.71 Normals (American students) «1,600 20.9 1069 2859 428 Dysthymics (Hospital Patients) 84.18 10.84 177.86 1002 Prisoners (Recidvists) 146 30.35 1073 2409 9.11 Hystercs (Hospital Patienis) 58 300? 1184 2691 9.26 Psychosomatics (Hospital Patients) 108 35.69 1089 2538 923 Psychopaihs (Hospital Patients) «35.35.58 1091 30.77 9.51 ‘SHORT SCALE Normats (English): Quota Sample 1,600 615 333 7.96 297VALIDATION OF SCALES In a sense, the data in Table 1, may serve as validation of the scales, because the general theory un- derlying the consfruction of the scales demands that dysthmic neurotics, i.e. those suffering from anxiety, re- active, depression, obsessional/compulsive symptoms, phobias, etc. should have high scores on neuroticism and introversion, while psychopaths and hysterics should have high Score on neuroticism and extraversion. Indeed, hysterics anc: psychopaths are found to be sig- nificantly more extraverted than the dysthymics, and all neurotic groups are higher on neuroticism than are the normal groups. The hysterics are no more extraverted than the normals; this finding is not quite in line with expectation, but has been repeated on several samples and must be accepted. The position of the recidivist prisoners and the psychosomatic group is close to the hysterics, a finding which is not unexpected. ” An alternative, and perhaps a superior method of validation involves the concept of construct validity, i.e. the validation of a given measure by its incorporation within a scientific system or nomological network. This has been attempted in the development of the M.P:I., and the results have been reported in The Dynamics Of Anxiety and Hysteria”on the whole, verification of pre~ dictions has been forth-coming in a surprisingly large number of instances. This work is still being continued, and reports of it are periodically appearing in the rele- Vant English and American Journals. ~ AGE, SEX AND CLASS DIFFERENCES White differences due to these factors have been found to exist, they are too slight to warrant the prepara- tion of separate norms. However, for general guidance the following facts may be noted. Women have higher scores on neuroticism by about one third S.D.; lower class and younger age group also tend to have higher score on N by quarter S.D. and one-fifth S.D. respec- tively. On E, men have been found slightly more extra- verted than women, but by only one-sixth S.D. interac- tions are insignificant throughout. Among U.S. student groups also, women have been found slightly less stable on N by one-sixth S.0. All these differences are statisti- cally significant, but with the possible exception of sex differences on N, are not of a size to make them psy- chologically significant. INSTRUCTIONS FOR THE USE OF THE SCALE Instructions for literate subjects are printed on each copy of the M.P.I.; these should be read aloud to groups of subjects, or be read silently by subjects tested Individually. They should not be amplified or alerted in ‘any way. When the questionnaires are collected after ‘completion, care should be taken to check that all ques- tions have been answered; where answers are missing subjects should have their attention drawn to the omis- sions. With illiterate or blind subjects the questions may be read aloud and the answers recorded. This should never be done when anyone but the subject is present, and on no account must the examiner change the word- ing of the question, amplify or interpret it, or give advice to the subject on how to answer it. Completed question- naires should be scrutinized for the number of "7" re-sponses; this number should be recorded on the front page. if this number exceeds 10, results are of very doubtful value; the subject may be retested, and part of tne instructions relating to use of “?" responses brought to his special attention, SCORING he inventory Is scored directly from the test forms. It should be made clear to the subjects that their answers must be marked inside the. boxes. If itis marked outside or above the boxes, proper corrections should be made before scoring. The test forms should be scanned to ascertain that only one answer has been marked for each question. The scoring sten- cil should now be placed along the anchoring points made on the test form. Tha raw scores may be checked and added to give a total of iN and E scores respectively. It may be noted that all the items answerable in category “yes” are assigned with a wattage of 2:scores.(except for item nos. - 14, 16, 18, 22, 24, 30, 36 and 40 - which belong to extraversion dimen- sion and reassigned 2 scores only when answered in category 'no"). All the “"2" responses are assigned with one score when the subject marks them in the middle category. The raw scores are thei converted into standard scores to compare the degree of neuroticism and extraversion with the norms. One translucent scoring key is available , for each of the two scales. The instructions to use this key are given on the key itgelf. Score page-1 first, then the second page. and add the scores. Make sure the key and the questionnaire which is being scored are properly aligned; the numbers in front of the questions on the Inventory should correspond with the numbers of the keen MEANING AND USE OF NORM TABLES (it one is to'ind out just where the subject stands in relation to other members of population, the raw scores will tell nothing about this, until converted into normative scores which indicate that proportion of the population will get a higher or lower score. To find out a standard score, it is always assumed that the mean of a set of sigma score or Z values is always equal to zero and the standard deviation is equal to 1. However, since half of the scores in a distribution will lie below and half above the mean; about half of our sigma scores will be negative and half positive. For these reasons, sigma Scores are equally converted in a new distribution with M and S.D. so selected so as to make all scores positive and relatively easy to handle. ‘Thus, a standard score is simply a linear transformation of a given measurement X to a new scale. Therefore, the transformation does not change the shape of the distribution in any way. The Raw Scores of N and E dimensions of M.P.l.'s long and short scales have thus been expressed as Standard Scores in a distribution of M=50 and.SD=10, which are given in Table 2 and 3. 7 Table 2: Norms for Hindi Version of M.P.I. Scale Raw Score M.P.I. Dimensions of Long Scale on the Standard Score Dimensions Neuroticism Extraversion 1 28 6 2 29 8 3 30 10 4 31 42 5 32 13 6 33 15 7 34 17 8 35 18 9 36 19 10 37 21° "W 38 23 9Rew Score M.P.1. Dimensions of Long Scale on the Standard Score Dimensions Neuroticism Extraversion 12 39 24 13 40 26 14 41 28 15 42 29 16 43 31 17 44 32 18 45 34 19 48 38 20 47 37 21 48 39 22 49 at 23 50 42 24 51 44 25 52 45 26 83 47 27 54 49 28 55 50 29 56 52 30 37 53 + 31 58 55 32 59 57 33 60 58 34 61 60 35 62 62 36 63 63 37 64 65 38 65 68 39 68 68 40 87 70 41 68 71 4 These norms are based on-6250 school/college students and mental hospitals and various mental health clinics. The age range of sample »eing 14 to 30 years with a mean age of 21.5 years. Norms for males and females have been combined. Table 2: Norms for Hindi Version of M.P.t. Scale Raw Score M.P.1. Dimensions of Short Scale on the Standard Score Dimensions Neuroticism Extraversion 4 30 21 2 34 25 3 37 29 4 40 sae 5 43 37 6 48 ai 7 50 45 8 53 49 9 56. 53 10 so 57 11 63 61 12 6 65 Raw Score Mean 7.202 8.312 Raw Score SD 3.214 2.492 Based on 4250 college students with a mean age of 21 years. Norms for the male and female are the same aINTERPRETATION OF SCORES Now, one can easily read these tables from the left after placing the obtained raw scores. The standard score value equivalent to that raw score will be found to the right in that row of the table relating to N and E dimensions. Thus, one may easily’ interpret that a standard score of 50 is equg! to the average. A difference of 10 standard scores is''+ or ‘-" 1 sigma from the average and is not significant. But a standard score of 70 or 30 needs attention. Similarly, a standard score of above 70 or below 30 indicates a very considerable deviation from the average. MEANING AND USE OF STEN SCORES . Another method of treating the raw scores with the standard scores to determine the individual's posi- tion in a redefined population is use of STEN scores. Since’ measurement in psychology is usually not on an absolute scale, the best that most measures can give is the relative standing of one individual with another, or with a group of individuals. ‘Two general approaches to examining these rela- tionships are ava'“ible, and in both cases, the raw scores typically unsergo certain transformations. In the first approach, the fast user transforms the raw scores for the group of individuals he has tested into some sort of standard score or centile, and then makes compari- sons among the individuals in that group, or makes them between this group end another that he may have previ- ously tested. 12 In the second approach, the jest user transforms the raw scores again, to some type of standard score. but this time he compares the individual, or infact, the whole. group, to a population whose typical response pattern is known, In both cases, the question is “now jual stands compared to others in the population?" In the former case, the population is quite specific to the aims of the test user. For example, he may have collected his data from management trainees, and may wish to compare the ability profiles of each sex. In the latter case, the population is one whose abili- ties have been investigated and reilability measured to the degree that it might be said with confidence, that a particular level of some ability either is, or is not charac teristic of that population. An exampie would be the pop- ulation of all boys of high and senior school level ‘Norms’, however, need not necessarily refer to ‘normal population, but only to population about which some minimum degree of knowledge is available, Obviously. in the early stages of a test's use most commonly avail- able norms are for the general 'normai’ population TRANSFORMATION OF RAW SCORES ‘The type of standard score used with thig is the STEN - a unit on a standard ten-point scale. Although sten will appeal to most users, it is important that each investigator have access to the kind"of standardized score he desires. The values of raw score means and standard deviations (from which sten can be derived) are given in a tabular form in the following pages. ctions for obtaining the raw Scores from the test form was discussed earlier. Here, we shall assume that the raw scores have already been obtained and that sten scores are desired. Standardization tables are giv- en for short and long scales of M.P.! 13The size =f the samples and the means and stan- dard deviations or raw scores are given in each table. The sample in the school students group consists of boys and girls of ninth standard to twelfth standard, within the agé range of 14-17 years. The sample in the college students group consists of males and females undergoing their graduation or post-graduation courses, within the age range of 18-22 years. No separate tables for classwise distribution have been constructed for any group of norms. Finally, sample in the general adult pop- ulation consists of mainly the occupational groups, stu- dents and housewives, within the age range of 22-30 years. n Once the test administrator has decided which ta~ ble (long or short scale) will be most appropriate for his/ her data, the use of the norms table is quite easy. The values within the table (i.e., the body of the table) are "raw scores" - the values obtained from the scoring key. In summary, the procedure for obtaining sten scores are: 1. Obtain raw scores from the test form using scoring key. 2. Select the appropriate norm table (long or short). 3. Convert each raw score in the table to sten score, written at the top row or the bottom of the table. The two norm tables for long and short scales are given in the following pages. 14 wean SD 10 Sten Score: 28 29-31 32-34 36-87 G8-40) 41-48 28.08 6.31 03 48 9413 14-18 19,29 24:26 20-93 (5498) 99-43 44-48 29.29 10.024 Norms for General Population (Combined) for Long Sc 1-21 22-24 25+ 0-45 16-18 19. Extraversion. Gat Sten Score 5 Neuroticiem 15 Table-S: Norms for. General Population. (Combined) for. Mean SO 10 1 Dimension > Sten Score 2.49 831 492 6-7 40 03 Extraversion 3.21 7.20 67. 8 O40 14-12 45 02 Neuroticism 10 6 Sten Score 5USES OF THE SCALES The scales have been used for several purposes, which however are not likely to exhaust their usefulness. 1) Experimental - Most of the phenomena in ex- perimental psychology, such as conditioning, figural af- ter-effects, vigilance, the constancy phenomena, massed practice work curves, reminiscence, P.G.R. and other autonomic responses, have personality differenc- es as important parameters; their discovery, measure- ment, and control are facilitated by the availability of readily administered inventories of the kind here pre~ sented. 2). Clinical - As routine measures, the administra- tion of the M.P.1. for both outpatients and inpatients may be useful in facilitating diagnosis, drawing attention to certain areas for further inquiry, and save time by pre- senting the psychiatrist with a condensed picture of the patient's personality as seen by himself. 3) Educational - Research on the success and failure of University students by D.W. Furneaux has drawn attention fo the important part which is played by neuroticism and extraversion in this connection, and em- phasizes the possibilities.of using simple questionnaire measures for the purpose of guidance and selection. 4) Market Research - In recent years, motivational research has become of increasing concer to organiza- tions concerned with publications to certain products, advertising campaigns, etc. Such work can be consider- ably facilitated by having available not only the views expressed by samples of the population, but also mea- sures of their personality dimensions. 5) Selections Problems - In conjunction with the measurement of a person's ability, measurement of his A 16 personality can be of considerablé'assistance in coming to decisions in problems of selection or allocation. The possibility of conscious falsification of answers probably makes the use of the N scale relatively useless here but the E scale is less likely to be sdvaffected, and the aif ferent personality qualities of extraverts and introverts make it probable that they will be best adapted to quite different types of work In all its applications, the M.P.l. should primarily be regarded as research instrument. Different firms, or- ganizations, hospitals universities and other bodies have different problems, deal with the different samples of the population, and aim at different solutions of their problems. Only applied research can be determine whether instruments such as the M.P.I can be success- fully used by them, and just what forin such use can best take. Separate norms will probably have to be built up by different organizations, to mirror the particular samples of the population with which they may have to deal. Such work can only be done by properly qualified and experienced psychologists; it cannot be too strongly emphasized that psychological instruments, even when so apparently simple and straight forward as a question- naire or inventory, cannot with advantage be used by the layman in the absence of supervision by a suitably trained person. 7