on on Os =e 6s en, an aS os a os te oes of os ob an ie Reprinted from JounaL oF Anwonseit. axo Soctar Paxewotocr, Vol. 68, No. 3, Mar. 1964 BODY POSITION, FACIAL EXPRESSION, AND VERBAL BEHAVIOR DURING INTERVIEWS '* PAUL EKMAN* School of Medicine, Univerity of California, San Fraxcico ‘The communicative value of body postion and facial expresion was evaluated ty woeaturing an O's ablty to detect 4 relatonship between nonverbal and ‘verbal ‘ebavior which bad been simultaneously enitied. The verbal and owverbal smal were collected during ? diferent standardized sires lter~ ‘ews, Judges (Js) were shown pairs of photographs together with short ‘written speech samples snd eequied on each tril to pick the photograph which Iatched the verbal behavior. In 4 separate experiments with diferent sroups of Js accurate judgments were obtained. Evidence for a teatonship between oaverbal and verbal behavior simultaneously emitted was replated acrow P'difeent samples of interview bebavior and wnder 3 cue condilons—seeag the head, body, oF whole person, ‘There has been relatively litle systematic lnvestigation of the information which may be transmitted. through spon:aneous nonverbal behavior shown during, interpersonal trans. action. Research on body movement and facial expression has had to deal with a Phenomenon which is continuously occurring, has no readily apparent unit of measurement ‘or method of evaluation, and is both dificult ‘and expensive to record. Despite the clinical conviction that nonverbal behavior provides important dues about an individual's emo- tional experience, such beliefs rest on litle ‘more than anecdotal evidence or speculation. In fact, research on interview bebavior ‘has increasingly focused on verbal rather than nonverbal interactions. ‘The major problem in exploration of the ‘nonverbal aspects of interview behavior, may at least intially appear to be the acquisition of permanent record. Actually, unless the ‘body movements and facial expressions are obviously meaningfal from simple inspection, the investigator is quickly overwhelmed with 4 mass of photographic stimull, which are just as balling and complex as were the ‘original behaviors themselves. The central dif This project was supported by a. postdoctoral research fellowship, MP" 6092, from the National Insitute of ‘Mental Health, United States Publle Health Service. The author is indebted to Sanford ‘Autuma for his maay valuable suggestions about ‘he Fecha Steal amy, and prepa Now at San Frandco State College, ficulty in research on spontancous nonverbal behavior is the development of fruitful tech riques for evaluating the information which may be contained in records of nonverbal behavior. Two approaches can be_distin- ‘gushed which although related, usually have been separately pursued. “The first approach has measured variations in nonverbal behavior as a response to some other factor, such as interview structure ot Patient mood. Sainesbury (1955) and Ditt- mann (1962), for example, have reported 4 ‘between changes in nonverbal behavior and the content oF structure of an interview. They did not study the related question, owever, of what may be com- ‘municated to an observer by these changes in nonverbal behavior. In the second research approach, nonverbal Dehavior is presented as a communicative stimulus and an observer's response to that stimulus is measured. Most of the experi- rents using this second approach have been Performed with posed behavior which oc- curred during isolation rather than in the midst of an interpersonal interaction (for example, Schlosberg, 1954). Moreover, these studies have focused upon judge agreement rather than accuracy in understanding com- ‘munication through nonverbal behavior. At tempts to measure the information trans- nitted through nonverbal behavior during an Interview ave been inconsistent. in their methods and results (Giedt, 1955; Mab, Danet, & Norton, 1959). These experiments 2s296 avn Exnan have not allowed specification of the actual nonverbal cues which served as the basis for the judgments required, A further criticism of the research studying nonverbal behavior as a communicative stimulus, is that it has relied on the application of ‘highly complex verbal concepts for evaluating the information transmitted by the nonverbal cues. A judg- ‘ment task which requires ratings on emotional scales, diagnostic or psychodynamic formula- tions, may well miss much of the information ‘which nonverbal bebavior can communicate to an observer. The present series of experiments combined features of both research approaches by meas- turing an observer's response to nonverbal behavior (Approach 2), in a task which re- quired him to relate nonverbal behavior to another aspect of the interview situation (Ap- proach 1). A relationship between what a person does with his body and what he is saying was assumed, and a matching pro- cedure was borrowed from the early research on expressive behavior (Vernon, 1936; Wolf, 1943) to test the following hypothesis: Judges can detect a relationship between nonverbal and verbal behavior which was simultaneously emitted during an interview, Four separate experiments with different ‘groups of judges will be reported. In each of these experiments judges responded to verbal and nonverbal stimuli which had been gathered during the course of two interviews. Genera Meron Interview Procedure ‘A standard rather than lineal interview was cplayed to increase the comparabllty of the two Interviews and to insure the presence of diferent ‘fective reactions, The interview structure was ‘achieved by programing both the syle and content of the interviewer's (the experimenter) behavior foward the interviewee (the subject). After an Introductory affectivly soutal peeled of 10 minutes, the experinenter sttscked and crtced the wbject oie of vccupation, competency, and motivation. Throughout this sess phase, the experimenter continually questioned and) iterupted ‘the subject, responding tangentally to Mis reps, and generally ving litle opportunisy for any delene, After 19 uch aioues the experimenter iidated the extharls phase by apologizing for being hose, episining i ‘had been seeesary in order to study the subject's feacdon to stress, The subject was praised for Bip reslleney and performance onder sre, and humor ‘minutes of catharsis, the experimenter was reat Soring in manner and attempted to bring about a teleae of tension. Interview Participants Two such standardized interviews were recorded, with completely diferent paricipanta in each of the Interviews All four. interview participants were from the Seid of pavcbology. The too ferent aperimenters were sal! research pevchologsts 38° veteran's hospital, having a pestion of some Authority over the two diferent trainees ‘ho served fs subjects, All of the participants were told. that ‘he experiment was concerned wilh interviewing, and tat thelr behavior would be tape recorded ‘and observed. The experimenters so” knew tat they would be photographed and nonverbal be- havior would be studied. Recording Method Photographs showing profile view of both the experimenter and the subject vere Laken through fpesway vison sereen with @SSumllimeter still ‘camera. The verbal behavior ‘war tape cecarded, ‘The verbal and nonverbal records were synchronised bya switching “device mounted on the camera Which ‘paced an audible signal over the tape fecorded voices whenever a peture wa taken ‘The frequency of photographic mp avery 30 seconds in Interview A and frery 18 seconds ia Interview By was dictated. by Tinitaions in the fm capacity of the camera. Selection of Verbal and Nonverbal Stimulé ‘The judgment task involved presenting 2 wetten Speech samples {fom the Interview along swith pairs of photographs. On each trial the judge svar required fo pick om pair of pictues the tne ‘which best Sted or matched ‘the verbal Behavior ‘The universe of verbal behavior hich could be sampled for presentation ‘ine Judging task wes delimited by’ the number of occasion’ during the interview in which a photograph nad. been taben, There were 26 such polats during Interview A and 71 point daring Tateview B. Speech sampler Of from 30" to 60 seconds tere choren from this pool of verbal behavior they” appeared intel~ pile sehen lifted from the contest of the total onversauion, ‘There wat 30. inspection af the Photographs during thie selection procedute, An Idee card. was prepared for each speech sample ‘whlch stated whether the verbal Interecion had been taken from the tres or catbarse.phave of che Interview; provided a onrsentence background synopsis; coataned the speech sample of one ov mo ‘atements by both the experimenter andthe sub: Jeet rich the aetsl moment which tad been photor Srapied indlated by a ccled word Pais of potographs, correct and incorret, were shen laced to) steompuny each speech sample ae egae Pes Pees ee ee ve PP woea fs os ee cece as on Cs as te ones oF, on on on Ee ‘Exenessive BEEAvion 297 ‘The correct photographs were the oves which aetialy had Sena taken dunog the speech samples correct photographs were randomly. selected (0m, "photograph pool delinited by two factors: Toe ‘mouth poston, open or closed, had to be the same 1 both the incorrect and correct photographs, and I the experimenter (or the subject) were speaking Jwben the correct photograph was taken, then the Jncoreeet photograph had to be Selected from photo- ‘apis. allo taken while the experimenter (or the ‘abjecr) was speaking ‘Two types of incorrect photographs could be chosen for pairing with each coreet one. The ise correct photograph could be taken from the same phase of the interview as the correct plctre, oF {aken from the other phase. Since speech samples ad bean selected (om bath the strest and catharsis phases of the interview, there were Tour types of ‘orrectincorrect. photograph pairs which could be Dresented to the judges With a verbal sample om the ses phase of the interview, the juages could be required to. choose Betveen two. sree. pboto- araphs of beewean sess and 2 catharsis ai. Stmiany, with a verb sample from the catharas phase of the interview, the judges could be presented With either ewe catharsis photographs of with 8 (Stharsis and 2 etree photograph Experiment I ‘The problem was to determine if any telation- ship between verbal and noaverbal behavior simul: {ancously eated is communicated to an observer. Hypothesis: untrained. judger can choove {rom 2 pair of photographs the one which wat taken during 2 even speech sample Subject. Eighteen college freshman, consisting of the entire introductory pryehology css taught by the author, served as judges. There were 10 males, S'lemaies, the median age was ‘Method. Fourteen speech samples and accompany ing correc-incorect photograps were selected from Interview B by the iethod described above, Eight of these speech samples were from the stress phase of the interview; st, from the eatharis pase Four % ve inch enlargements were made of exch photography showing a protle view of both the fperimenter and the subject. The Judgment ‘ace ‘was individealy adminitered In ofder to. alow ‘random variation af the order of peseatig the ‘peechphotograph items. "The judge was shown Dire of photographs, each of which came tom the aime Interview phase on ball ihe lems hile oF fe remaining tals the incorrectcorect photoreaph alr wece drawn fom diferent interview poaces ‘Able of random numbers was sied to determine ‘woether for a particular judge the colee on 2 ba Heclar speech would be between, photograph pai ‘om ihe sare oF diferent interview phaser Prior t the task, the udee sas given the folowing instructions to ret ‘Tis is a ses of your sil im imerpretng and understanding gestures snd Sody moverests ou vil be showa some photographs which were taken every 15 seconds during’ a thirty minute Inter View. The interviewer, or Examiner, was 2 sail Dpevehologst at a hospital, and the person inter~ “lowed, the Subject, as 4 student Ia taling to become a psyehologist The Subject was told that ‘be was paricpatng ina research project on inter- Viewing techniques, and that he’ would be ob Served through a oneeway vision sereen and tape fecorded. The interview was prearranged £0" the Examiper to ask facial questions during. the Inwoductory’ Sst 10 misutes, and then to bee ome hostle and chaleging, questioning. the Subject's academic training, the adequacy of bis reparation to take his examiaations coming up {he folowing month, and to eoatnually Interrupt ‘him. This Stressful period lasted about 10 minutes In the Final phase of the interview the Examiner fplained what he had been doing; that it had been part of the research to try and provoke the Subject, and generally attemplad to reassure the Subject. Ta actuality, this Interview plan did aot work out perfectly. ‘There was aot clearly de- Sed a diference tetveen the three phases of the Interview a had been expected. ‘There was some sires for the Subject throughout ‘he Interview ‘The final phase was not completely successful ia producing some relef from the sess since the Subject “knew that the experiment wat sl ‘continuing, and perhaps was apprehensive about ‘what might’ be Coming nest, Neverbeles, there ere some important diferences between tie Yarious phases of the interview. The stress phase Aid ‘have more overt expression of hostlty” and tension; and in the Gal phase the Subject it fxpeieace some ‘ele, at least knowing that the worst was over. You will be given to Tead 1 ckeps from the tapereconded conversatioe, ‘The verbal excerpts cach took less than minute of interview time. One photograph wa taken nuting ach of the veal exerpis, and le actual “rds Spoken waen the photograpa was taken are ci ed in red, ‘After you read carb varbal excerpt, ‘sil be showa's pale of puotegraphs, and. your job as a jodge wall be to pick which of the two Diceares in ihe pair Gis we bese oF matches the Serbal excerpt. ‘Tall of the. photorrspte. she Examiner appears on the let aod the sect 30 he Tight, ‘Your lark then is tu ‘uterine for ach tet of two photorrapia which ose is mest ‘ikey "te" have Been taken during ‘the veroal Results. The number of curtect of ace curate choives for each jude wa9 tabulated, sition of obidive arcimacy scores ‘was evaivisted by comparicon with @ theorst feal_melian representine ike score night Ee expected if only shance factore were chit nan simply298 Paut EKMAN TABLE Inteview A (6 eral ‘Sained median Sabre cepted maton tS | Tnteviow B (14 tials) ‘Gioines mean Gestion "spect median of 7 Interviews A and B combined (20 vials) “Ohained median Sp Shave expected median of 10 ee flected chance factors, then there would have been an equal probability of either photo- raph being selected on each trial. The ex- pected median was therefore equal to a cor- rect choice on half of the trials. Significance fests were derived by applying Wilcoxon's matched-pair ranks test (Siegel, 1956, p. 75) to the difference between the obtained scores and the expected median. It can be seen in Table 1 that the obtained median accuracy score was significantly better than might be expected by chance. Table 1 also shows that more than three-fourths of the judges scored above the expected median. Experiment IT Problem. The purpose of this experiment sas to replicate the findings reported above with a diferent ‘roup of judges and to extend the generally of the Fults by" tiling. verbal and. nonverbal” stimuli ‘drawn from an additonal interview. A. secod Durpose was to determine if it were sl posible to Fateh verbal and_soaverbal. behavior with more restricted. nonverbal eve formation. Tn. splot Study (Ekman, 1961) aa attempt was made 10 Specly the nonverbal cues required Tor the judge ‘ment task by Uiitng the judges to seeing either bead or body. The results indented lite ierence Im accuracy Ifthe judge saw head of whole person cues, Bat accuracy ‘was ‘much lower ‘with. phato- ‘raphe showing only. bod)” posidoa, Hypothes Sudges can choose from a pair of photographs the tne which was taken during a siven speech sample (when responding to pictures of the whole person. This diserimination between photograph pai is 3 2 ner | # B tae zs | sae] oi) sa 3 Be se wor |r juz fase) os | a a ea ako posible when the Judges are limited Showing only body postion ‘Subject Since the plot needy had suggested that working with body posidon photoseaphs vas a Scale task, an attempt was made to select as Judges individuals who might be able to best under~ Stand. this typeof nonverbal behavior. Sixteen Drofesioaal modern dancers verved as judges: 1d fwere female, the median age tas 25, the median tduestlon sas 16 years Method. ‘The same task was employed. but the procedure was modifed for group presentation {ther than individual administration. A Booklet was Prepared. which contained an iastuction shee, Silla to the one used in the Srst experiment, and 20 speech samples. Fourteen of these were the Samples used in the Gist experiment, the other 6 speech. samples were drawn com "Interview | Toeve of the 20 samples were (com the stress phases of the interviews, the remalning & samples ‘were drawn from the catharsis phases. Half of the forrect-incorreet photograph pairs were compoved of pictures taken during the Same interview phase, the remaining 10 pairs of photographs consisted of ictures from diferent Interview phases. The poto- fraphs were mice into two sats of postive trane- fareaces The beads of both the experimenter std the subject were covered with opacue ii in. one Set of transparencies, s0 that only the body postion foul be seen ‘Two projecion screens and two 3s-millimeter side" projectors were used 10 show the comect- Incorrect” photograph pais simutaneousiy tothe sroup of judges alter they rexd each af the spewed Samples in the booklet. The judges Ast completed ‘the 20 tials ceponding to photographs of body posidon, and then after 20-minute intervening {ask repeated. the procedure sith photographs showing the srbole person, ‘The appearance 91 che we pe ST ewes Pepe Pp e.s eee SPen Os as es os on af an ts as te anes af os of an Ee Exenessive Brnavion 299 correct photograph on either the lft of ight pré- jection fereen dred under tbe body and whole person cue conditions and was balanced across the iota, Results, The technique of data analysis was the same as described in the first experiment. ‘The results on Interview A trials and Inter- view B trials in this experiment and in Ex- periments III and IV will be interpreted together in the discussion section of this paper. The results in Table 1 show that ac- ‘curacy on all 20 trials (Interviews A and B combined) replicated the findings of the first experiment. The judges were able to pick the photograph which matched a verbal sample ‘when limited to seeing body position and also when shown the whole person, While it appears that the level of accuracy was greater when the whole person was seen, the differ- ence between the two cue conditions was not tested since there had been no control for practice effects. Experiment III Problem. In interpreting the results from Experi- meat Ife was not car whether the judge? ability to match “Body. only" photographs with speech Simpler was due to any special characteristics Which ‘might be associated wilh dancers or if such ac- forecy might be found in a more unselected. group of judges, The major purpose of this experiment ‘ras then, to determine if @ heterogeneous group of Judges could match body position photographs with Speech samples. A secondary purpose was to replete the findings of the two earler experiments. Hy= pothesis: A heterogeoeous group of judges can choose from a pair of photographs, showing only the body postion, the one which’ was taken uring & ven Speech sample Subject, Fifteen undergraduate stndents in the authors introductory paychology evening "cas served as judges. Twelve were males; the median ge was 22.5 witha range from 18 to 5S. ‘Method. Exaclly the sime procedure described in Experiment IT was followed, Twenty speech samples were presented in booklet form. The Judges rst fesponded to body posidon photographs, and. then ‘o pictures showing the whole person Results. The techniques of data analysis I shows that for the results of all 20 trials, the hypothesis was confirmed and the easlier findings were replicated. No comparison was made between the accuracy achieved by the dancers in Experiment IT and this group of evening psychology students, since the two groups of judges differed in age, education, and sex distribution as well as in thelr choice of vocation Experiment IV Problem. Tae design io Experiments I and IIL id not permit study of relative accuracy in match ing nonverbal and verbal bebavior as 3 function of the nonverbal cue information availble. The Durpore of this experiment was 10 test impresions from the plot Suady televed to earler by com- paring accuracy between groups of judges who we Inlted to seeing ether bead oF body position con verbal cuss. Hypothesis: Judges can choose from 4D pair of photographs te one which was taken uring 4 given speech sample when responding to father heat) or “Body postion cues, The judges Showa bead cues will acleve greater accuracy 1Ban the group of judges seing body poston cues ‘abject Female students in two freshman psy- chology clases served. ax judges ‘There were 27 Judges in the group seeng body position photo- sraphs, 29 judges in the group responding to bead Photographs. The median age in both groups was 1S ‘Method. ‘The general procedure ascribed ia Ex. periments If and IIT was followed exeepe that to Separate groups served. as Judes, each coup Fespoading’ to oly one cue condition, either nex or body, Results, The technique of data analysis described earlier was again used to evaluate the results which are shown in Table 1. Con- sidering the group which responded to head cues, the results on all 20 trials combined gen- erally conform to the findings reported earlier for judgments of the whole person. The group responding to body position cues, however, did not achieve a significant level of ace ccuracy in their scores on all 20 trials. Exar nation of the results on Interviews A and B separately revealed that while accurate judg- ments were achieved on Interview B, the judges were significantly (two-tailed ‘test) below the chance level of expectation in their responses to body position photographs from Interview A. There was a tendency for body photographs from Interview A to be similarly misjudged in Experiment IZ, although it had not reached statistical significance. Differences in accuracy under the two cue conditions were evaluated with a Mann- Whitney U test. The group responding to ‘head cues achieved greater accuracy than the body position judges in their performance on all 20 trials (p <.05) and in their judgment of photographs from Interview A (p< 001)300 Pave EKMAN Drcussion, Accuracy in matching materials from the ‘wo different interviews was related to the type of cue information provided—head, body, or sthole person cues. Although there was some inconsistency across the four ex- periments, by and large a significant level of accuracy was achieved on both Interviews A and B if either the whole person or the head were shown, There were marked difer- ences in the accuracy obtained on Interviews A and B, however, when only the body Positions were shown. The body cues in Interview B seem to have provided suficient information for matching in two out of three experiments, Table 1 shows that such accu- racy was never reached on body cue judg- ments of Interview A. This difference in accuracy sas evaluated with the MeNemar test for the significance of change (Siegel, 1956, p. 63) by comparing the scores on the two interviews for each cue condition in Experiments TI, TI, and IV. Judgments on Interview B trials were more accurate than Interview A trials only under the body cue condition (Experiment III, » <.02; Ex- periment IV, p<.001). The results of Experiment IV on differences between judg- ments made on the basis of head or body cues provide related information. Head judgments were superior to body position judgments for Interview A (p< .00i), while 42 tend in the reverse direction was seen for judgments made of Interview B. There are a number of hypotheses which might explain why in one of the interviews hhead but not body cues could be related to verbal behavior. These explanations involve differences in sending behavior, that is, the ‘extent to which people may differ in the information which is conveyed to an observer through their nonverbal behavior. The ex- Perimenter and the subject in Interview A may have been head rather than body ders: or, they may have been sending information’ in their bodies which was un- related .or contradictory to the verbal be- havior: of, the body cues while related to the verbal behavior may bave either anticipated ged behind it. The choice between vpotheses will have to avait further ‘A communication paradigm involves con- sideration aot only of sending behavior but also of receiving behavior, that is, possible differences across observers in theit skill or sensitivity in understanding nonverbal be- havior. While these experiments were not Gesigned to study this question, one aspect of receiving behavior can be analyzed by determining whether accuracy in judging one interview was related to accuracy in judging the other. Spearman rank-order correlations vwere calculated between accuracy in judging Interview 4 and Interview B for the various cue conditions in the last three experiments. Of the six possible correlation coeficients, only two were significant and these were both in a negative direction, Thus, it would seem that accuracy in matching verbal and non- verbal behavior from one interview was In- dependent of such accuracy in working with materials from another interview. The more ‘general question of whether sensitivity as a receiver is specific to the behavior of a par- ticular sender, or group of senders, must also await further research. ‘The overall results on the four experiments suggest that some information related to the verbal behavior is conveyed by spontaneous nonverbal behavior during interviews. In at- tempting to specify just what may be co ‘municated through body position and facial expression, both specific and general classes ff information can be outlined. The most obvious example of specific aonverbel com- munication would be a gesture such as a smile ot fist shake which has a direct veroai equivalent or translation, Even less symbolic nonverbal acts such as swaying of the body for tapping of the foot may have a specific communicative value by emphasizing or focusing attention on a particular part of a verbal message. Although not studied in these experiments, nonverbal acts, such as movements towards or away from another or direction of eye gaze, may not necessarily bbe related to the verbal message, but instead ‘may communicate specific information in the language of the relationship (Ekman, 1962) In addition to these specific meanings, son- verbal behavior may siso communicate more general or gross information about the sende Examples ould be information & oe oe ee ee oe ee ae oe oe oe oe -~=2 Peon op Gs as ue on an af as Gs as Ge apes OF os of an ie Expuessive BEHAvion 301 disckarge of tensioa, In the present experiments, specific in- formation must have been communicated by the nonverbal behavior. The judges were not simply responding to the gross. differences between the behavior shown under stress and catharsis, since they had been able to dis- ccriminate between photograph pairs taken from the same emotional phase of the inter- view. This diserimination between two stress ‘or two catharsis pictures was as accurate as the judgments made of pictures from different interview phases, Although the majority of the results ob- tained were statistically significant, and the distribution of scores was skewed towards accuracy, the absolute level of correct choices ‘was far from perfect. There were a number of factors which made this judgmental task extremely difficult. If the nonverbal behavior photographed had contradicted or was not directly related to the verbal behavior, accu racy would suffer. Similarly, if the nonverbal behavior anticipated or lagged behind the verbal behavior, the photograph could not hhave been matched with the short speech sample provided. Presenting the speech sample asa typescript without any vocal cues may have served to increase the am- Diguity of the verbal behavior. Finally, the use of still photographs rather than motion pictures eliminated any cues from sequence for movement patterns which are customarily available to an observer. Tn light of these limiting factors, it is particularly impressive that consistent evi- ddonco was found for at least a partiat relation- ship between verbal and nonverbal behavior. ‘These experiments indicate that body position and facial expression spontaneously. shown during an interview are not random act or noise, but have specific communicative value related to the verbal behavior. Further ‘more, this relationship is not obscure or available to only the privileged few, but can be detected by untrained observers. REFERENCES Drrewamis, A. T, The relationship between body ‘movements and moods in interviews. J. consul Peyehoy 1981, 2, 460. Exxtis, B. Body language during interviews, Paper ead at Tater American Congress of aychology, Dexco City, December 1961, Banwax, P. The communication of interview stress ‘through ‘body language. Paper read at Westero Peychologieal Awocation, San Francisco, Apdl Gtion, F. H. Comparison of visual, content, and ‘editry cig in interviewing. J. conell. Psychol, 1955, 19, 407-416, Mam, G,' Daxer, By & Norrox, Nes, Redection of major pesoualliy characteristics ia. gestures fad body movements. Amer. Peychologit, 159, 14 357. (Abstract) Saneesauny, P. Gesturd movement during psych tne interview. Prychosom. Med, 1985, 17, 458 ie. ‘Three dimensions of emotion 1956, Gh, 31-88. ‘mate, §. Nonjorametri itatiale: fr the behaviors Tclences, New York: McGraw-Hill, 1956. Vezwon, P. E. The matching’ mediod applied t favesigatons of personality. Pryekal, Bul, 1996 33, 4-177 Worry, W. Tie expression of perionality.New York Harper, 193. (Received November 19, 1962)