THE JOURNAL OF ALIED SYCHOLOGY Vol. 7, No. 6, 19 The Measuremen of ersonaliy and Behavior Disorders by he I.. A. T. Music reference Tes Raymond B. Caell and Jean C. Anderson Laboraory of ersonaliy Assessmen and Group Behavior, Universiy of Illinois On he wide research fron which is roughly designaed by "projecive ess," bu perhaps more accuraely by "mispercepion ess" (1), few recen advances have been so promising as ha conneced wih music percepion. The powerful and immediae connecion of musical simulaion wih emoional experience, and he many indicaions ha unconscious needs gain saisfacion hrough his medium, have long poined o measures of musical preference as effecive avenues o deeper aspecs of personaliy. Moreover, he lack of verbal conen is iself, on general principles, a promise ha he verbal, cogniive defenses of he censor may be by-passed and he emoional needs probed more direcly wihou disorion by defense elaboraions. The Music reference Tes ersonaliy ess which proceed from he esheic reacions of he subjec, or from likings and dislikings which canno be based on logical, explici relaionships o he subjec's purposes and senimens, occupy an area inermediae beween ha of projecive ess and ha of oher objecive personaliy ess. For he liking or disliking is evidenly due o characerisics impored or projeced ino he physical sounds by he lisener, ye he "projecions" are no so explici as in he imagery evoked by he Rorschach or he inerpreive sories which he subjec is asked o weave around he T.A.T. I is possible, herefore, ha furher research and clinical experience wih his relaively unexplored class of ess (which may be called ess of "affecive mispercepion") will show hem o have cerain advanages over he sandard projecive or mispercepion ess. For sophisicaed subjecs inuiively realize ha heir cogniive projecions sand in need of defensive disguise, whereas heir likings and dislikings make no more sense o hem han hey do o he psychologis before his saisical analyses are made. As in all es consrucion involving "iems," i would be foolish here o design psychological measures hinging on he luck of a single rresponse and o aemp o relae such a single response o personaliy dimensions. Insead we firs seek reliabiliy in he es measuremen iself by composing i of scores on several iems, hereby diminishing he effecs of chance and specific hisorical associaions. This may be conceived as discovering he dozen or more "iems" ha can be validly added ogeher o give a score on some single dimension of emoional qualiy or musical-emoional reaciviy. Aemps o find hese groupings by inrospecion or by psychiaric judgmens mus be se aside, for hey are shown by preliminary research o be highly unreliable and o consiue an amaeurish approach o he problem. Insead i is necessary o find he dimensions of musical choice by submiing a number of musical excerps o a large populaion and correlaing he responses, hereby discovering empirically which responses "go ogeher." This firs sage of research in he area has already been carried ou by Caell and Saunders () using 120 half-minue musical excerps under condiions described elsewhere. The psychologically ineresing and reassuring hing abou his facor analysis of a marix couched in a new variey of response correlaions, namely, in music preference responses, is ha simple srucure was as definiely obained here as wih abiliy ess, and ha a comparison of wo facorizaions revealed a very graifying degree of invariance of he facors. Wih his assurance from an iniial sudy i is o be hoped ha psychologiss will be encouraged o face he vas amoun of exacing work required by his approach insead of being beguiled by merely esheic inuiions in es consrucion. 6
Measuremen of ersonaliy and Behavior Disorders 7 The wo doveailed facor analyses yielded eleven sable facors (). Bu before-hese basic findings could become a pracical foundaion upon which furher "applied" research could readily go forward, in clinics and guidance ceners generally, i was firs necessary o consruc ou of he above research findings a convenien rouine insrumen. This was done under he auspices of he Insiue for ersonaliy and Abiliy Tesing by he senior auhor and has issued in a 12-inch long-playing record, reproducing 100 half-minue music excerps (SO on one side, Form A, and an equivalen SO on he oher, Form B). Excep for he firs and he las hree facors in his es here are en iems provided o measure each facor. These iems were chosen from he 120 facorized, according o he usual es consrucion principles; a significan loading on he facor concerned; a balancing (suppression) of loadings on facors no concerned; a balancing of "like" and "dislike" responses in he score for any one facor; no use of any iem for more han one facor. A cyclical order of sampling of iems from he various facors is used in he es as finally presened. The es so consruced, when cross validaed on a new populaion, was found o have consisency (spli-half reliabiliy) and equivalence (Form A Form B) reliabiliy coefficiens (2) ha were adequae on only seven or eigh of he eleven facors. See Table 1. This inadequacy arises largely from some facors being measured on a bare minimum of or iems in one form. Accordingly i is advocaed ha only seven or eigh independen facors be rouinely measured in sandard clinical use and ha he remaining hree or four measures serve an exploraory purpose, as "locaed nuclei" from which furher research can, by exension ino new iems, build up beer facor scales. Meanwhile he es has been iniially sandardized for every facor on a normal populaion of 80 suden and non-suden aduls ranging from 18 o 68 years of age. The insrucions, which are given in sandard form by he voice on he record, are se ou below. The I..A.T. Music reference Tes of ersonaliy () is hus normally presened simply as a es of musical preferences, bu he implicaion ha we were psychologically ineresed in he resuls from he sandpoin of personaliy measuremen was realized, a leas by he normal group, in his paricular experimen. Firs Issues Needing Research Now ha such a measuring insrumen is available, a number of researches immediaely sugges hemselves, especially in applied psychology. Concerning is promise as a personaliy es i is a once apparen from inspecion of he acual musical excerps found o be highly loaded in he various facors, ha hese facors are no merely culurally-deermined groupings, corresponding o musical "schools" or periods (wih one possible excepion among he eleven facors: F 1). Wih his superficial inerpreaion rejeced we may nex examine he hypohesis ha hese facors correspond o wha have been called major "hidden premises" in he logic of personal preference (1). For hese hidden premises of choice decision, according o our hypohesis as saed elsewhere (1), should be emperamenal and early-environmen-deermined dimensions of personaliy iself. If his is correc, here should be some subsanial correlaions beween hese facors and he facors on he 16 ersonaliy Facor Quesionnaire or any oher measure of he primary personaliy facors. This a leas is he hypohesis upon which he whole of he presen invesigaion has been carried forward. If he musical choices are deermined by personaliy facors, i.e., by emoional needs and consiuional empers, we should expec, furher, ha various neuroic and psychoic syndromes, which are hemselves explicable in erms of combinaions of personaliy facors, and someimes in erms of single personaliy facors, should show correlaions wih he musical choices. The immediaely needed invesigaions, herefore, seem o be: (1) a sudy correlaing he music facors wih primary personaliy facors, in a normal group; and (2) a comparison of psychoics and normals in erms of musical preference facor profiles. The hypoheses ha he music facors cor-
8 Raymond B. Caell and Jean C. Anderson respond o needs or o emperamenal facors can be esed by his design, bu one should also recognize ha a hird possibiliy exiss namely ha he discovered music facors represen affecive mood saes, emporary dynamic simulus condiions, physiological influences, ec. This alernaive, however, need no be invesigaed unless he presen search for sable personaliy associaes proves aborive. Some "funcion flucuaion" associaed wih mood will almos cerainly exis and i will aenuae our correlaions. Bu if our hypohesis is correc ha he major associaions will be found in relaion o relaively sable personaliy srucures, hen i could seem beer o rack down his residual, "flucuaion" variance laer. A ha poin no only he associaions of he music facors wih mood, bu also he individual endencies o high or low flucuaion on he music facors will bring in relaionships of furher imporance for undersanding musical preference and personaliy. A fourh design of research which is also immediaely needed is a facorizaion of a populaion of psychoics, o see wheher he srucure of facors is he same here as in a normal group. Unless here is some fairly close resemblance of he facor srucure in he wo groups, i would indeed be illogical o measure psychoics on he same dimensions as hose found among normals. Accordingly, we have also gahered daa for facorizaion of he same 120 excerps on a populaion of 100 psychoics, and his will be inercorrelaed and facorized if saisical man-hour resources can be provided by he Music Research Foundaion. The general reacion of culivaed liseners o he above proposiions has been ha our hypoheses neglec he role of inellecual and cogniive funcions in musical appreciaion. Our argumen is ha hese funcions are no primary bu are only means o ends echnical raionalizaions of he aeshee, perhaps changing superficially wih culural climae for saisfacions which are deeper and more sable. Iniial experimenal suppor for our posiion is given-by he fac ha he music facors do no apparenly correspond in conen o culural or echnical dimensions. A research designed o ackle his quesion more posiively has meanwhile been se in moion. I consiss of an experimen in which fify choices in picorial ar, hiry choices in archiecure, and fory choices in sculpure are inercorrelaed and also correlaed wih he facors in musical choice. If he same facorial dimensions appear here, aligning hemselves wih he music facors, and cuing across periods and culural inegraions, here will be addiional evidence ha we are proceeding beyond echnical, culural or hisorical paerns. Design of he Experimen The firs par of our invesigaion, ha wih normal subjecs, called for he adminisraion of he Musical reference Tes o a normal populaion which should be : (1) well varied in personaliy; and (2) simulaneously measured on a sufficienly reliable and valid measure of he primary personaliy facors. The main conribuion o he es populaion consised of 102 male and female subjecs, 76 of whom were Universiy of Illinois sudens, ages 18 o 29, and 26 of whom were "general aduls," ages 0 o 81. The remainder were esed in a second sub-group consising of sudens, boh male and female, ages 17 o 28. Since we needed o apply a personaliy es which deals wih primary and independen personaliy dimensions of known associaions we employed he I..A.T. 16 ersonaliy Facor Quesionnaire, which is also convenien for group adminisraion wih reasonably lierae populaions. The 16.F. includes inelligence as one dimension. Each of he 17 subjecs, herefore, ook a one-hour music preference es in which boh forms A and B of he music es were adminisered, and a half-hour silen session in which Form A of he 16.F. Tes was adminisered. The insrucions in he Music reference Tes are on he beginning of he record, and are as follows: "This is a es of your likings and dislikings in music. Your score has nohing o do wih how much you agree or disagree wih popular ases, bu only wih how much you agree wih yourself;
Measuremen of ersonaliy and Behavior Disorders 9 ha is, wih how consisen 1 you are. So ry o say, as each piece is played, wheher you yourself like i; wheher i is pleasan, so ha you would like o hear more of i, or wheher you would jus as soon have i swiched off. "On he score shee before you are numbers for he fify pieces ha will be played, each for less han half a minue. As each comes o an end, underline L, I, or D, opposie ha number, indicaing you like i, or have an inermediae, indifferen reacion, or dislike i. Dislike does no mean ha you hae i, bu only ha you don' paricularly like ha kind of music. In fac you should aim o have jus as many D's as L's underlined when you ge o he end. Try no o use I for inermediae more han you need. In fac, you should expec o end up wih very roughly one-hird L's, one-hird I's and one-hird D's. Bu don' boher abou ha oo much. Jus give your reacions as ruhfully as possible...." The adminisraion of he Music reference Tes o a group of psychoics ook place a Kankakee Sae Hospial, Kankakee, Illinois. In his case he subjecs were aken in small groups of hree or four a a ime, in order ha i migh be ascerained ha hey were appropriaely responding on he answer shees o every piece of music. I is well known ha diagnoses in differen menal hospials do no agree very highly (as shown on he individual cases ransferred from hospial o hospial), and ha he very proporions of manic-depressives, schizophrenics, hyserics, and oher psychoic syndromes, as diagnosed in differen insiuions, may vary considerably. As usual a good deal of difficuly was experienced in obaining a sufficien sample of some psychiaric syndrome groups. In acceping he group divisions finally used he crierion for classificaion was naurally he hospial diagnosis as reached in case conferences. A oal group of 98 psychoic paiens was obained consising of 6 alcoholics, 22 schizophrenics of mixed ypes, 10 manics, 7 paranoids, and 2 of oher caegories each no sufficien in number for separae use in our sudy. The subjecs were boh male and female, he age range being approximaely 2 o 60 years. 1 This obviously asks he person o be "rue o himself" and o give his considered judgmen; wih advanced music sudens on he oher hand i migh be inerpreed as being consisen wih regard o musical "schools," bu our subjecs were no music sudens. Resuls for he Normal ersonaliies The findings for he normal group will firs be described. Our iniial ineres urns on he reliabiliies, a minoriy of which, as menioned above, were low enough o sugges dropping cerain facors. These correlaions are presened firs as consisency (spli-half) coefficiens in Table 1, ar A and secondly as coefficiens of equivalence (correlaion of Form A wih Form B) in ar B of Table 1. The equivalence coefficiens perhaps do no do jusice o he ess because he highes loaded iems were in every case pu in he A form, since, when psychomeriss are unable o use he full lengh es, i is he A form ha hey will use. This reduces he equivalences (columns and 6) below he consisency coefficiens (columns 2 and ) which more ruly represen he inernal consisency, and are defecive for a 10-iem lengh of scale only on facors, 9 and 10, recommended o be dropped. The correlaions beween he sixeen facors of he 16. F. Tes and he eleven facors of he Music reference Tes were worked ou separaely for he wo populaions, as a muual check. For economy of represenaion he values in Table 2 are blanks excep where he correlaions on he wo samples are of he same sign and boh beyond he 1% level of significance. Then a single value he mean correlaion (Fisher's z) has been correced for aenuaion, by he given reliabiliies of he Music reference and 16. F. Tes measures, and recorded in Table 2. None of he correlaions is large enough o demonsrae a one-o-one relaion beween he music facors and he personaliy facors. Bu he se of 16. F. Tes facors associaed wih any one music facor has a psychologically consisen and compaible characer among he members in every case. For example, he personaliy facors correlaing significanly wih music facor No. 1 are dominance, surgency, oughness, radicalism and self-sufficiency all possibly relaed o some second-order, comprehensive facor of emperamenal oughness. Furhermore (and alernaively) he relaive magniudes of he correlaions are such as could be compaible
0 Raymond B. Caell and Jean C. Anderson Table 1 Reliabiliy Coefficiens for Facor Measuremens Facor 1 2 6 7 8 9 10 11 ar A ar B Consisency Coefficiens Equivalence Coefficiens (Whole Group) (Form A wih Form B) Half- Lengh Coefficien (.71) (.62) (.06) (.1) (.10) (.27) (-1) (.6) (.00) (.1) (.7) No. of Spearman-Brown Iems Correced o in Half Full Lengh.8.77.11 (used only experimenally).9.18 (used only experimenally)..8.6.00 (used only experimenally).2 (used only experimenally). Sample of 102 ersons.7.2 -.10.02.11.8.1.8.16.0.28 Sample of 71 ersons.6.7.2.19.9.27.11.26 -.01.11.1 wih a one-o-one relaionship of music and bly by longer scales for each facor and by personaliy facors if chance experimenal dropping iems in one facor scale having any error and he exising specious correlaions correlaion wih anoher facor). A es of among he facors wihin boh he personaliy his possible explanaion mus awai much and he music area could be eliminaed (noa- furher work on he purificaion of he pres- Table 2 Correlaions of Music reference Facors and ersonaliy Facors 16.F. Facors A Music reference Facors I 2 6 7 8 9 10 Tl 2C Jj C -. -.0 i!/,ry ' ~ ~~".OO ' "~~".ou F.6 -.8. fjyj -2A H.68 -.6. I -.6.70 -.7 L.1 -.9 M N.60 -.7 Qi.8 -.6.6 Q 2.8.7.8 o! - -.2
Measuremen of ersonaliy and Behavior Disorders 1 en facor scales. Meanwhile, however, his explanaion ress on he indicaion ha he highes correlaion for a given music facor wih any personaliy facor is also he highes correlaion for he personaliy facor wih any music facor. For example, facor 2 has is highes r wih Q, which r is also Q's highes r wih anyhing; facor 's highes is wih H, which is also H's highes; he facor column has is highes wih M, which is also he highes in he M row, and so on, wih very few excepions (noably facor 8). As o he consisency of psychological meaning among personaliy facors associaed wih a given music facor we may menion, in addiion o facor 1 above, ha facor 2 correlaes negaively boh wih paranoid endency and nervous ension, which endencies have been previously found, associaed by Darling (2); and ha facor, which correlaes essenially wih M ("Unconvenionaliy racical Concernedness"), also has some associaion wih Q2 ("Independen Selfsufficiency"). The alernaive possibiliy is hus indicaed, as suggesed above, ha where a music facor does no align iself wih a firs-order personaliy facor i may prove on furher research o correspond o a secondorder facor uniing he personaliy facors in some underlying common influence. For his reason music facor 1 has been called "Tough Sociabiliy Tenderminded Individualiy," which coningenly resrics he meaning prey closely o he psychological bi-polariy of personaliy facor I, wih which i is mos associaed, bu also suggess feaures of he oher facors wih which i has some degree of associaion. The over-all descripion of he personaliy dimension associaed wih his paricular music facor hus becomes remarkably similar o he Tender Tough-minded coninuum described by William James (6). Resuls for he Abnormal ersonaliies As saed above, in he accoun of design, he es was adminisered o 98 hospialized psychoics, divided ino hose four major syndrome groups which had each a sufficien number of well-diagnosed cases o promise some significance of differences, if such should exis. The means and sigmas on all 11 facors are shown for normals, for abnormals as a whole, and for he four abnormal syndrome groups, in Table. The differences are examined below by he es, firs wih respec o he differences beween he main psychoic group and he psychoic sub-groups, on he one hand, and he normal group on he oher, wih resuls as shown in Table. Nohing below a 10% probabiliy is recorded in he column. Table Scores of Normal and Abnormal Groups Normals» = 69 Abnormals w = 98 Alcoholics w = 6 Schizophrenics (D-) Manics «= 10 aranoids «= 7 Facor 1 2 6 7 8 9 10 11 1.6.7 1.1 9.6 2.7 6.8. 12.2 2.8 8..1 8..2 8.0 2.6 7..0.6 2.1 6.1 2.1 1..1 8.7.7 8.9 2..8 2. 11.6 2.7 1 2.8 7.0 2.6 7.9 2.2 9.2 2.1.9.6 2.9.7 6.. 9.0 2.1.1.0 12.2 2. 11. 2.7.8 2. 7. 2.1 9.0 2.1. 1.9 6.6 2. 1..7 10.0.7 8.6 2.. 2. 11..0 1 2.8 7.8 2.6 8. 1.9 9. 6.0 1.7..0 1.8.6 1.6 9.0 1. 8.2 2.6 12.7 2. 1 2.2 7.2 1.6 9.1 2.6 9.0 2.2.9 2.1.6 2. 12. 6. 12.1.6 9. 1.7. 2.1 1 9. 2.6 7.1 2. 8. 2.7 8.1 1. 6.6 1.. 2.
2 Raymond B. Caell and Jean C. Anderson Table Significances of Differences of Abnormal Groups from Normals Facor 1 2 6 7 8 9 10 11 Toal Abnormal Alcoholics Schizophrenics Manics aranoids 1.7.9 2. 6.7 7...1 8.0 1.7 lev. sig.) 0 * 1-2 2-0.6 1..1 6.2.8.8 1.0 lev. sig.) 2-0 1 (% lev. sig.) 0. 1.8 0 2. 2-.7 0.9 1.0.7 0.9 1. 16 0.6 2. 2. 0.1 lev. sig.) 2-0 2-1-2 0..6 1.0 0. 1. 1.6 2.7 v. sig.) 2- * This indicaes "beyond he 1% level." The psychoics differ, beyond he 1% level, in being lower on facor 2, lower on, higher on 6, lower on 7, and higher on 9. These differences similarly characerize he alcoholics, who happen o be he larges group, hough sill consiuing only 6 ou of 98 psychoics. The schizophrenics differ a he 1% level only by being higher on 6 and 9. The manics have similar endencies on hese facors (2-% level), bu also come up wih a new difference (1-2%), by being lower on facor 11. The paranoids have no resemblance o he alcoholic and schizophrenic majoriy, bu share he manic's lower score on 11 and show a new paern in being lower (\% level) on S. Before commening on hese findings le us examine, finally, he capaciy of he es o discriminae among various psychoic syndrome groups hemselves. The es examinaion is presened in Table. Table Significances of Differences of Syndrome Groups, from Toal sychoic Group and One Anoherf sychoics Alcoholics n = 98 & 6 sychoics Schizophrenics n = 98 & 22 sychoics Manics n = 98 & 10 sychoics aranoids n = 98 & 7 Manics Schizoids n = 10 & 7 Facor 1 2 6 7 8 9 10 11 (% lev. sig.) 2.6 1-2 12 2. 1-2 1. 0. 1.0 (% lev sig ) 0. 0.1 0.1 1. 1. 0. (% lev. sig.).8 1. 1. 0.0 2. * (% lev. sig.) 1.8 2.1 0 1. 0.1 0. 1.7 2.1 0 1. 0.6 0. 2.6 0. 0.1 1.7 v. sig.) 1- fonly he noeworhy levels of significance (beyond 10%) are enered.
Measuremen of ersonaliy and Behavior Disorders I will be seen ha he schizophrenics above have no significan differences and comprise, as i were, he prooype of psychosis. The alcoholics, in spie of being he larges group conribuing o he "mean psychoic," differ very significanly by being lower on 2 and 7. Manic-depressives show a disinc, characerisic paern which, in spie of he small numbers, is saisically significan, boh in relaion o general psychoics and o schizophrenics. From boh of he laer hey differ by being higher on facor, and from he oal psychoics by being lower on facor 11. Wih respec o facor he manic-depressives and he schizophrenics fall on opposie sides of he normal mean, which suggess ha his facor has close connecion wih he dimension envisaged by Bleuler, Kreschmer, and ohers. I is ineresing o noe ha he pararioids share some of he characerisic differences of boh schizophrenics and manics, bu have one addiional divergen facor and finish wih a uniquely characerisic profile. These resuls, if confirmed on anoher sample, indicae ha he es is a powerful means of psychiaric diagnosis, for if differences on single facors exis a such levels of saisical significance he predicion from he combinaion of facors in his paern should yield subsanial separaion of he wo groups. For example, since he facor measures are in principle independen, he difference of he normals and abnormals would be significan approximaely a he (1/100) level, and he resuling absence of any subsanial overlap beween he wo disribuions should make predicion even on he individual case highly reliable. As far as an exploraory sudy permis we can roughly indicae he diagnosically useful paerns as follows: (1) o disinguish psychoics from normals: low 2, low, high 6, low 7, high 9; (2) alcoholics should be similarly disinguished, bu also by being especially low on 2 and 7, which paern should furher disinguish hem from oher psychoics; () paranoids disinguish from normals by low and low 11; and so on for oher pairs of groups. Examined in erms of he meanings of he correlaions found beween hese music facors and normal personaliy facors hese psychoic paerns are psychologically consisen and recognizable. Bu he psychoic associaions also hrow furher ligh on he psychological meaning assigned o he music facors. Thus, in erms of he labels now assigned o he music facors in he handbook () he paranoid paern combines he facor of "aranoid Imperviousness Over Anxiey" (Low ) and "Schizohymia" (Low 11). The alcoholics combine "Frusraed Emoionaliy" (Low 2) and "Wihdrawn Schizohymia" (Low 7), incidenally corresponding o he 16. F. Tes facors (see Table 2) known as Q (Nervous Tension), and H (Wihdrawn Schizohymia), which paern well fis he published descripions and analyses of he dynamics of alcoholism. The manics disinguish from normals by being high on he facor of Eccenriciy, on Dominance, and on Frusraed Emoionaliy (i.e. on music facors 6, 9, and 11 [ ], corresponding o 16. F. facors C[ ] [wih ohers], E and QJ. The original general inerpreaion of he Q facor of "Jieriness" or "Somaic Anxiey" as "Frusraed Emoionaliy" is srenghened by his associaion of he facor wih boh alcoholism and mania, and by is absence from he schizophrenic profile. Similarly ligh is hrown muually on he alernaive escapes of alcoholism and manic exciemen, by he associaion of he "Wihdrawn Schizohymia" (16. F. facor H [ ]) wih he former, and of "Eccenriciy" and "Dominance" wih he laer. Wih increasing invesigaion of he physiological, social and dynamic meaning of such uniary, measurable facors, as esablished in normal populaions, he way oward causal explanaion of he psychoses could become much more clear. Space does no permi here any exensive discussion of he relaion of he personaliy associaions of he music facors o he characer of he music per se, in he facor iems. However, one may noe ha he psychoic group seems o prefer, according o he musical iems in facors 2, and, music ha is relaively slow and simple (and also relaively "sad"). Furher, from he difference on facor 7 i can be added ha hey end o avoid brighly colored (harmonically and
Raymond B. Caell and Jean C. Anderson exurally) music in favor of clear harmonic progressions, swee melodies and subordinae accompanimen. The excepion o his paern is he manic group, which, on is disinguishing facor (No. ), prefers fas, exhilaraing, simulaing pieces wih exural complicaion, rhyhmic variaion and less obvious melodic oulines. These associaions migh roughly be explained in erms of empahy, bu as more evidence accumulaes hey should receive more direc research invesigaion, especially in he ligh of such research approaches as hose of Rigg (7, 8). Summary 1. A previously compleed facor analysis of 120 very diverse musical excerps was used as a basis for consrucion of a Music reference Tes of ersonaliy, se up o measure eleven facors by 100 iems on wo sides of a long-playing (% R..M.) record. As he equivalence of he A and B forms is inadequae for hree or four of he facors, i is recommended ha hese be reserved for research improvemen, by iem analysis, and ha he remaining seven or eigh facors alone be used as inernally valid measures in rouine applied psychology, noably in seeking exernal validiies by predicions in clinical and guidance psychology. 2. Since he esablished groupings of iems do no correspond o musical schools or periods (hough possessed of some consisency of musical characer) i is hypohesized ha hey represen dimensions of personaliy (especially of emperamen) deermining ase. Correlaion wih he 16 ersonaliy Facor Quesionnaire Tes, on normal populaions of 102 and 71, confirmed his by yielding many significan correlaions. A one-o-one relaion of music preference and personaliy facors canno be proven by hese resuls, since boh measures of facors are imperfec., Bu he correlaions, correced for aenuaion, are a leas consisen wih he hypohesis ha, bu for conaminaion, he same personaliy dimensions deermine, in all bu wo cases, boh he verbal and he music preference facors. Coningen iles have been given o he music preference facors in accordance wih he personaliy associaions. These iles proceed on he probabiliy ha mos music facors are primary personaliy facors hough some may be second-order personaliy facors.. Applicaion of he Music reference Tes o 98 paiens in menal hospials revealed several facor measure differences, significan a he \% level, beween psychoics and normals and beween various psychoic syndrome groups. If confirmed on furher samples, hese paern differences are so marked as o make he es a valuable adjunc o psychiaric diagnosis. The meaning of he music facors as indicaed by he personaliy facor correlaions agrees well wih he meaning as found independenly in erms of he associaions wih psychoic syndrome groups. These scales migh herefore have value in hrowing furher ligh on individual psychoic syndromes. Received February 2, 19. References 1. Anderson, H. H. and Anderson, G. H. roecive echniques. Chap. 2. New York: renice Hall, 191. 2. Caell, R. B. A guide o menal esing. London: Universiy of London ress, Third Ediion, 19.. Caell, R. B. and Anderson, J. C. The I..A T. Music reference Tes of ersonaliy. The Insiue for ersonaliy and Abiliy Tesing, 1608 Coronado Drive, Champaign, Illinois, 19.. Caell, R. B. and Saunders, D. R. Musical preferences and personaliy diagnosis. I. A facor analysis of 120 hemes. /. gen. sychol, 19, in press.. Caell, R. B. and Wenig,. W. Dynamic and cogniive facors conrolling mispercepion. /. abnorm. soc. sychol., 192, 7, 797-809 6. James, W. ragmaism: a new name for some old ways of hinking. London' Longmans, 1911. 7. Rigg, M. G. Musical expression: an invesigaion of he heories of Erich Soranin. /. exp. sychol., 197,, 2-. 8. Rigg, M. G. Speed as a deerminer of musical mood. J. exp. sychol., 190,, 66-71.