Anxiety Darkness

download Anxiety Darkness

of 9

Transcript of Anxiety Darkness

  • 8/3/2019 Anxiety Darkness

    1/9

    Journal ol Consulting an d Clinical Psychology1976 , Vol. 44, No. 1, 83-91

    General Versus Specific Trait Anxiety Measures in the Predictionof Fear of Snakes, Heights, and DarknessMartin Mellstrom, Jr., George A. Cicala, and M arvin Zuckerm anUniversity of DelawareT he relations between general and specific trait anxiety tests and fear measuresin three actual situations were investigated. Both types of test were adminis-tered to 76 u n d e r g r a d u a t e females early in the semester. Later, each subjectw as exposed to each s i tuat ion, where observer 's ratings, behavioral , and sub-ject ive fear measures were obta ined. The results indicate that th e specific testswere clearly superior to the general ones in predicting fear of snakes but onlyslightly superior in predicting fear of heights and darkness. It was concludedtha t th e overall superiority of the specific m easures supported th e current t rendt oward s i tuat ion specificity in personality assessment. The roles played bythrea t to self-esteem, th e form of the general trail anxiety measure , and sensa-t ion seeking are also discussed.

    Most at tempts to measure anxiety havebegun with th e assumption that it is a t rai t ,that is, an enduring disposition of a personto act in a reliable manner in a wide varietyof situations (Allport , 1937). For example,Taylor's (1953) Manifes t Anxiety Scale(TMAS) was intended to measure a person 'scharacteristic level of anxiety, or drive,which was thought to be relatively constantover time and across situations.Recently, the trait concept of anxiety hasbeen questioned. Mischel (1 9 6 8 ) has arguedthat trait tests lack predictive validity be-cause they do not take into account th esituational specificity of behavior. Hodgesand Spielberger (1966, 1969) , Houston andHodges (1 9 7 0) , and Katk in (19 66 ) havefound trai t anxiety (A-Trait) to be unrelatedto autonomic response and performa nce mea-sures in stressful situations. Spielberger,Gorsuch, and Lushene (1970) noted that h isA-Trait scale (State-Trait Anxiety Inven-tory; STAI) predicted state anxiety (A-

    This report is based on a thesis submi t t ed by thefirst author in part ia l fulf i l lment of the requi rementsfo r the MA degree f rom th e Univers i ty of Dela-ware . T h e authors would l ike to thank the followings tudents fo r the ir work as experimenters in thepresent study: E l m e r B o u m a n , Chip Carpenter ,R ichard Dark, Scott Eiler, Jeff Gibson, TerryHerbein , John Jones, and George Laskaris.Requests for reprints should be sent to MarvinZuckerman , D e p a r t m e n t of Psychology, Univers i tyof Delaware, N e w a r k , Delaware 19711.

    State) in situations in which there w as somethreat to self-esteem but did not predict insituations in which there w as threat of painor physical harm.Some researchers have done studies inwhich specific trait i tems, referring to thefeared object or situation, were comparedwith general A-Trait scales in the predic-tion of responses to specific fear situations.Hodges and Spielberger (1966) found that asingle fear of shock item correlated .50 withchanges in heart rate made in response tothreat of shock, but there was no relationshipbetween the T MA S and the heart rate mea-sure. Mellstrom, Zuckerman, and Cicala( 1 9 7 4 ) found that i tems from a new specificA-Trait test, the Zuckerman Inventory ofPersonal Reactions (ZIPERS), and a ques-tionnaire using snake items predicted fearreactions to the exposure of a snake, whereasgeneral A-Trait measures did not. Similar re -sults have been found fo r public speaking(Lamb, 1973) and test anxiety (Sarason,1961) situations.Although th e previous studies all seem tosuggest that specific A-Trait measures arebetter predictors of situational responses thangeneral ones, a finding by Zuckerman (inpress) raised doubts about this impor tan tgeneralization. Using th e ZIPERS he foundthat general and specific trait measures wereequally predictive of A-State responses in anactual classroom exam ination si tuation. The

    83

  • 8/3/2019 Anxiety Darkness

    2/9

    84 M. MELLST ROM, JR. , G. A. CICALA, AN D M. Z U C K E R M A Nimplication of this finding is that reactionsto some anxiety-provoking situations may beequally predictable from general and specifictests. If general trait measures have predic-tive validity fo r certain situations, it is oftheoretical and practical importance to iden-tify th e situations and discover how theydiffer from other situations. The currentmovement toward situation specificity in per-sonality assessment (viz., Endler, Hunt, &Rosenstein, 1 9 6 2 ; Mischel, 1968; Zuckerman,in press) is critically dependent on theanswer to this question.The purpose of the present study was todetermine th e differential predictive power ofgeneral and specific A-Trait measures, withth e addition of two types of general A-Traitmeasures. The two types, distinguished byEndler and Shedletsky ( 1 9 7 3) , are (a) thosethat rarely specify stimulus situations, suchas the TMAS and STAI A-Trait scales, and(b) those that do specify situations, with aperson's total score across all situations rep-resenting a general A-Trait measure. Mostcriticisms of "general A-Trait" scales havebeen directed toward th e first type of mea-sure, and the comparative validity of thesecond type, relative to the first and to spe-cific A-Trait measures, has not been ade-quately established. Fear situations involvingsnakes, heights, and darkness were usedbecause they involve fairly common fearsthat are amenable to behavior measurementtechniques.A finding by Segal (1973) suggests thatsensation seeking might be related to anxietyin these situations. The Thrill and AdventureSeeking subscale of the Sensation SeekingScale (SSS; Zuckerman, 1971, 1974) mea-ures th e approach tendency as opposed to theavoidance tendency in situations involvingmoderate danger. Although the scale does notcorrelate with general trait anxiety scales,Segal (1973) showed that it does correlatewith responses to hypothetical situations inEndler et al.'s ( 1 9 6 2 ) S-R Inventory ofAnxiousness, which contains scale items withelements of physical threat from inanimatesources. Since the SSS was shown to correlatewith responses to hypothetical situations, itwas expected that it would predict re -

    sponses to real situations that contained someelements of danger.METHOD

    SubjectsSeventy-six female students enrolled in introduc-tory psychology at th e Universi ty of Delaware inth e fall of 1973 were subjects. They participated inth e study in order to f u l f i l l a course requirement .Predictor Tests

    Three of the pretests have been regarded tradi-tionally as measures of general t rai t anxiety. TheSTAI A-Trait Scale (Spielberger, Gorsuch, &Lushene, 1970 ) was originally intended as a researchtool for selecting subjects who vary in their disposi-tion to respond to psychological stress with differentlevels of state anxiety. The TMAS (Taylor , 1953)was designed to measure manifest anxiety as definedby Cameron's (1947) description of chronic anxietyreaction. It has been th e most widely used measureof A-Trait. The Neuroticism scale of the Eyscnckand Eysenck (1964) Personality Inventory ( EPI )was used as th e third measure of general A-Trai t .Prior studies have found very high correlations be-tween th e T M A S and the EPI Neuroticism scale,and both tests may alternately be considered asmeasures of "anxiety" or "ncuroticism."

    Two pretests provided indices of both general andspecific t rai t anxiety. The Geer Fear Survey Schedule(FSS) consists of self-rated fear reactions to 51commonly feared objects and si tuat ions and hasbeen s hown to be related to behavioral and self-report measures of fear (Geer, 1965). A subject 'sto ta l fear score across all 51 i tems might be takenas a general T e a r f u l n e s s or A-Tra i t measure; herscore on any one particular item can be used as aspecific A-Trai t measure . Another test that providesmeasures of bo t h general and specific A-Trait is amodified form of the ZIPERS. It consists of 12positive- and negative-affect-arousing situations inwhich subjects indicate, on a 5-point scale, thedegree to which each situation elicits in her each of13 reactions. Factor analysis of the response dimen-sions of the scale has yielded several factors in -cluding th e fear arousal factor used in the presentresearch. By summing a subject's scores across allsituations on the responses comprising th e fear fac-tor, a measure of general A-Tra i t may be derived.A subject's score on those reactions to 1 of the 12situations yields a specific A-Trai t measure. The testwas modified by substituting a "heights" and a"darkness" situation for the two "success-acceptance"situations in the usual form. The test alreadycontained a "snake" situation.To provide additional specific t rai t tests, the firstauthor devised three fear questionnaires, one for eachfear used in the study. The snake questionnaire asksth e subject to indicate, on a scale from 1 to 8, thedegree to which he/she would, or would not , liketo be in 27 situations involving snakes. This scale

  • 8/3/2019 Anxiety Darkness

    3/9

    G E N E R A L V E R S U S SPECIFIC TRAIT ANXIETY M E A S U R E Sw as used in an earl ier s tudy (Mel ls t rom et al.,1974) and appears to be a val id index of snake fear .Addi t ional fear of heights an d fear of darkness ques-t ionnaires were const ructed to be similar to the fearof snakes questionnaire. The total questionnaire iscalled th e Situation-Specific Questionnaire (SSQ).T he Thri l l and Adventure Seeking subscale of theZuckerman (1971) SSS was described in the intro-duct ion . It is a forced-choice scale. A sample itemis as fo l lows:

    43. A. I would like to try parachute jumping.B. I would never want to try j u m p i n g out ofan a i rp lane , wi th or wi thout a parachute .Fear Situations

    In the snake s i tuat ion , a 1-m long Bal l Pythonwas housed in a plywood box with a sl iding glassf ron t .The h eig ht si tuation was on the 1 6th story of adormi tory a t the univers i ty . On each floor, a 2 mX 1.5 m landing connected the main building toa fire t ower . These landings were fenced in , exceptfo r 8-c gaps where th e landing joins the buildingand tower , and were impossible to fall f r o m . Situ-ated on the ground below were four 3S X 27 cmposter boards on which th e letters L, V , T, and Zwere pain ted .In th e darkness situation, the 3 X 4 m room w aswindowless and became total ly dark when th e l ights

    were turned off . A smal l wooden box with a toggleswitch was constructed so that pushing the switchturn ed off the l ights and sim ultaneously a ctivateda timer located in an adjo in ing room, where th eexper imente r was present .Subjective Situational Fear Measures

    To measure subject ive fear in each s i tuat ion , threestate measu res were used. Behavioral measures usedar e described in the Procedure section.1. The A-Statc scale of Spielberger ct al.'s (1970 )STAI w as devised to measure th e ins tantaneousanxiety level of persons at any given t ime. It isassumed that anxiety f luctuates over t ime, and thisscale is designed to be sensitive to such changes. Itrequests subjects to describe h ow they feel "now."2 . T he A-State form of the ZIPERS (Zuckerman ,in press) w as administered with th e STAI A-Statescale to provide evidence of thei r concurrent valid-i ty . This test consists of the i tems of the 13 reactionson th e ZIPERS, with ins t ruct ions to indicate th ereac t ions being experienced "now."3. A third state measure w as provided by thefear t he rmomete r (Walk , 1956) . Subjects are askedto place a check on a 1 to 10 scale to indicate th ea m o u n t of fear they are experiencing at that m o m e n t .Procedure

    During the first 4 weeks of the semester, subjectswere adminis tered th e predictor tests. Six weekslater, subjects were scheduled for the three fearsi tuations. Each subject individually part icipated ineach fear si tuation once, and self-report , observer's

    ra t ings , and behaviora l measures of fear wereobta ined.Subjects were selected randomly, with th e restric-tion that they not reside in either of the two16-story dormitories where th e height si tuation w aslocated. It was felt that living there might affectth e subjects ' behavior in that situation. In addi-t ion, subjects were randomly assigned to the sixpossible temporal sequences of fear s i tuat ions todetermine if sequence, a "nuisance variable" (Kirk1968 ) , had any effect on fear in the situations. Tominimize such carry-over effects f rom one si tuationto the next , subjects were scheduled for no morethan one si tuation on a given day.Snake situation procedure. When th e subject ar-rived, with th e snake in view, she was seated andgiven her instructions by tape recorder. They in -s t ructed her to "pe r fo rm th e task" of approachingthe snake and lifting it up into th e air. As in th eheight and darkness si tuations, she was told t ha tif sh e could not perform th e task it was all right .After th e instructions, th e subject was adminis teredth e STAI A-State and ZIPERS State scales.A f t e r the state tests were completed, th e behav-ioral fear test w as begun. T he exper imenter told th esubject to stand on a part icu lar spot that served asa s tandard s tar t ing point for all subjects and thensaid, "You may now begin your task." T he t ime ittook the subject to touch the snake, i f she could,was recorded by the experimenter as latency. If sh ecould not touch the snake, she was given the maxi-m um latency (300 sec) . The experimenter rated thesubject 's behavior on a 1-12 ra t ing form describinga cont inuum of behavior f rom 12 ("cannot approachsnake closer than 6 feet") to 1 ("picks up snake andholds for 30 sec"). The subject 's score on this taskw as called th e task score. T he experimenter alsorecorded th e snake's behavior, another nuisancevariable, during th e subject 's approach and then gaveth e subject a fear t he rmomete r t o fill out as shestood at the locat ion of her closest approach to thesnake. After th e subjec t departed , the experimenterrated her on a 1-7 scale of anxiousness, which wascalled the observer 's ra t ing .Height situation procedure. The taped instructionstold the subject that her "task" was to step out ontoth e landing, record the let ters that were on theground on a piece of paper , till out the question-nai re ( the fear t h e r m o m e t e r ) , and re turn in to thebuilding. T he sub j ec t w as then adminis tered th eSTAI A-State and ZIPERS State scales.

    After completion of the state scales, the behavioralheights test w as begun . T h e subject started f rom astandard location. T he latency was the lime it tookfo r her to reach another standard location and beginlooking down f rom the landing. The distance ofher toes, in IS-cm intervals , from th e edge of thelanding comprised her task score. The total amountof t ime spent by the subject on the landing w asrecorded as total t ime . After th e subject left, theexper imente r m a d e the 1-7 observer's rat ing andverified t ha t she had correct ly recorded th e lettersthat were on th e g r o u n d .

  • 8/3/2019 Anxiety Darkness

    4/9

    86 M. MELLSTROM, JR., G. A. C I C A L A , A N D M. Z U C K E R M A NIt w as t hough t that we a the r condit ions m i g h tinfluence the subjects' behavior on the landing sincei t was open, except fo r metal gri l lwork, on twosides. To observe the effect of this nuisance variab le ,the experimenter recorded the outdoor temperature ,precipitation, and wind conditions, each on separatescales of intensity, after each subject departed.Darkness situation procedure. Upon arrival , sub-jects heard taped instructions tel l ing them that theirtask was to sit in the adjoining room, t u rn off thelight, and remain in the darkness until the experi-men te r re tu rned . T h e STAI A-State and ZIPERSState scales were administered at the end of the tape.A f t e r complet ing th e scales, the darkness fear testw as in i t ia ted. T he subjec t was lef t alone in the

    room, with the experimenter in the adjoining room.The t ime f rom the door being closed to the subject 'sturn ing off the l ight w as recorded as the la tency .The length of t ime the subject spent in darkness,up to a l imit of 300 sec, w as recorded as to ta ltime. At the end of the behaviora l test , th e subjectcompleted the fear t h e r m o m e t e r , describing the fearshe had ex perienced in the darkness. After th esubject lef t , the experimenter made th e observer'sra t ing of her anxiety in the s i tuat ion .

    RESULTSThere was a truncation of range in thebehavioral measures of the heights (taskscore) and darkness (total t ime) situations.In the heights test, all but 10 subjects re -ceived the lowest possible fear score, whereasin. th e darkness test, all but 4 subjects re -mained in darkness the maximum time,receiving the lowest possible fear scores. Asa consequence of the restricted range, thecorrelations of these variables with the pre-dictors were expected to be low. Fortunately,th e other measures taken in the fear situa-

    tions were not as seriously restricted in range,permitting higher correlations with the pre-dictors. It should also be noted that th e totaltime measures taken in the darkness andheights situations were subjected to a recip-rocal transformation so that high scores on allfear measures indicated greater fear.Only one of the recorded nuisance vari-ables affected subjects' behavior in the situa-tions. A correlation of .28 ( d f 75, p < .01)was found betwe en precipitation and taskscore in the heights situation, so the relationsbetween th e task score measure in the heightssituation and all other measures were assessedby partial correlations, thereby removing th eeffect of the precipitation variable.

    Relations Among th e PredictorsHigh intercorrelations between the TMAS,STAI A-Trait, and EPI Neuroticism scales

    indicated that there is considerable overlapin what is being measured by these threetests. The correlations fo r these scales rangedfrom .73 to .86. Since they intercorrelated sohighly, these three tests will be called, forpurpose of discussion, the "ncuroticism tests."The subjects' total scores on the FSS andZIPERS, which will be called th e "omnibus"measures, correlated more highly with eachother (r = .68) than with the neuroticismtests, suggesting that tw o different forms ofgeneral A-Trait were measured. Though basi-cally unrelated to the neuroticism tests, th eThrill and Adventure Seeking Scale showedmoderate ly high negative correlations withthe FSS and ZIPERS.

    The intercorrelations of the specific trai tmeasures of the FSS, ZIPERS, and SSQ (notshown in these tables) revealed good conver-gent validity (Campbell & Fiske, 1959) fo ral l three fears. The correlations between mea-sures of the same specific fear from differenttests we re high, ran ging from .48 to .78.

    Thus, the various A-Trait m easures fellinto three classes: (a) neuroticism measures,consisting of the TMAS, STAI A-Trait, andEPI Neuroticism scales; (b) omnibus mea-sures, comprised by total scores on the FSSand ZIPERS; and (c) specific A-Trait mea-sures, represented by the specific trait mea-sures of the FSS, ZIPERS, and SSQ. Sincecomparisons among the individual measureswere not as impor tan t as comparisons amongclasses of measures, th e standard scores ofth e measures comprising a given class weresumme d to provide a composite measurerepresenting that class.

    The intercorrelations among these compos-ites and the Thrill and Ad venture SeekingScale, the fourth predictor, are shown inTable 1. The omnibus composite measureshowed moderate to strong correlations withall the other predictors, probably as a resultof its "general" nature. The correlations be-tween th e neuroticism composite and most ofth e other predictors were low, whereas th eThrill and Adventure Seeking Scale correlatedhighly with everything but neuroticism.

  • 8/3/2019 Anxiety Darkness

    5/9

    G E NE R A L V E R S U S S P E C IF I C T R A I T A NX I E T Y ME A S U R E S 87Though al l three specific A-Trait compositeswere significantly correlated, th e relation be-tween th e composites of heights and darknessfear was the strongest.Relations Among the SituationalFear Measures

    Overall, the response measures taken ineach situation intercorrelated quite highly.STAI A-State and ZIPERS-Sta te correlatedbetween .43 and .49 with the main behavioralmeasure in each situation and between .48and .64 with the clinical rating. The feart hermometer correlated between .53 and .62with th e behavioral measures and around .7 0with STAI A-State and ZIPERS-State . Thelast tw o measures correlated .86, .85, and .77in the snake, heights, and darkness situa-tions, respectively. Some of the situationalmeasures, fo r example, total t ime, simply didnot correlate with the other measures. How-ever, most of the response m easures did inter-correlate highly within each situation, in-creasing confidence in their construct validityas measures of fear or anxiety states.Somewhat similar to the findings for thespecific trait predictors, responses in theheights and darkness situations were corre-lated, whereas responses in these situations

    T A B L E 1I N T J O R C O R R E I - A T I O N S A M O N G T H E FOUR C L A S S E SP R E D I C T O R M E A S Q R E S

    ScaleNe ur o t i c i s mSSS-TASSnake A-TraitH e i g h t A - T r a i tDarkness A - T i a i t

    O m n i -bus.45*-.57*.55*.6.!*.74*

    N c u r o t -icism-.21.l.i.29*.2 0

    SSS-TAS

    -..14*-.57*-- .56*

    SnakeA - T i a i t

    . .T O *.42*

    H e i g h tA - T r a i t

    .59*Ntilt:. O m n i b u s Summe d s t a nda r d s c o r e s of the total scoreson the Geer Fear Sur ve y S ch ed u l e and Z u c k e r n i a n I n v e n t o r y ofPersonal React ions . N e u ro t i e i sn i S u m m e d s t a n d a i d scores onth e T a y l o r M a n i fe s t A n x i e t y Seale, State- Trait A n x i e t y I n v e n -t o r y Tiait Scale, a n d K y s e n c k P e r s o n al i ty I n v e n t o r y N e u i o t -ic ism scale. SSS-TAS = T h t i l l an d A dve n t u r e Se e k i ng subscaleof th e Se ns a t i on Se e k i ng Scale. Snake A - T t a i t S u m m e ds t a nda r d scores of the s na ke A - T r a i t mea su res of the Geer FearS u r v e y Sc he du l e , X u c k c i m a n I n v e n t o r y o f Personal Re a c t i ons ,an d SSS. Hei g h t s an d da r kne s s A - T i a i t c omp os i t e s w e i e s imi -la r ly der ived.*J > < .01.

    did not correlate with responses in the snakesituation.Relations Between th e Predictors an dSituational Fear Measures

    Table 2 shows th e correlations betweenth e four classes of predictor measures and thesituational fear measures. The validity co-efficients of the specific composites representth e relations between th e situational fearmeasures and the specific com posite designedto predict that fear. To pro vid e a single mea-

    TABLE 2C O R R E L A T I O N S A N D H E T A W R I G H T S F O R F O U R C L A S S E S O F P R E D I C T O R S

    Si t ua t i ona l f e a rme a s u r eSnake s i t u a t i o n

    A-StateFear t h e r m o m e t e rO bs e r ve r s ' r a t i ngBe ha v i o r a l i n d exH e i gh t s s i t ua t i on

    A-StateFear t h e r m o m e t e rO bs e r ve r s r a t i ngBe ha v i o r a l i n d e xDarkness s i t ua t i onA - S t a t cFear t h e r m o m e t e rO bs e r ve l s ' r a t i ngBe ha v i o r a l i nde x% Signi f i cant r" s

    I : O m n i b u s

    ,42**(.00)..18** ( .04).42** ( .00).23*(-.22)

    .2S*(-.10.25* (-.24)

    .16(-.19),29**( - - . ! (> )

    ..18**(-.02).4 !**(-.02).28*(.04).09 (.00)50%

    PI I :Neuiot-icism

    ,25*(.I6).1.1 ( .04).15 (.05).04 ( .03)

    .3/)** (.32"), 2 7 * ( . 2 1 ).12 (.06).15 (.04)

    ..17**(.2H").25**(.14*).08(-.01).07 (.0.1)25%

    r cd i c t o rIII : Specific

    .61**(.57),62**(.59).62** (.50").64**(.73). 29 * * ( . I3 ), 4X**( .48), 3 7**( .4 M.44** (.34").35** ( .13),49**( .2 '>).30** ( . 1 2 ).06(-.09)92%

    SSS-TAS

    -.28**(-.05)-.22* ( .00).36**(-.l6)- . 2 0 ( . 0 8 )

    ~.3t**(-.23)-,34**(-.t6)-.24*(-.10)-.**(-.27")

    -.4.1**(-.32'0-.45**-(.28")-.36**(-.28)-.21(-.25)67%

    /-test c ompa r i s on\vs.I l l I I vs.I l l

    .s I I I > I II I > III > I I I I I > I I I I > I I1 > 1 1 I I I > I 1 1 1 > I IH J I I I > 1 I I I > I I

    I I I > [I I I > I

    111 > I I

    .6.1

    .62

    .64

    .66

    .44

    .53

    .39

    .49

    .53

    .53

    .58

    .22

    Note. The values in p aren t h eses are beta weights . O m n i b u s = Summe d s t a nda r d s c o r e s of t h e total scores on the Geer Fear S u rv eyS ch ed u l e and Z u c k e i m a n I n v e n t o i y of Pe r s ona l Re a c t i ons . Ne u r o t i c i s m S u m m e d s t a n d a r d s c o re s ot i th e T a y l o r M a n i f e s t A n x i -e ty Scale, State-Trait A nxi e t y Inve n t o r y T r a i t Sc a l e , a n d Eys e nck Pe r s ona l i t y Inv e n t o i y N e u ro t i e i sn i scale. Speci f i c = F o r eachof the three fears , sum m ed s tandard scores on the specific A-Trai t m easures of the Geer Fear S u r v e y Schedule , Zucke i m a n I n v e n t o i yof Personal React ions , and SSQ. SSS-TAS = T h r i l l and A d ve n t u r e Se e k ing s ubs c a le of the Sensat ion Seeking Scale.* T h e regression o f the cr i t e r ion var iable on thi s predic tor is s igni f icant (p < .05) a f t e r t a k i n g t h e effect of the o t h er p r ed i c t o r sinto account.

  • 8/3/2019 Anxiety Darkness

    6/9

    88 M . M E L L S T R O M , J R . , G. A. C I C A L A , A N D M. Z U C K E R M A Nsure of A-State in each s i tuat ion , th e s tandardscores of the STAI A-State and ZIPERS-State measures were summed for each situa-tion. In the snake situation, the two behav-ioral measures were similarly combined. Thebehavioral indices in the heights and darknesssituations were task score and total t ime, re -spectively. The latency and total t ime mea-sures obtained in the heights situation weredeleted because they did not correlate witheither the predictors or other sitnational mea-sures. The latency measure of the darknesssituation w as similarly deleted.To compare th e validity coefficients of thethree types of A-Trait measures a t test forth e significance of the difference between tw ocorrelation coefficients fo r correlated samples(Ferguson, 1971, p. 171) was used. Theright-hand portion of the table shows theresul ts of these comparisons.For the snake situation, the f- test com-parisons show that the specific predictors cor-related significantly higher with all foursituational m easur es than either the omnibusor neuroticism measures . I n addi t ion , for thefear t h e r m o m e t e r and observer's rating, th eom ni bus composi te had significantly highercorrelations than the neurot icism composi te ,which had the lowest validity coefficients ofall predictors for all situational measures.For the heights s i tuat ion , the specific com-posite predicted the behavioral index signifi-cantly better than the neurot icism composi le ,and it also predicted the fear the rmomete rand observer's rating significantly be t t e r thanthe omnibus composi te . No other differenceswere signif icant, although the specific m ea-sures had the highest validity coefficients forall s i tuat ional measures but A-State. T heomnibus and neuroticism measures seemed tobe equal ly predict ive of the measures inthis situation.

    in th e darkness s i tua t ion , no differencesbetween th e A-Trait composites were signifi-cant , and no pat te rns of superiori ty werereadi ly apparent . However, an overview ofth e correlat ions for all three situations revealssome impor tan t differences. First, the onlysi tuat ional measures the neuroticism compos-it e pred icted were self-repo rts. In contrast,th e specific predictors showed signif icant cor-

    relations with all types of situational mea-sures. The percentages of significant correla-tions (p < .01) for each type of measure,shown in the bottom row of Table 2, showthat 9 2% o f the validity coefficients for thespecific A-Trait measures were larger thanwould be expected by chance. T he corre-sponding percentage for the neuroticism com-posite was only 17%. The specific measuresshowed significantly higher validity coeffi-cients than the omnibus measures for thesnakes and heights situations but not for thedarkness situation. T he overall validity seemssomewhat better for the specific measures, asdemonstrated by the percentages of signifi-cant correlations ( 9 2 % versus 5 8 % ) , bu t thisdifference is not nearly as great as that fo rthe specific versus neuroticism comparison.Only for the snake situation did the omnibusmeasures show significantly higher correla-tions than th e neuroticism measures. How-ever, in all situations the omnibus composite,in contrast to the neuroticism composite, cor-related with other situational measures, inaddi t ion to A-State.

    Table 2 also shows that th e Thrill andAdventure Seeking Scale, which does not mea-sure anxiety directly, shows considerable va-lidity in the three situations. It seems thatth e val idi ty of the scale was as good as anyother predictor, with the exception of thespecific A-Trai t measures in the snake s i tua-tion. To shed more light on the compositionof each situational measure and the relativecontributions of the predictors, 12 stepwiscmultiple regression analyses using the fourtypes of pretests as predictor variables andeach situational measure as a criterion variablewere performed.For the measures in the snake situation,the beta weights of the specific A-Trai t pre-dic tors were universally the largest, suggest-in g that anxiety and fear in this situationwere caused primari ly by fear of snakes, thuspermi t t ing the specific composite to be mostpredictive. For the heights situation mea-sures, th e results were not as simple. For theA-State criterion variable, the beta weight ofthe neuroticism composite was the only oneto reach significance, although the Thrill andA d v e n t u r e Seeking scale approached signifi-

  • 8/3/2019 Anxiety Darkness

    7/9

    G E N E R A L VERSUS SPECIFIC TRAIT A N X I E T Y M E A S U R E S 89cance, suggesting that these traits, not fearof heights, were related to A-State. For thebehavioral index, th e weight associated withth e Thrill and Adventure Seeking scale w asas large as that of the specific A-Traitcomposite, and both were significant.In the darkness situation, th e weights forth e behavioral measure showed that th e Thrilland Adventure Seeking scale was the variablethat best explained th e subjects' behavior.Subjects' levels of A-State in this situationmay be explained by Thrill and AdventureSeeking and netiroticism, with a small compo-nent (f t .13) contr ibuted by fear of dark-ness. For all three situa tion s, the w eights ofth e omnibus composite were near zero ornegative, suggesting that when the otherthree predictor variables are in the regressionequation the omnibus measures provide onlyredundant informat ion .

    DISCUSSIONThe results of the present study suggest

    that measures of specific A-T rait perm itmore accurate prediction of behavior in spe-cific situations than do measures of generalA-Trait. Particularly supportive of this inter-pretation were the findings for the snake situ-ation, in which for all four situational me a -sures, th e correlations of the specific com-posites were significantly higher than boththe neuroticism and omnibus general A-Traitmeasures. For the heights situation, threesignificant differences were found between th evalidity coefficients of the specific and gen-eral predictors. Also, th e percentage of sig-nificant correlations for the specific A-Traitcomposite measures was considerably higherthan that for either general composite. Al-though the re was little doub t abou t the supe-riori ty of the specific measures re la t ive tothe neurotic ism measures , th e difference inpredictive power between th e specific andomnibus measures was not as great.

    Comparisons between the neuroticism andomnibus composite measures revealed slightlygreater validity for the latter measures. Forthe snake situation, tw o differences betweenvalidity coefficients were significant, andacross all three situations; th e neuroticismmeasures predicted self-report measures only,

    whereas the omnibus measures predicted othermeasures as well.

    Thus, for the situations of the presentstudy, th e specific A-T rait m easures showedth e most validity, th e neurotic ism measuresthe least validity, and the omnibus measuresintermediate validity. However, the fact thatfor the heights and especially the darknesssituations both types of general measureswere nearly as valid as the specific ones,demonstra tes th e potential usefulness of thegeneral measures. It is theoretically importantto determine which aspects of these situa-t ions perm it the general measures to be nearlyas predictive as the specific ones.O ne such aspect of these situations may bethreat to self -esteem. Spielberger ( 1 9 6 6 ) andSpielberger et al. (1970) have pointed outthat tests such as the T M A S and th e ST A I -Trait scale are maximal ly predictive of si tu-ational anxiety when the situation poses somethreat to self -esteem. I f the heights and dark-ness situations of the present study werem o r e th rea ten ing to self-esteem than thesnake situation, this would explain the near-equal accuracy of the neurotic ism and specifictrait composites in predicting responses tothose situations. Since th e om nibus compositecorrelated .45 with th e neuro tic ism composite,th e omnibus would also be predictive ins i tuat ions posing threat to self-esteem. Theheights and darkness s i tuat ions could haveposed more threat to self-esteem if subjectsfelt .the tasks involved were so simple thatinability to per form them indicated that thesubject w as childish, foolish, or t imid . Thisattitude could easily develop in the heightsand darkness situations because the taskswere very simple and involved absolutely norisk of physical pain or injury. Support forsuch a view comes from the fac t that only4 subjects could not complete the darknesstask and only 10 could not maximal ly com-plete the heights task. In the snake si tuat ion,however, subjects m ay have experienced littleor no threat to self-esteem because of an atti-tude that the task involved was difficult anddangerous enough that no one, includingthemselves, really expected them to per formit . The cultural stereotype of women's fear-fulness of snakes m ay also have helped pre-

  • 8/3/2019 Anxiety Darkness

    8/9

    90 M , M E L L S T R O M , JR., G. A. CICALA, A N D M . Z U C K E R M A Nvent loss of self-esteem fo r failure to performin this situation. The fact that only half ofth e subjects touched th e snake and evenfewer picked it up suggests that many sub-jects did find the task difficult to per form.Thus, we might attribute anxiety arousedin th e snake situation primarily to the pres-ence of the snake, whereas anxiety in theheights and darkness situations m ay havebeen a product of threat to personal adequacyand a fear of heights or darkness. This ideais supported by the relative sizes of thebeta weights of the neuroticism and specificA-Trait composites for the three situations.

    This interpretation helps explain the find-ings of the present study. First, assumingdifferential threat to self-esteem in the fearsituations explains why the general measureswere almost as predictive as the specific mea-sures in the heights and darkness situations.It also suggests that if the tasks of theheights and darkness situations had beenmore difficult, th e differences in predictivepower between the general and specific pre-dictors m ight have been as great as was foundfor th e snake situation. Second, th e in terpre-tation explains why the responses in theheights and darkness situations were corre-l a ted : They both threatened personal ade-quacy. Third , the interpretation explains whythe neuroticism com posite predicted the self-reports but not the behavioral measures inth e heights and darkness situations. The be-havioral measures, by their very nature, wereconfined to the specific fear of heights ordarkness, whereas th e self-repo rt A-State testsmeasured any anxiety being experienced byth e subject, including that aroused by "egothreat ."Another impor tan t finding was the surpris-ing degree of predictive power of the Thrilland Adventure Seeking scale. The predictivepower suggests that persons scoring high onthis scale are less likely to report anxiety inanxiety-inducing situations. The multiple re -gression analyses corroborate this finding andsuggest that fear of a given situation is notth e only variable controlling a person's ap-proach toward that situation; thrill andadventure seeking is also important. Conceiv-ably, someone high on this trait m ay even

    approach rather than avoid a feared situationto obtain arousal. Thus, much behavior usu-ally thought to be under the control of fearor anxiety, like that in the present fear situ-ations, may also be a function of other vari-ables such as sensation seeking. Isolationof the effects of these other variables en-hances our ability to describe and understandanxiety phenomena.R E F E R E N C E S

    Allport, G. W. Personality, a psychological interpre-tation. New York : Hol t , 1937.Cameron , N. The psychology oj behavior disorders:A bio-social interpretation. Boston: HoughtonMifflin, 1947.Campbell, D. T., & Fiske, D. W. Convergent anddiscriminant validation by the mul t i t r a i t -mu l t i -mcthod matr ix . Psychological Bulletin, 19S9, 5 6,81-105.Endler, N. S., Hunt , J. McV., & Rosenstcin, A. J.An S-R inventory of anxiousness. PsychologicalMonographs, 1962 , 7(5(17, Whole No. 536) .Endler, N. S. & Shcdletsky, R. Trait versus stateanxiety, authori tar ianism, and ego-threat versusphysical threat. Canadian Journal oj BehaviouralScience, 1973, 5, 347-361.

    Eysenck, H. J., & Eysenck, S. B. G. An improvedshort quest ionnaire fo r the measuremen t of extra-version an d neurot ic ism. Lije Sciences, 1964, 3,1103-1109.Ferguson, G. A. Statistical analysis in psychologyan d education. New York: McGraw-Hill , 1971.Geer, J. H. The development of a scale to measurefear. Behaviour Research an d Therapy, 1965, 3,45-53.Hodges, W . F., & Spielbcrger, C. D. The effects ofth rea t of shock on heart ra te for subjects whodiffer in manifes t anxicy an d fear of shock.Psychopkysiology, 1966, Z, 287-294.Hodges, W . F., & Spielberger, C. D. Digit span:An indicant of trait or state anxiety? Journal o fConsulting and Clinical Psychology, 1969, 33,430-434.I touston , B. K ., & Hodges, W. F . Siluational denialan d performance under stress. Journal o f Person-ality an d Social Psychology, 1970, 16, 726-730.Kalkirt , E. S. The relationship between a measureof transitory anxiety and spontaneous autonomicactivity . Journal o f A bno r ma l Psychology, 1966,71, 142-146.Kirk , R. E. Experimental design: Procedures for thebehavioral sciences. Bclmont , Calif . : Brooks /Cole,1968.Lamb, D. H. The effects of two stressors on stateanxiety fo r students w ho differ in trait anxiety.Journal of Research in Personality, 1973, 7, 116-126.Mells trom, M. Zuckerman , M., & Cicala, G. A.Anxiety : General versus specific trail in predicting

  • 8/3/2019 Anxiety Darkness

    9/9

    G E N E R A L V E R S U S SPECIFIC TRAIT A N X I E T Y M E A S U R E S 91.fear of smite. Psychological Reports, 1974, .75,317-318.Mischel, W. Personali ty an d assessment . N ew York:Wiley, 1968.

    Sarason, I . G. Test anxiety and th e intellectual per-f o r man ce of college students . Journal o f Educa-t ional Psychology, 1961, 5 2, 201-206.Segal, S., Sensation seeking and a n x i e t y : Assessment

    of responses to specific s t imulus s i tuat ions . Journalo f Consul t ing an d Clinical Psychology, 1973 , 41,135-138.

    Spiclberger, C. D. Theory and research on anxiety.In C. D. Spic lberger (Ed . ) , Anxiety an d behav-ior . New Yo rk : Academ ic Press , 1966.Spielberger, C. D., Gorsuch, R. L ., & Lushcnc , R. E.STAI manual fo r th e State-Trait Anxiety Inven-tory. Palo Al to , Calif . : Consulting PsychologistsPress , 1970.

    Taylor , J. A. A persona l i ty scale of manifes t anx-iety. Journal o f Abnormal an d Social Psychology,1953, 48, 28S-290.Walk , R. D. Self - rat ings of fear in a f ear - invok ings i tuat ion. Journal o f Abnormal an d Social Psy-chology, 1956, 52 , 171-178.Z u c k e r m a n , M . Dimens ions of sensation seeking.Journal o f Consult ing an d Clinical Psychology,1971, 36 , 45-S2.Z uc ke rma n , M. The sensation seeking m o t i v e . P r o g -ress in Experimental Personali ty Research, 1974 ,7, 79-148.Z uc ke rma n , M. State and t r a i l anxiety: New ap-proaches to assessment wi th implications formo t ive m easures . In M. Zuck erm an & C . D.Spielberger ( E d s . ) , Emotions an d anxiety: N e wconcepts, methods, and applications, in press.(Rece ived May 27, 1 9 7 5 )