)V1ESI.J RETURNING MATERIALS:
PIace in book drop to
LIBRARJES remove this checkout from
.‘Innr‘lalL. your record. FINES wiII

 

 

 

' be charged if book is
returned after the date
stamped below.

 

 

o 0*-
JUN 0 212005

 

 

THECDNSIRJCTIWOFASQIETOMEASURECINFIICTOVEREXERESSWG

MON: ‘IHEAMBIVAIENCEOVEREDGRESSIQJ QIESTIGINAIRE

By
Laura King

A‘IHESIS
£ktndtted.to
Nfichigan.$tate‘University
in l' 1 .
fer the degree of

MAS'I'EROFARIS

Department of Psychology
1989

Am

'IHEQIS'IHICTICNOFASCEIETOMEASURECINFIICTOVEREDCPRIBSDIG
WITCH: 'IHEAMBIVAIENCEOVEREDG’RESSIONGJESTIONNAIRE

By

laura King
Msflfliesemminedtheinportanceofanbivalerneabaxtexpmsing
amticn. A questionnaire measure of anbivalence over expression
(Am) was designed and administered to 292 subjects along with the
Emotional Expressiveness Questionnaire (E512), measures of social
desirability and intense anbivalence. Factor analysis of the AH:
indicated that it is unidimensicnal . The Am was negatively
correlated with the Em and social desirability, and positively
correlated with intense anbivalence. Wcmen scored significantly
higherthanmmboththeAEDarfltheEm. Asecondsampleof 48
albjects participated in a 21-day study of mood and health.
Ecprassivenass was positively correlated with sane measures of
well-being and with daily negative affect. 'Ihe AR) was positively
correlated with a several indices of psychological distress. Neither
theEBertheAmwascorrelatedwimdailysynptanreports or
umber of different illnesses. Results surport the contention that
espressivenessisminsulatimagairsttheadverseconsequencs
traditionally associated with imcpmssivernss.

Dedicatedtomypam1ts,RwertC.andIdaS.King

iii

'Iheautlnrexpressasgratitudetothematbersofmyccmnitte, Dr.'s
JoelAmmoffarflSusanFrankfortheirhelpfulcaxmentscnthis
research. VeryspecialtharflcstoDr.RobertEhmsforhis
assistarnemthisthesisarflhiscmtimiingaxpportofmycareeras
arasearcher.

iv

MECFGNI'ENI’S

ListofTabl95 .............. ....vi
Listof Figures ...... ...........vii
Traditional Views of Emotimandmpmsim 9
Mimardfibcpmsicn: ADynamic Interactimll
Eb<pr9551m= ASccial Issuels
'IheSocialization of Ecpr95s10nl6
Emotional Ebcpr95$imarriArmsal= AnInverseRelaticnship.....18
Stmmary .............. ..38
Study1:FactorAnalysis
Developent oftheQue5ti0nnaire............. ............ ..40
Prediction ............ ........41
Method. ...... .42
R95ultsandDiswssion.. ..... ............... ..43
Studyz: Application
Predictions .................... 67
mthod... ......... ...................... .................. 68
ResultsandDiscussmn ............... 72
GeneralDiscussicn ......... . ...... 87
Listof Refereno95 ...... ........96

Amendic95...” ............................................... 104

Table
Table
Table
Table
Table
Table
Table
Table
Table
Table
Table
Table
Table

IISTOF'IRBIE

1.... ........ .0. 0000000000000000 0 OOOOOOOOOOOOOOOOOOOOOOOO 45
2.0......... .................. 0.... ...................... 49

3.....O..........0...................................0...53
4....0.........0...............0..00.....0... ............ 56
5..............................0...0........ ............. 58

6................0.......0.....0.......0...0.......000...60

7............ ..... .......................................62
8............ ............ ............. ................... 73
9.............. ..... ........................ ........... ..74
10............................... ........................ 77

11. . . . . . . . ............................................... 78
12 . . . ..... . .............................................. 80
13 ....................................................... 81

vi

Figure 1.

Figure 2.
Figure 3.
Figure 4.

LIST OF FIGJRE

CHD risk and expr9551ven955 (Fran Freidman, Hall, and

Harris, 1984). ...... 3o
Diagnosis ofT3rpeAbased0nexpressiven95s...... ....... 30
Cells II and IV are at risk for illn955.. ............ ...31
Measurement model with path coefficierrts.. .............. 54

vii

mmowcnou

Shieldshasomrtedthat "asaallmremareanbivaleltabmt
amticn (1987, p. 231)." This ciltural an'bivalence is apparent in many
cannon expressions which subtly denigrate emotion and emotional
display. Expre5sion of auction is cannonly termed a "giving in" to
passion. We canon-mt that disaster is avoided when "cooler heads
prevail". Mistakes occur when sentimentality ru195 and we think with
"cur hearts iratead of car heads". Hodtsdlild (1983) has concluded
that action is often portrayed as an inpediment to societal progr955.
Along side this inplicit disdain for auction are societal expectations
forardencwragenentoftheexpressimofawtionaswellasthe
assmptim that actions mast be "let out". There is a belief that
expressionisthehealthyendtoanotiaaal response. Onenustbe "in
touch" with one's true feelings, and "follow one's heart." This
eq1ivocal cultural view is of particular significance to psychology,
for this anbivalenoe inpinges on a microcosmic level of experience.

Mimarxiamtional eaqara5$imareisa195whid1pervadenany
aspects of hulnan eacperiaace. The enotional life of an individual is
daviously inportant to his or her psychological well-being. Emotion
ard expression also contribute to physical well-being (e.g. , for
exanple, Derogatis, Abeloff, and Melisaratos, 1979) and have social
inplications as well (e.g., Tavris, 1984).

While popular alphasis on expressim may be dwindling sanewhat,
the assunptim lingers that expression provid95 an insulation against

1

2
psychological and physical ailnent. One nust let one's feelings out or
bear the brunt of a neurosis or an ulcer: emotions will find
expr955ion, one way or another. Freudian repre55ion is an example of
the early bias towards expr95810n for its own sake. Freud (1917/1977)
portrayed repressed inpulse5 as constantly pushing for expr955ion.
AccordingtoFreud, inpllse5m11devenmally findameanstomanif95t
ttmselveseitherintheiractualcontentorinneurotic symptoms.
"Catharsis"wasthemeansbywhid1thisawtional energycouldbe
released. later psychotherapeutic techniqn5 have also emphasized the
inportanoe of abreactive experienc95 (e.g., Rogers, 1961) . In th95e
techniqu95 we can see the inplicit assrmption that emotions nust be
ventillatedinordertoreleasepentupenergy.

Althmgh mid apnsized the psychological ills that could be
visited upon repre5$ive individuals, current assuuptions about the
benefits of expre55ion aphasize the adverse physical consequenc95 of
irhibition. The acknowledgnent of the relationship between mind and
bodyhasledtotheassertionthat irhibitedemartionsneedmtbe
expre56edthroughmental illness. Instead, smirinpulsescanfind
release intheautormicnervoussystenardinvariws psychosomatic
disorders. The inverse relationQIip between anotional expr955ion and
autonanic activity, which has been supported to some extent by r95earch
to be reviewed later, represents a return to Freud's homeostatic model
of anctim, substituting psychosanatic 11111955 for neurosis.

'mis tensim-release view of auction, while it is inplicit in many
cannon assmptias about expression, is certainly ripe for criticism,
ignoring as it does, individual differences in expr953iven95s styles
and the social functions and inplications of emotional expression.

3
Bcduthefiafliancorneptofrepressimandthemrerecentlypmposed
camection between emotional expre5510n and physical illn955, arphasize
an inevitability of expression that we are beginning to realize was and
is unwarranted. It may be that sane individuals can be inexpressive
without suffering any adverse consequal095 (Permebaker, 1985) . In
addition, the nature of anctional "energy," itself, may differ across
irdividuals with regard to tendenci95 to feel intense emotion (larsen &
Diener, 1987) , to engage in different kinch of emotional control (Roger
& Nesshoever, 1987) as well as to e.xpr95$ or inhibit emotion (Bell,
1978) . Also, as we shall see later in this discussion, physical
synptars have been associated with both inhibition and overt expr95sion
(e.g., Friedman, Hall and Harris, 1984).

An examination of emotional expression from a sociologically
informed perspective (e.g., Hochshild, 1983; Baumeister & Cooper, 1981)
reveals that expression is a cmplex task. Kluckhohn and Murray (1953)
cannented that "the most difficult and painful function of personality
(is) that of acccmnodating its expressions, needs, choice of
goal-objects, methods, ardtime—programstothepattems thatare
conventionally sanctioned by society (p. 19) ." Each individual
functions within his or her own individually determined constraints
whicharethemselvestheproduct ofandwhichexistwithin larger
social norms. Toassumea feltemotimmstor shouldbeexpr95$ed is
to ignore thousands of years of evolution for our speci95, hundreds of
years of civilization for our society, and tens of years of
socialization for each individual .

Thecrmcofthefollcwingargmrtisthattherearefactorsthat
intervene between auction states and emotional expression which may

4

influence the potentially pathogenic quality of inexpressiven955.
'Ihe5e factors include individuals’ choices about whether or not an
auction should be expre53ed. These choices will be influenced by
individual styles as well as social expectatiors regarding emotional
expression. The following diswssion and inve5tigation will examine
theissueofanotional expr'95simapartfruntheassmptionthat
expression of auction is always a positive action. Our enphasis will
be not on expr95sim per se but rather upon individual expre55iven95$
styles: social evaluation of and individual canfort within these
sty195 and their inplications for psychological and physical
well-being. Mcaweptswhidlareasentialtothepment
inve5tigation are personal strivings and display rul95.

was (1986) has described the role of ”personal strivings" in
personality and subjective well-being. A personal striving refers to
"what a person is characteristically trying to do" (1986, p. 1059).
mans has found that anbivalalce about one's strivings is associated
with low subjective well-being, especially high negative affect. He
has suggested that a preporrlerance of his subjects’ anbivalent personal
strivings, i.e., those strivings whose adliaraterrt would cause
Winessasvell ashappiness, cmcemedanctimnanagenent.
Anbivalence over those strivings which deal with erotiors is sugge5ted
hereasaninportantfactorintherelatiarshipthathasbeen
danonstrated to exist between emotional expr95$ion and well-being.
Personal strivings are particularly suitable for a study of ambivalence
cverexpre55ion sincewhatapersmstrivestodoneybeverydifferent
fran what a person actually does (Ems, 1986) . Within this
framework, we can examine the relationship of anbivalence about

5
expression apart fran the basic issue of expression itself. Meehl
(1964, p. 10) defined intense aubivalence as "the existence of
sinultanecus or rapidly interchangeable positive and negative feelings
towardthesameobjectoractivity, withtheaddedprcvisothatboth
the positive and negative feelings be strong." Ambivalence can also be
tmught of as an approach-avoidance conflict—wanting sauething but at
thesametimenctwantingit. Withregardtopersonal strivings,
individuals may have goals about auctional expression and be an'bivalent
abaltthesegoals—towishtoadlievethauwhile atthe simultaneously
acknowledging that achievement iuplies negative as well as positive
consequences.

It may be only fitting that expre55iven955 be examined through a
goal framework, as Bamueister and Tice (1987) state that emotions are
caucernedwiththesamegoalsastheself. Emotiorsareagaugeofcur
progressincurgoalpursuits. Wefeel anger, joy, etc., inr95ponse
to the frustration or fulfillment of our d95ir95. Not only do auctions
and euctional expression arise out of goals, emotion and expr95$ion can
beusedintheserviceofgoalsaswell. Personalstrivingswhich
pertaintoamtion revealmxixabcutthenatureof individual auction
managemlt. Exauples of action-related personal strivings collected
in a study by more and King (1988) include, "control my taper," "be
in touch with my feelings," "almys amear cool," "not let my auctions
take over," "hold my feelings in if they will hurt others," "always
wear a smile on my face," "be honest and open about my feelings," "let
my anger out before it all builds up inside me," and "learn to expr95s
my love toward my family better." The juxtaposition of these strivings

6

provides a pointed denustration of the tramlation of cultural
aubivalence into individual lives. Evident in th95e strivings is the
knowledge that auction must be controlled, that expression inpli95
vulnerability but also that auction should be hon95tly expressed.
The5e strivings exemplify the fact that individuals possa55
expr958iven95sgoalsandseektoadlievecertainendsthm1ghtheir
expressive behavior. We will posit that expr95$ion is a motivated
Immanbehaviorandthat ithasrecognizedpurposesinhtmlan
interaction. Wewill proposethat itmaynctbeasharmful foran
irriividualtoforegotheexpressionofanemctimasitisforthe
iniividualtofeelpro5o1refrmotherstoexpraswhentheydomt
really wish to do so.

Also central to this discussion is the concept of "display
rules" . These are the inplicit and explicit ru195 that govern
auctio'lally expro58ive behaviors. Display m195 have been defined by
different authors as both personally and culturally determined and
defined (lewis & Michalson, 1985; Malatesta & Haviland, 1982), gender
related (Malat95ta & Haviland, Hochsdlild, 1974; Shields, 1987) ,
governing deception, drama, sinulation, and masking of expression
(81mm & Bugental, 1982: lads & Michalson, 1985). The influence of
display rules is apparent in sale of the personal strivings mentioned
earlier. Our lmofledge of and confidence abort th95e rul95, their
applicability, clarity, or aubiguity, as well as the extent to which
or goals conflict with these ru195 are crucial to anbivalence over
expre58ion and, therefore, to the relationship between health and
expression. Murray (1938) characterized developuent into adulthood as
a process of learning the5e societal ru195. The individual "must learn

7
to ixhibit or redirect certain inpulses, talpcrarily or permanently, as
wellaslearntoforcehim (orher) selftoperformcertainotler
actioewhidlattletimearereplgnanttohis (orher) feelings (p.
20) ." Constant failure to conform with display ru195 could lead to
social censure and foster ambivalence. Constantly striving to nest
these rula5, with little or no regard for ore’s genuine emotional state
could lead to aubivalence as well.
Afinalconcernthatdeservestobeaddre5sedhereisthereasons
why expression of auction is particularly likely to be fraught with
anbivalence. Why should individuals wish to expre5$ their emotion and
yet sinultareoisly not wish to expr958? First, Batmeister and Tice
(1987 ) have contented that the subjective feeling state that comprises
an action ireludes a motivation to express. This d95ire to exprl955,
colpled with beliefs that expre55ion is recessary for psychological and
physical well-hairy would dictate expression. At the sane time, the
expro55icn of auction entails self-disclosure. The sane sorts of
conflicts and vulnerabilities which are inherent in the proc955 of
self-disclosure are also a part of auctional expr955ion. Emotional
expressimreqrirestleriskof intimacy. Concernscver intimacyare
conuoninoirculture (McAdam, 1985). Mdaushascomuentedthat
individuals are often aubivalent abort becouing intimate with
others—that often it is those individuals who crave intimate
relationships who also fear such involveuents. Emotional expre53ion
requiro5thatweallodotlerstoseethatsideofuswhichis1953than
civilized, asidewhichtesbeendenigratedbcthbyolrolltureardour
science (as later discussion will reveal). Individuals are faced with
adilm. mtleo'etemexpressimispresentedashealthyand

8

favorableardmtleotleritispcrtrayedasasignofwealcms.

Inordertoexamiietheroleofaubivale'eecveracpmsionin
psyd'ological arr! prysicel well-being, a oestiomaire was created wt
of the personal strivings collected previolsly by moors (1986) . This
questiomaire was administered along with a neasure of expressiveness
toagroipofsubjects. 'me5edatawereusedinafactoranalysisof
thesescales. Inaseco'ldstudy, thequestionnaireueasureof
aubivalence over expression was administered along with measur95 of
expressiveness to a second groips of subjects frou whou well-being
informationwasalsogathered. Thus, thetwostudi95tobereported
here will address tie contributions of aubivalence over expr955ion and
expressivere58 to psychological and physical well—being. The central
thesis of this research is that expre5sion is, in fact, of little
iuportance to well-being. One can be expressive or inexpressive of
auction and be healthy. Expressive behaviors thenseIV95 will rot
predict well-being. Rather, frou tte oirrent perspective, it is
aubivalence abort ore’s expre5$iveness style that fosters ill-being.

The relationship between aubivalence over expro55ion and
ifllibitim that is inplied by this model is inportant. A point that
willbenaderepeatedlythrougtnrtthisinvestigatimisthat
individuals may be inexpressive without being idiibited. Aubivalent
arriirexpressive irriividualsarelikelytobecmfusedbyauodel that
stresses the iuportance of expression, gee. More and King (1988)
fond that it is those goals abort which an irdividual is ambivalent
upon which he or she is unlikely to act. 'nms, anbivalent individuals
maybelikelytobe irexpressive, howevertteseindividualswillnotbe
distirguisl'ed frou realthy inexpressive individuals via ueasures of

9

expr95sivens5s aloe. Before discussing tie actual construction and
test of the Anbivalence over Expressiveness Qus5tionnaire it may be
helprl to reviov sons of the basic issles of auction r95earch,
afiesizing throughort points at which anbivalence my play an
inportant role in expr953ion manageuent.
mm. M 9f Beetle: am Wig

Geerally speaking, historically, awtional expre5sion has not
been viewed in a positive light. It has been portrayed as primitive
and uncivilized. mly relatively recently has its value been
acknofledged by r95earchers and tleorists. Hodechild (1985) has
categorized theorists in emotional expr953io'1 into two groips:
"organisuic" vs. "interactive." min, Fred and Jane5, (althoigh he
seats to fit this category less clearly) are placed in the former
category by Hcd'lsdiild while sociologists such as Goffman and
Wild herself mild fit into the latter category. Organisnic
thinkers stress tle basic fixity of emotions—the universality of
auctionsandexpressions. Theyviewmanageuentandassasmentas
extrinsic to the felt motion itself.

Daminisoftencitedastlefomderofrs5earchinemotiorel
expression (Moltgouery, 1985). According to Darwin (1872) emotional
expressiore are actions whidi occur whenever specific states of mind
areinduced. Thus, euctimardexpressimwerepmentedbyDarwinas
basically inswarable. He presented the inhibition of emotional
expression as essentially inperfect, stressing that even behaviors
mriertakentosuppressexpressimaretlelselvaexpressiveand
revealing. Itmaycoueasasmprisetoreadersoflhefiigjggf
mmtmmmmgmmmmesm

10

(1872) , Darwin did not present aotional expr955ion as directly
adaptive forhumanbeingsandthat, inthisvoluue, her950rtedto
leuarckian use-inheritance to justify tl'e existence of "universal" (see
Enouan and Friesen, 1971) facial expr95sions. In his principle of
"serviceable associated habits, " mrwin smested that atotional
expressioe arose oxt of the inheritance of acquired characteristics .
mile he viewed animal expre5$ions as adaptive and communicative, htmlan
expressions, like vestigial organs left by generations long past, were
seen by Darwin as actswhich occur "...tl'nrgh theymay not...be of the
least use (p. 28)." Darwin rejected the idea that there could exist
sous d95ign behind tnuuan expression (Mortgouery, 1985) . Darwin
explaired the existence of expressions as [hysiologicel imitations of
the actiore that would occur if we were still primitive people.

JanetandSartrealsocameoItofanorganismicviewofetotiorel
expressiolarrirepresentfurtlertleviewof motional exprwsionasan
uncivilized permemn. Jaret viewed euotimal reactions as a
"disorder" (Sartre, 1948). Sartre (1948) viewed awtional reactions
rot as pure disorder but ratter as a form of "l95ser adaptation".
Euotioial expressions, particularly expressions of negative euotions,
for Sartre, involved a "falling back" or primitive mechanisus.
W, Sartrerecognizedthattleseprimitivemedlanisusservea
specific purpose. He maintaii'ed that erotional displays are a form of
primaryprocesscoping—auearebymichtteselfcenbelonered in
ordertomeetreedsatalower level. Thus, evenwithinanorganismic
perspective, Sartreportrayedauctional expr955ionasanersatzmeans
to an end.

In contrast to the organismic view, the interactive view of

11

anotional expression stresses that expressions are a product of biology
and psychology. The interactionists aclcnolledge that management
inpacts upon am shapes the feeling itself. Frau an interactionist's
perspective, enotion mnagauent entails tte creation of experience.
Withintleinteractiveperspective, theinnerstateofeuctionisas
susceptible to social forces as the external expr955ion of enuotion.
mum ans Mariel: A uremic Mien

Intuitively, the relationship between auction and emotional
responsemayseeuasiupleone—ore feelsaneuctionandexpr955es it.
However, a plethora of research has proven this uncorplicated instance
ofeuctional responsetobe farfroutherule. Infact, manyfactors
are involved in the relatiorehip between an emotion state or feeling
(these terns will be used interdnangeably in tl'e present investigation
and discussion) and overt emotional displays. As our understanding of
the relationship between feeling state and expre55ion has increased, so
has our appreciation of its cuplexity. For Darwin, enotion and
display were inseparable. A felt auction was expressed. This basic
tenet, that an inner feeling leads to a particular emotional display,
hasbeenchallegedbyttecryandreseardnindicatingtrettheact of
eqar'955ion inpacts significantly upon tie quality of erotional
experience. The inpact of e¢pr95$ive behavior on enotional experience
isrelevanttoadisoissionofanbivalenceoverexpressiontothe
extentthat individualsareawareofthisinpactarrinrtilize itinthe
douains of euotion arnd inpro5$ion manageuent.

Jaues (1884) siggested that human enotional experience was the
peroeptionoftl'ebodilydnangeswhidnocoirinro5ponsetoa
stinulus. AccordingtoJames, fieremsnoeuotionecceptthatwhich

12

found expression, or more accurately, that which $1141 3 m _i_n
expression: "A purely disabodied human action is a nonentity (p.
194) ." According to James’ theory, eadi auction possesses a particular
setofskeletallmsclencvalentswhich infornsthe individual ofthe
quality of a particular action. 'Ihus, within this perspective,
actions thenselves are inferences based on perceived bodily changes
ardairowncognitiorsaboutwhatisapprcpriatetoapartiwlar
situation. Although James was placed in the category of "organismic
theorist" by Hochschild, we will introduce our discussion of the
interactionist view with an overview of his theory. James might be
consideredanorganianicthinkerirsofarasheassmedthateadi
actionpossessedafixedsetof sleletalwsclemvementswhidi
provoked the attribution of the specific action. Still, to his credit
James recognized the intimate interaction of action, physiology, and
actional expression, or, inhiswords 'Wmmdmourmental life is
knit up with our corporeal frame (p. 201)." According to James, not
onlydoestheintemalsubjectiveactionalstateinpactonovert
bdiavior but these behaviors actmally define actional experience.

Rick (1984) has suggested that, "the skeletal nuscle activity
associated with actional expression plays a direct role in regulating
emotional pm (p. 48)." 'Ihis connection was not missed by James,
himself. In addition to the nany titles allotted to him in psychology,
James might also be dubbed the father of action nenagaient, for, as an
extensim of his basic idea, James proscribed certain behaviors which
were intended to assist an individual in avoiding negative affect and
naximizing positive affect. Specifically, James maintained that if we
would refuse to express our actions they would eventually dissipate.

13
In ad'litim, James held that posing positive actions vmld cmquer
negative moods. In short, James surmised that the manifest expression
of an action could produce that action. Nearly a century later,
fascinatingreseardibylaird, IanzettaardotherssugaortedJanes’
intuitions about the inportance of expressive behaviors to actional
euperieme.

Iaird (1974) asked subjects to perform various facial contortions
which were designed to match aotional expressions of smiling and
frowning. Subjects' subjective emotional experience mirrored these
nanipulated facial expressions. mile smiling, subjects reported
thaeelvesto feel happierandwhile frowning, theyreportedmore
negative mood. Ianzetta, Cartwright-Smith and Kleck (1976) instructed
subjects to disguise or exaggerate their reactions to an anticipated
shock. Whenhidingtheirreactimtotheshock, subjectsshowed
reduced skin conductance (i.e., lowered arousal) and reported the shock
tobelesspainful relativetotheconditicn inwhichtheyexaggerated
their reacticms. More recently, McCanne and Anderson (1987) reported
that subjects who suppressed smiles while thinking of positive scenes
reportedlessenjoynentardnoredistressthanwhenpermittedto
acoapany their thonghts with congruent facial expression. ‘Ihus,
expressive behaviors can create, enhance, or mitigate action states .

mcently, the results of studies such as Iaird’s have been
criticized on nethodolcgical bases. mtsumoto (1987) perforned a
meta-analysis on these expression manipulation studies and concluded
that the actual effect size for expression manipulation was only small
tomderate. EVenso, mtamtomaintairsthat it is inportantto
inprove these studies methodologically, in order to examine the

14

cartribution of different cmpments of expression an emotional
experience. It is important that the expression manipulation studies
have bean criticized on methodological (vs. tteoretical) grounds.
Certainly, the results of these scidies are intuitively appealing and
standthetestofdailylife. Werecognizethatwedonanageaotion
by posing expressions. Indeed, the inportance of expression to
subjective actia'ial experience apparatly is recognized even by
children.

Cole (1985) reports that children are capable of posing happy

eupressionsatayomgagebutthattl'eabilitytoposesadorangry
expressions depends on developnental change. Anecdotally, Cole
daserved that children were often reluctant to pose negative
expressions. 'Ihesechildreiwould statethattheywerereluctantto
posesadfacesbecausetheydidmt'wmtofeel sad." 'Ihisevidence,
whileonlyanecdotal, doesaiggestfliatdiildrenlcowtlettheycan
poseregativeeaqaressionsbutmlyattlecostof feelingthe
camitant regative auctions.
'Ihisresearchhasinportantinplicatimswithregardto
anbivalence over actional expression. Inhibition of action can have
deleterious effects (:1 anoticnal experience in that it my diminish
ae’s experience of positive action. Hoooever, this research would
also suggest that, contrary to popular assunpticn, inhibition is not
always a bad thing. Indeed, inhibitim of auction and even dissembling
antimcanbeviaedinthiscontextasatool foremotionnenagaent
(see Sackheim, 1983) . Despite the usefulness of deception as a
legitimte means of action managaent, there is potential for
iniividuals to perceive thenselves as dishonest when dissembling. ‘Ihis

15

cmld lead to guilt and ardoivalenoe. It is regrettable that little
W has been done in this area.

films, the "causal arrow" betweel action and expression may be
double-sided. 'Ihe relatiorehip betweel state and display can be
caiceptualized as an interaction. 'Ihis interactive relationship does
notexistinavacuum, however, andisitselfaproductofanother
interactive relationship: "'me actioml eupressicn-aotional state
relationship is a functim of the interaction between biological and
envircnnental factors (Iewis & Michalsal, 1985, p. 166)." Certainly no
actional display occurs without social influences impinging, either on
the form of the display or its subsequent reinforcement.
Miss: A $221 lag

'nlerecognitionofaotionasasocialisalehasbralghtanev
legitimacy to this facet of human experience. Baumeister and Tice
(1937) suggest that it is difficult, if not inpossible to justify the
adaptive existence of aotional expression on an individual level.
However, withregardtothecmtinledeidstenceofasocialgmlp,
actional expression can be viewed as useful and adaptive. Euotional
expressim plays a key role in group cohesion, communication, and
loyalty (Bamneister s. Tice) . As early as 1924, Allport recognized the
social significance of actional expression and appended Darwin’s
exception of actional expression taking into account the purposive,
motivated nature of expression. Allport mintaired that although
expressive behaviors my not have been originally purposive or
canmicative they have m so. According to Allport, anotional
expression is not a product of [mysiological similarity but of social
cattext. ‘Ihe validity of Allport's assertions is demonstrated everyday

16
inhumninteractim. Weexpectcertainexpressions franothersandwe
knowthatwearembjecttosocial expressiveness expectationsas
well. Evidenceofthisawareressisprcvidedintieindividual
strivings that subjects have listed in our investigations. Further
evidence for Allport’s prophetic stataents is provided by Bauneister
ani Cooper (1981) , who found that the public expectation of action can
causethat action. Subjectswhoweretoldthat, basedonsane
criterion ofpastbehavior, theywereexpectedtobeabarrassedand
amcialsabwtsinginginpiblicwerencreabarrassedardamdwsthan
thosewhohadrctbeenprinedinthisway.
lbs gram 9.: We!)

'Be socialization of expression has been characterized as the
process of gradually separating action state fran expressive
behaviors, the asynchrmization of feeling and display (Cole, 1985;
I-Iilfle, 1985; Iewis 8: Midlalson, 1985; Malatesta, 1985; HOC'hSChild,
1983; Jones, 1935). Lewis & Michalscn trace the developrent of facial
expression fran sign (i.e., an innately meaningful expression, bearing
an isomorphic relatimship to a feeling state) to synbol (i.e., an
arbitrary expression whose neaning is iniividually and socially
determined). LevisandMidlalsmhavefandthatncsttm-yearolds
are capable of producing positive and regative expressions at
will—wholly apart fran their subjective feeling state. By age 3,
children are aware of display rules and are able to match appropriate
expressions with a variety of situations (Lewis & Michalscn, 1985) . In
his early studies with childrel, Jones fourd that the older subjects
showedlessanotimaldisplayardgreatergsrresponseto
enotiar-provoking stimli. Jaes suggested this as evidence for the

17
socialization of internalizing emotion-e rather than displaying them.
Jones terned this phermetm "leaned imjbition."

According to Field (1982), development brings an increased
capacity for expressive control. Children learn to produce a variety
of expressions apart frun feeling states and to inhibit expressions
cacanitant with certain feeling states. Field poses an interesting
question as to the ultimte inplicatim of the learning of actional
cartrol. We do not know if ncre self-management actually results from
thistrainingorif, insteadchildralacquireagreaterrangeof
expressive freedan. Perhaps the result of the socialization of
actional expression is flexibility. this wild allow children to
becane adults capable of using their expressive skills to their own
advantage in interactims with others. 'Ihese skills could also allow
irdividuals to avoid abarrassnent and vulnerability. In addition,
skills in action managaent could prove useful in the workplace.

Ehctimmanagenenthasbecaeanaclowledgedcamcdityinthe
narketplace. Hodlschild (1983) studied staaaniesses on the job and in
trainingprograms. Notsurprisingly, shefmndthatnuchofthe
trainirqthattheyreceived involvedthemanaganentof actional
dimlay. Stewardesses are traired in caplicated tactics for dealing
with their genuire actimal responses. vaicusly, this sort of
canercializatim of emoticnal display occlrs rather routirely in our
society. Unfortunately, because of the foregoing disolssion we can see
that artificial managaent of action can actually inpact upon
experience. lbdlsdlildsuggeststhatthiskindofpressureonthejob
can result in the alienatim of individuals fran their own actional

experience. 'Ihis kind of sale of affect display could be a source of

18
anbivalence over expressim.
Mime; miss am Ansel: An m Messy

tactical responsemyrctmlybevitaltoactional andsocial
experience but to [ilysical well-being as well. physiological arousal
is an obvious am acknowledged caponent of action. The relationship
between ewtional expression and physiological arousal has provided the
foundation formanyofmrassxmptiaeabaltthe linkbetween
expressiveness and illness. Ultimately, this relationship may prove
inportanttotlelimcbetweenanbivalerceovereupressicnard
ill-being.

Inhibition of acticnal expression has been associated with
psydiosanatic synprtcms and low levels of subjective well-being
(Permebaker, 1985). mny of our asslmpticns about the advantages of
emoticral expression are based on the inverse relationship between
enoticnal expression and autonanic reactivity which has been supported
by research (Buck, 1985) . ‘Ihis inverse relationship is perhaps ncst
familiar to us as the basis for lie detector tests—not telling the
truth causes an increase in galvanic skin response (gsr). Jones (1935)
foxmdthatwhenanenotimisexpressed, evenwhenitisemressed
intensely, tierenaybercchangeinarcusal. I-IokansonandShetler
(1961) and I-lokansm and Burgess (1962) found that the expression of
aggression was associated with lower arousal levels than the inhibition
of such expressim. Blocking aggressive inpulses toward a frustrator
resulted in increased physiological arousal, including increased heart
rate and blood pressure (Hokanson and Shetler, 1961) . 'Ihus, typically,
the inhibition of enotional expression has been associated with an
increase in autamic activity (Pemebaker, 1985) . Expression of

19
actimpermitsamrsal levelstoreturntobaseline. Inthissense,
anotialal expressial can be sea as a safety valve (Anderson, 1981).

'Ihe inverse relatimship between actional expression and
autamic reactivity has been translated into an individual difference
characteristic, asreseardlhasrevealedthatsanepeople, terued
externalizers, characteristically acpress their actions overtly while
others, internalizers, typically react autonanically to emotional
stinuli. Notarius arr! Ievenson (1979) identified natural externalizers
and internalizers by observing their reactiots to an action-provoking
film. Subjects were divided into the categories of externalizers and
internalizers by their consistent modes of reactions to the film.
Internalizers characteristically respalded autonanically to the film
while externalizers tended to change their facial expressicre and had
little autonomic reaction to the film. In a later threatening
situatim, externalizers exhibited ncre facial response while
internalizersstmedncreautolmicarwsal. NotariusandIevenson
found that externalizers were generally less physiologically reactive
trenintermlizersandthattlereherebaseliredifferercesbetveenthe
two groups. The distinction betweel externalizers and internalizers
has heal associated with other persalality differences as well.
Externalizershavebeenfwndtobencreeoctraverted, tohavehigher
self-esteenardtobencreacalrateserdersofacticnalncssagesthan
internalizers (Back, 1984).

Field (1932) filned infant facial expressions and found that the
externalizer-internalizer distinction could be amlied to infants as
well. Sale children naturally internalized their actional experience,
enmibitin; ncre autonanic reactivity while others manifested actional

20
expressions and showed little dlange in autonomic arousal. The
categories of externalizers and internalizers are not free fron social
evaluatiol. Field found that externalizers were viewed as acre
requensive arc had "better” interactions with parents than irhibitors.
Externalizers were also more popular in preschool.

Given the research on the inverse relationship between autonomic
reactivity and expression and the individml difference dinension of
externalizer-internalizer, it would seen reasonable to conclude that
expression of auction is healthy, lack of expression is unhealthy, and
ambivalence over expressiol is an unnecessary construct. Although the
inverse relatiotship between expression and autononic activity is
well-documented, this uncaplicated view has proven inadequate as our
mrierstanding of emotional expressim and inhibition has grown. 'Ihis
inverse model ignores individual expressiveness goals. Such a model
(emotenccnpassthecolplacitywhidlisapparent inhealthyadult
ewtim nanaganent. For instance, while there is evidence that facial
display attenuates arousal (Rick, 1984) remll that evidence already
reviewed (Ianzetta, Cartwright-Snith and Kleck, 1976) would suggest
that it augments arousal. Also, previously discussed research suggests
that inhibitiol can be a form of action managaent ained at inproved
fmctimirganiiecessaryforcocertratimandperformrcemvariols
tasks. lbsearch has daustr'ated that autolonic reactivity may be
related variously to nonexpressicn fig expression as a function of
situatimal factors, individual differences in characteristic styles of
expression (Sackheim, 1983; Roth & Cohen, 1986) and the quality of the
actiol itself ('Iavris, 1984; Murray, 1985) For exanple, with regard
to situational factors, Hcakanson and Shetler 1962) did find that

21
subjectsvtcaggressedagailstfrustratorswlcwereseenaspossessing
higherstacusthanthesubject, stuonedatrendtowardelevated levels
of arousal. Along similar lines, in her work on anger, Tavris (1984)
hasfoundthat itisstressful foranindividualwhofeelsrcdesireto
expressacticntobepressuredintodoingso. Lastly, researchers
(e.g., 'I'avris, 1984; unfray, 1985) have pointed out that negative
antics such as aggression, anger, and hostility present special
prdalas for individuals because the expression of these auctions may
hold inportant ramifications for the social group. Furtherncre, Bell
ard Byrne (1978) present evidence that repressors (i.e., internalizers)
actually report fewer health probl- than sensitizers (i.e.,
externalizers) .

Perhaps the key flaw in the inverse conceptualization of the
relationship between expressiol and arousal is that it fails to
distinguin betweal confortable inexpressiveness and inhibition. The
asamptimthatactimmstorshouldbeexpressedcarrieswith itthe
corollarythatthosewhodorctexpressanotimareinhibiting
expression. Clearly, such an assunption may be erroneous. Soretines a
blank face or a seemingly inamropriate smile is a genuine indication
of an imividual’s actional state. As we shall see this problal has
beenercomteredandaclcmledgedbyreseardersinthefieldofhealth
psychology and it is a problal for which the Anbivalence over
Expression Questionnaire is suggested as a solution.

An apprcpriate point of tralsition fron the siltplistic inverse
relatioship between expression and well-being to a ncre sophisticated,
goal based model of this relationship is provided by PeImebaker (1985)
inhissuggestimthat lackofexpressionpersemaynotbe

22
pathogenic. Rather, according to Pemebaker, lack of actional
expressiclcoupledwiththedesiretoexpressewticlisthe
(literally) fatal cabinaticl. Pcmebaker has found that individuals
whoirhibittteirdesiretocclfideinolhersaboutenoticlal life
eventsareatanincreasedriskfortledeveloptentof laterhealth
problu. Perlnebaker has coired the phrase "active inhibition" to
refertotl'eprocessofwillfullyprercrtingccself frcnadesired
action. Active inhibition is thought to be related to obsessive
flloghtsabouttleirilibitedactimanidlrclicautcmicarousalwhich
leads eventually to physical breakdovn.

Penrebaker’s nodel of the relaticship between actional
expressiveness and physiological stress accomts for the variety of
individual styles of regulating acticl which exist in the healthy
Innnan experience. It is possible, within this nodel, for an individual
whofeelsnostrongdesiretoexpressanacticltobeiuexpressive
withoutanyadverseconsequences. If, however, cewishestoexpress
an action but lacks the appropriate means of expression or believes
that expressicl will be censured, then ce is likely to actively
imiibit expression In this latter case, the individual is ncre likely
toexperierceplysicalsyaptasirdicativeofalclgtemircreasein
autclcnic activity as well as lowered life satisfaction (Permebaker,
1985) . ‘Ihrough Permdoaker’s ccceptualizaticl, ncexpressicl is
clearly distinguished from inhibition. mile nonexpression may refer
toeitheranaturalteniccyrcttobeveryexprossiveortoadaptive
acticl managaent tactics, inhibiticl is an active process of denying
ceselfwhatcetrulywishestodo forreascswhichmay involve
guilt, shame, or uncertainty.

23

Individuals have different styles of dealing with emotion, styles
whioharetoscneextentlearned (ques, 1935) andtoscteextent
prosaic at birth (Field, 1932). When these styles come into conflict
with social daarcs anbivalence my develop. As nick states, "'Ihe
socialsupportardpersclalcosistercyofacopingstyledetermines
its protective effect (p. 251) ." With this in mind, it seas most
inportant that we look at individual styles of expression and
irrlividual expressicl repertoires.

Pennebaker's foouscntheintentordesiretoexpress action
irfiicates that it is appropriate to look at individual’s goals with
regardtoenoticlaleuquressivetessasinportantfactorsindetemuning
the potentially pathogenic quality of certain characteristic styles of
emotion regulation. Based cm the research of Pennebaker, ‘I‘avris,
Elms (1986), and Elmore and King (1988), we may conclude that when
thesegoalsareinconflictwithsccialnornsorwithothergoals, an
irriividual may experience the detrimental effects of psychcscnatic
reacticl, whether they are expressive or inexpressive

'Ihescopeoftheinflucceofcmflictovereumressiclisrct
limited to the psychological. Exotical expression as been linked to
physical health as well as psychological well-being. Because the
goal—based model of the relaticship between anhivalence over
expressici and psychological well-being should also generalize to
physical realm, literature will now be reviewed which will culminate in
the inclusicl of the cccept of aubivalence over expressicl in the
relationship between expressiveness and physical health. A careful
examination of the applicable health psydiology research will show that
oftenalthough itappearscnthemrfacethatexpressiclisan

24
insulaticl agairst rhysical illnesses, in fact the relaticship
betwecl expressicl and illness is far ncre caplet.
W Series @ filth

Given the relaticship between cwticxal expression and autoncnic
arousal, it is not surprising that expressiveness styles have been
associated with a variety of 111nm. Inhibition of expression and
restrictive acticl managenent styles have been associated with a
umber of illnesses. As we broaden our lmowledge of the relaticship
between acticl and illness, Fridlund, Newman and Gibson (1984) suggest
that the term "psydloscnatic" may becae redundant since the role of
acticlal co-mediators will be inplicit in every disease state.
Eroticlmanagenentisoueshavebeclbroghtupinrelaticltocarcer
(e. g., Pelletier, 1985; Cox 3 McCay, 1982), corclary heart disease
(cm) (Friedman, Hall, and Harris, 1934), arthritis (Udelman s. Udelman,
1981) , and chrclic pain and dquression (Bertler et al, 1986) .

Pelletier (1985) imluded among the persclality characteristics of
carcarvictinsttetendcsyto"bottleup"arritleixsbilitytoeuquress
negative actics. 'Ihe "cancer-pros" persclality has been
characterized by a mencpressicl of negative ncod, suppression of
negative traits and coping by perseverance (Fridlund, Newman, & Gibson,
1984) . Time generalizatics are based cl numerous studies which have
fond that expressiveness styles can be used to predict diagnosis of
malignancy as well as prognoses.

Intheir1982 review, Ooorarrimytracethelclghistoryofthe
searchforaccmecticlbetweenexpressivenessardcancer. Galen
(cited in Cox and McCay) observed that melancholic woren were ncre

likelytodevelopbreastcarcerthansanguixewmen. Intheeighteenth

25

century, Guy (inCoxsrncCay) maintairedthathystericcaplaintswere
linked to carncer, and in the nineteerth century, Srcw (1883) declared
that intense emoticn drained the body’s vitality, lessening resistance
tocarncer. lbrerecently, inthe50’sand60’s, rosearcherssonghtto
connect particular psychological ccnflicts with careers in particular
body sites, alcng similar symbolic lines as those described in Freud’s
ccuversicn disorder sufferers (e.g., Tarlau & Snalheiser, 1951). More
recently, studies have begun to pirpoint conflict over actional
expressicn as a culprit in the etiology of cancer. Bacon et al. found
that breast cancer patients possessed an inability to deal with anger,
aggression arnd hostility (cited in Co: and McKay) . "Limited"
prospective studies (i.e., usirng wcnen who have "snspicious" lumps that
havemtyetbeendiagnosedasmalignant) havefoundthatthe
dnaracteristicthatdistinguishesbetweeumuenwithbenign lumpsand
those with maligrnancies is low neurotician scores arnd restricted
outlets for the expression of regative action (Cox & McCay, 1982)
Greer E: lbr'ris (1975) found that wcnen with breast carncer were likely
to possess abnormal patterrs of actional behavior, particularly
extras suppression of anger. Although few "true" prospective studies
havebeeldcne, thosethatexistsupporttheprenisethattheinability
toexpressnegativeacticuisassociatedwithbreastcancer (Cox&
McCay, 1982) . while feelings of hopelessness have also been associated
withcancer, Corandlbcayccncluiethatthestrcngestpredictorof
cancer is an anti-actional attitude, particularly an inability to
express negative acticn.

Derogatis, Abeloff, arnd Melisaratos (1979) evaluated long- and
short-termsurvivorsofbreastcancer. 'nneyfourrithatthosewcrenwho

26
wereevaluatedbycserversasshowinghighadjustnenttotheir illness
arnd as having uncre positive attitudes toward their physiciars died more
quickly than less well-adjusted patients. 0n self-report neasures,
log-term survivors scored signifimntly higher than short-term
survivors in hostility, anxiety, psydnoticism, depression, guilt, and
negative affect. long-term survivors also reported themselves as
feeling more vigor than short-term survivors. Inportantly, the two
groups did not differ in physical characteristics. Derogatis, et a1.
concluded that long-term survivors were better able to externalize
their negative nook. Morris, Greer, Pettingale & Watson (1981) found
that cancer patients tended to score significantly lower in anger
expressicnthanwcnenwhohadlumpswhidnwerebenign. Cancerpatients
tended to score high on "lie" or "social desirability" scales. These
differences were especially pronounced ancng younger subjects. The
portraitofthecarcerperscnalitytl'natenergesoutofthesestudies is
that of a well-adjusted, content, complaining individual—in short,
just exactly the opposite of our expectatics for one diagnosed with a
potentially terminal illness.

In an intriguirng analysis, Beutler, Bugle, Cro-Beutler, Daldrup
and Meredith (1986) conclude that the most ccsistent psychological
link between depression and dnronic pain is the "inability to modulate
or express interse unacceptable feelings (p. 752) ." Beutler et al.
maintainthat "thosewhoposmthis inabilitymaybeat increased
risk for both chrmic pain and depression (p. 752)." Beutler et a1.
identifythencstobservedcorrelatesofdnrcnicpainandcoexisting
depression are sonatoform arnxiety and conflict over expression of

anger. 'nneyproposethatprolcngedbloddngofinteseangermaybe

27
relatedtoircreaseddiseasesusceutibilityarrisuggestthat
therapeutic intervention be aimed at providirg outlets for unexpressed
auger. They state, "the nature of action, its target of expression,
ardtheewircmentinwhiouitisecpressedareinportantdeterminants
of its helpful or deleterious effect on pain (p. 757) ." To this list,
we might add more specifically, the importance of the congruence
between environmental display denands and expectatios and irndividual
styles of expression.

Udelman and Udelman (1981) aphasize the inportance of enotional
ecpressicn in the inprovaent of arthritis. They identify the problen
of "contaired hostility" as typical of arthritis sufferers. Udelman
and Udelman arinasize the inportance of case-by-case decisions for
therapeutic uses of action arnd expression. They stress that knowledge
of and respect for individual styles of expression may be key to
therapeutic success.

Enctior managaent has also been associated with coronary heart
disease (CHD). Friedman et a1. state that the Type A Behavior Pattern
(TAB?) is "intinately related to. . .copirg with and expression of
enotios (p. 151)." Inthis area, evidencehasbeenequivocal as to
the beneficial effects of expressiveness (Friedncn, Hall, & Harris,
1984) . The criteria for the identification of '13an include nonverbal
expressics of auger. Unlike cancer-prone individuals, Type A
personalities tend to be overtly expressive of hostility. The
relaticship between expression of hostility and TABP is not so
claret, however. Striieshavesumortedtheideathatmed
hostility was associated with CHD (Gilded, 1949; Cady, et a1. , 1961,

cited in Friednen, et al.). At the sane tine, Friedman, et a1.

28

acknowledge that there are individuals who are overtly expressive of
theiractics (innoluflinganger) wtlnoarerctTypeAbutarenaturally
charismatic. '"l‘rue" Type A’s terl to have a regative, hostile,
conpetitive style. Thus, individuals who express hostility as well as
thosewho"bottle itup"maybeatriskfoerD.

Thecaseofamdaunstratestheimportanceof individual styles
of emotion managaent in the develqmuent of pathology because
expression and inhibition can both be related to on). Obviously a
costnxztlikeambivalenceoverexpressiouisvitaltothestudyofthe
connection between physical well-being and actional expression.
Expression itself will rct predict which individuals develop health
problas. I-bwever, the relaticships between emotional experience and
expression as well as between situational demands and individual styles
of emotion mannagaent may do just that.

aeofthemcstimportanttasksforaquestiomairemeasureof
actioumanagenentistodiscrimirstebetweenthosewhosestylesare
protective and those whose styles are rct. A measure of expressiveness
canmotperformthiscriticaltaskbecause ittapsonlywhetherornota
person cnaracteristically expresses enotics, not how comfortable or
confidentheorshe feelswiththischaracteristic style orwhether
nonexpression is actually a product of inhibition. A conparisol of a
measure of expressiveness, the Affect Communication Test (ACT:
Friedman, Prince, Riggio, and DeMattea, 1980) annd the proposed AEQ will
clarify the differences between the ultimate goals of the two scales.

As previously stated, ce of the difficulties in identifying Type
A’s is that they may be either expressive or innhibitive of negative
action. Figures 1 and 2 are the Friedman et al. conceptualization of

29
thissituation. Notethatthegroupsmnoarehighincoronary
pronenesscamotbedistinguishedbytheirscoresontheAcr. In
Figure 1, individualswhcareeupressive oftheirhostilityand
coupetitiveressanithcsemnotendtoovercontroltheirregative
emotionsareatriskforCHD. 'nnesegruupsmaybeccnfusedbytheACI‘
becametheybothfeelinterseemotion. InFigure 2, iundividualswhc
arequietardrelaxedareconfusedwithpecplewhoarerepressedarfl
Wolled because of their similar motion-management styles.
‘Ihese individuals nay differ in their perceptios of their styles as
wellasintheirattitudestowardenotion. QnepurposecftheAmis
tcdistinguishbetweenrelaxed quiet people (i.e.,healthy inexpressive
irdividuals) and repressed, tennse people (unhealthy inhibitors).

30

Figure 1. CH!) risk and expressiveness (Fran Freidman, Hall, arnd
Harris, 1984).

 

 

High
Healthy Hostile
charisnatics conpetitiveness
(hi ACT) (10 ACT)

Ecpressiveness

W, W!
quiet persons overcontrolled
(lo ACI') inhibitors

low (possibly hi ACI')

Low High

Coronaryproneness

Figure 2. Diagnosis of 'I‘ypeAbased on expressiveness.

High Repressed tense Hostile
may be prone to corpetitiveness
other diseases

CoronaryProneness

Relaxed, Healthy
quiet persons charismatics

 

 

Diagnosed'rype B Diagnosed‘IypeA

31

Figure 3 presents an illustration of the rationale behind the
Anbivalence Over Expression Questionnaire. Individuals who fall in
either cell I or cell III would be at risk for physical disorder
because, regardless of whether or not they characteristically express
emotion, they feel highly anbivalent abort it. 'Ihis conceptualization
of expression and anbivalence abort expression makes no a priori
assnmptionabortthehealthfulnesscfenotioalecpressioninandof
itself. Rather eurinasis, is placed on irndividual satisfaction within

either characteristic erotion management strategy.

Figure 3. Cells II and IV are at risk for illness.

 

 

 

 

 

 

High
I II
Expressiveness
low
III IV
Low High

Anbivalence

32

mt 211m

Afascinnatingsubareawithintheareaofenntionreseardnisthat
of sex differences. In turn, emotional expression and auction
nenagenentareanangthemostcentralissuesinthestudyofgefler
differences. It sone ectent, proposed gender differences in auction
annd enoticnal expression are prototypical of all gender differences.
Cultural stereotypes dictate that wonen are often perceived as more "at
hone" with ennotion than nen. 'Ihis perception seens justified by
Shields' findings (1987) that among her subjects, wonen were more
likely to value enotional expressiveness, more readily reported
themselves as emotional, and reported experiencing a wider variety of
enotios conpared to men. In addition, wonen viewed male and female
expression as healthy while men viewed nale expression negatively arnd
fenale expression positively. The prediction that wonen will be more
anbivalentoveremotional expressionthanmenisbasedonthree
inportarntaspectsoftherelationshipbetaeenset, genderedeneticnal
expressiveness: 1) oultural deference to auction managenent over
enotion expression; 2) sex differences in ewtionality, the
socialization of emotion display rules and the institutionalization of
these ru1es;and, finally 3) the special case of woten annd anger.

‘Ihefirstoftheseaspectsisthemostobvious. Inanysociety,
enotion Inst be managed (Allport, 1924). Shields (1987) suggests that
inoursocietythevaluewhidnisplacedonlcuvingoe’s feelings is
nnotsomuchanencouragenentcfexpressionasanearstoesureprcper
managenent arnd regularization of emotion. Clearly, anbivalence is not
sonudnanissuecfeanessionasitisofmanageuent. 'Ihus, the
cannon assumption that wonen are more "in touch" with their emotions

33
andarebetteratcoweyingtheseemotiosthannenisnotnecessarily
pertinenttotreissueofanbivalenceinwonen. Sexandgender
differencesinexpressivenessarenotasrelevanthereassexand
gerierdifferen-ss inmanagenent. Dothesedifferencesexistand ifso
hovdotheypertaintothequestioncfandaivalenceoverenotional
expression?

Develcpnental psychology has produced provocative research which
speaks to this question of gender differences in enotion managenent
skills. Various researchers have videotaped and annalyzed infant
expressive behaviors and enotional interactions between infants arnd
nether-s. 'nnisreseardnhasshovnthattherearedifferencesin
enotionality between boys and girls and, more inportantly, that boys
and girls are given different forms of training in display rules.
Malatesta (1985) has found that in the opening weeks of life, male
infants demonstrated more labile affect than fenale infants. Boys
sinned more grimacing, more changes in expression, and were more easily
startledthangirls. Fennale infantstendedtobemore attentiveto
social stimuli (e. g., faces annd faces accoupanied by voices) than
male infants. Fenale infantsspentnoretinemaJdnganisustainingeye
contact than male infants. In sum, male babies were more emotional and
fetale babies more sensitive to social ones. An interesting side note
totheseresults indicatesthatthelatterpointmstbetakenwitha
grain of salt. Malatesta and Haviland (1982) point out that physical
differences in the facial structure may lead adults to conclude that
fenale infants are more attentive than male infants. Females’ eyebrows
are set higher on the forehead thannnnales', givinng females a more
”interested" look than mles. 'Ihe inplicntio's of this simple physical

34
difference could be profound. Feuunle infants may be perceived as more
"sociable" and may be treated differently fronn male infants as a
result. Perhapsastrogerarhasisonthesocialaspectscf
ecperienceasdnildrencouldleadadultwanentofeelmore anbivalent
abontactional equressionthannen. Wonenmaybetrainedtobemore
avarecfthesocialcosequencesoftheecpressioncfactions,
partioularly negative auctions.

In addition to differences in the earliest part of infancy,
differences have been deserved in the socialization of expression rules
for boys and girls. With regard to early socialization, Malatesta and
Haviland (1982) foundthat forbothsexes, nctherstendtorespod ina
fashion that is contingent with their baby's expressions. However,
ncthers typically respod differently to male and fennale infannts. 'Ihe
ncthersstxiiedbyMalatestaandHavilandtededtcengage inmore
expression matching behaviors in dealing with sos than in dealing with
daughters. mummrecftenmtdnedtheirsosezpressiosand
followed their daughters' expressios with dissimilar responses.
Mothersalsoinnhibitedrnmtdningrewoseswithboysmtnctwith
girls. Generally, ncthers were more variable with daughters than with
sos. Itwouldseennthen, thatdaughtersareexposedtoawiderrange
of expressions (whichmaybeconnectedtothebrcaderrangeof actions
that wonen report experiencing) while sos are exposed to reinforcaent
ofalimitedsetofexpressions. Contingentrespoding increasedwith
age forboysbutdecreasedwithage forgirls. Onemightsurmisethat
girlsaregivenmoreenotional indepedencethanboys.

Ievis and Michelson (1985) found that ncthers increased

resposiveness to displays of positive action and decreased

35
resposiveess to diwlays of negative action for both boys and
girls. But, inportantly, nncthers became less resposive to the crying
of male babies relative to fanale babies. ‘Ihus, male babies are given
differential treatnent with regard to negative action. Perhaps, male
infannts learn that their displays of negative action rust be managed.
'Ihis would indicate that, for better or worse, boys are given explicit
action managaent training particularly for negative actions while
girls are nnot.

Shennnum and Btgental (1982) report that amog grade sdncolers
sanpled, girls were ncre adept than boys at masking their actions
(i.e., displaying the action opposite tte oe they felt) while boys
were superior at inhibiting actional expression. In a study of fifth
graders, Cole (1985) found that boys were better at deception than
girls. Boys were particularly better able to neutralize their negative
actions (Cole, 1985) . In the sane sanple, Cole found that the oldest
girlswerebestatappearingpleasedwitharewardmnentheywere
actually disappointed. ‘Ihuus, fron infancy to grade school girls are
ncreadeptatmanipulatingenotional expressionwhileboysarebetter
able to sinply inhibit their expressions, especially for negative
feelings.

Differences in socialization may nct inmediately appear
significant to adult lives. However, if he cosider that males and
faeles can experience different socialization processes for display
rules, then we have only to recall that in our society, typically, what
ismale is often heldup as the standard (see Chodorov, 1978), wewill
havesetthestageforagreatdealof fanaleaunbivalenceover
actional expression. If we conclude for tie sake of argument that the

36

results of research on socialization differeces are a reflection of a
realdifference, manassnnethatnenandwonenarencttatghtthe
saneactionmanegaentsldllsasdnildren. Menaregivenncre
explicit rules while wonen are given ncre expressive freedou. This
feminine freedonn would seen an advanntage except that a "freeness" with
suction is net valued by society—actional control is. Rules that as
infannts may be considered to be "male" rules becone "the" rules in
adulthood. 'Ib make matters worse, ttese adult (formerly finale")
display rules are implicit in adult culture and are only made explicit
oncetheyarebroken. 'Iheedresultisthatwonenarenottrainedin
the display rules of the culture whidn they will later infiltrate as
students, adults, ncthers, and workers. Particularly, the research
Ieredenostratesthatnenaregivenspecifictrainingastothe
ammpriate manner in which to deal with negative actions. 'Ihe
actionofangerhasbeendescribedbymanyasapotential probletn for
wonenn. Angermaybeviewedasoeoftlencstanbiguos actions for
wonen, in that wonen are not specifically trained for dealing with it
and it is portrayed socially as an "nm-faninine" emotion.

Although actions in general have been idenntified alnncst entirely
as a faninine dotein (Shields, 1987; Keller, 1985; Hochschild, 1975),
the action of anger, which Shields regards as "an action of
entitlaennt," is net traditionally associated with wonen or "feminine"
roles and as sudn presents a special problan for wonen as individuals.
Onanabstractlevel, wonenhavebeenvieedastheconmmal,
connective force in a culture (Keller, 1985) , while anger can be viewed
astheactionofdisconnection. Iookingattheexpressionofangeras
anactionrenaganentissue, momseethattteroleofmurturer

37
allotted to wonen is a specifically actional role which carries with
it particular feeling and display rules. Fnurtheruucre, because the
faele/faninine role is an explicitly actional role, wonen are
routinely placed in situatios in which display rules which govern
ecpcressios conflict.

Hcchsdnild (1975) states that, for vmen, feeling rules conflict
withinandacrosscontects—mare, ingeneral, expectedtobe
caring and nurturant while specific contexts (e. g., in the workplace)
daend "affective neutrality" or even outright hostility.
Bernnardez-Bonesatti (1978) has suggested that the problenn of anger
could leadto furtherproblans forwaeninternlsof achievenenntand
powernctivation. Ste linnksthe inabilitytoexpressangerto
pathological suhnissiveness, self-defeating behavior, self deprecation,
anddepression. "'Ihecapacitytoexpressangeriscrucialtothe
develqnnent of goal-directed behavior" (1978, p. 217).
Bernnardez-Boesattiseesvnuenastedingtowterdom" theiranger
inordertopreserveinterpersonalbods. 'Ihebalancewhichmustbe
mainntained (if it can ever be adnieved) between general, pervasive,
cultural expression expectations and the feeling requiraennts of daily
life is a delicate oe. It is nnot difficult to imagine how negotiating
this balance and the conflicts it inplies could ultimately breed
ambivalence about actional expression. 'Ihis annbivalence is predicted
to inpinge on fenales’ realth and psychological well-being in the ways
which would be expected, given the previous discussion.

This disossion of set differeces in actional expressiveness and
anbivalence over expression dannstrates the different predictions that
arisefronttegcal-basedncdelofexpressionasconparedtothencre

38

traditional View of action. mile the older view of action would
predictthatthosethoarencreecpressivewillbencrerealthy, in
this discssion of sec differeces, the conclusion is reached that
mnilewaeunaretypicallymoreexpressivethannenntheywill alsobe
more anbivalent over expression. Further, because they are none
ambivalent over expression waen are predicted to suffer the negative
health consequences which have been associated with ill-being and which
are lnere attributed to underlying ancivalence. Thus, it is precisely
thosewhoarencsteuanessivewnoarepredictedtosuffergreater
distress. Men, who are inexpressive (but not necessarily inhibited),
arepredicted, becauseofthesocialapprcvaloftheirstyleaswellas
the training received in that inecpressive style, are predicted to be
better off physically and psychologically.
mg

Individually and culturally, actional expression presents a
dil“. Ebcpressionisviewedasreesssaryandhealthyandatthesane
tine as self-indulgent and unncivilized. Anbivalence over anotional
ecpressionissggestedtereasaninportantfactorinthe
relatioships between actional expression and psychological and
physical well-being. While traditionally expression has been treated
astherealthyandnaturaledtoenotionandlackofexpressionhas
beenconcludedtobeunhealthy, thepresentperspectivemaintainsthat
neitrer expressiveness nnor inexpressiveness is, in and of itself,
healthy or pathogenic. lather, anbivalence over o'e's style of
emotional expression is presented as the underlying factor that is
associated with pathological outconnes. In order to examine the

inportance of anbivalence over expression and expressiveness to

39
well-beingtmqestionnaireneasureswerecostructed—oetoneasure
eadnofthesecostructs. 'Ihedevelcpnenntandapplicationofthese
questionnaireswillbedescribedinthetwostndiesreportedhere.

Inthefirststdythescaleswerecreatedtoneasureanbivalence
overactional expressionandactional expressiveness. 'Ihesescales
wereadministeredtoalargegroupofstudentsatMidniganState
University. 'Ihis data will beused in a factor analysis of the scale
andaconparisonofsexdifferencesinactualecpressivenessand
anbivalenceoverexpression.

40
Studylz FactorAnalysis
magnum

A list of 600 personal strivings that had been collected from
subjectsinanearlierstudyflumnos, 1986),wereeucaminedandthose
strivings that dealt with actionwereconpiledintoalist. 'Ihis
listwasusedasabasisforthegenerationoflmitensabout
euctionalbehavicr. Itminthislargelistweredividedintothree
categories according to their major focus: strivings, behaviors, and
attitdes. The behavior itans were placed in a separate
qestionnaire, the Functional Expressiveness Questionnaire (Em), which
featuredaratingscalefronlto7with1indicatingthesubjectdid
nctagreewiththeitennatallanda7indicatingthatthesubject
stroglyagreedwithtteiten. AnecnpleofanitennfrontheEEnis
"Pecple can tell fronmy facial expressions hot I am feeling." High
scoresontheEmindicateatedecytoenpressaction(SeeAppedix
A).

‘nerenainiugstrivingitaewererewordedinordertoaddress
their potentially ambivalent dnaracter. 'Ihis rewording was performed
with an effort to preserve the originnal wording and intent. For
instance, the striving, "I typically strive to express my actions
honestly"becane"Iwannttoeq>ressmyactioshonestlyoutIam
afraid that it may causemeabarrassnentorhurt." Instructions for
tleAEnaQnasizedthattlesubjectanswereadnitenwithaviewto
itsoverallneaning. This, ifastataenntcosistedoftwothoughts,
subjectswereecouragedtogivetteitenahighratingonlyifboth
thoughts applied to then. ‘lhe rating scale for the questionnaire was
ascalefronnltos,with1indicatingthattherespodentneverfeels
likethestatanentsuggestsandasirdicatingthattherespodent

41
frequently feels that way (See Appendix B).

It may be ncted that re itas on the Am are negatively worded.
responsebiashasbeenaneduringcocerninpsyonological research.
W, becauseoftheconplennatnreoftheitasreversewording
was deaed infeasible. 'Ihe potential contamination of scores by
acquiesceceisnctlikelytobegreatrcweverassudnbiasisncst
likely to influece dichotonos judgments nct rating scales
(Schriesteim 8: Hill, 1981) . In addition, Schriesheim and Hill (1981)
have reported that negative itas may inpair response accuracy. They
found that negatively worded itae actually inpaired reliability. In
tlecaseofaquestionnairewithitaethatarealreadycapoundin
nature, the disadvantages of negatively worded itans were believed to
outweigh the advantages.

Predictios
1- Mwmmm It ispredicted that
neither the AR) ncr tie Em will be positively correlated with the
Marlon-Crowns Social Desirability Scale. It is further predicted that
the Alan will be significantly positively correlated with the Raulin
Intese Anbivalece Scale (Raulin, 1984) . Such a correlation can be
taken as evidence of the validity of the AR).
2. g; M. For reasos specified previously, it is
predicted that overall, wonen will be significantly ncre expressive
and ncre ambivalent over expression than men. More specific
predictioscanbenadewithregardtotteoriginofthisoverall
differece. With regard to expression, Hodschild, Shields, and
Bernardez-Boesatti suggest that female predoninance in the actional
realm does not generalize to negative actions. ‘Innus the overall

42
differece in expressiveness should be due to differeces on positive
and inntimacy actios rather than negative actions.

With regard to anbivalence over expression, wonenn should be ncre
annioivalent over expression of negative action, since develcpnental
literature cited previously would suggest that nen are taLght explicit
rules for dealing with negative action while wonen are nct. 'Ihus,
theoverall differecebeteennenandvmenshouldbeaccounted for
by differences in the anbivalece over the expression of negative
auction and net in referece to positive actios.

3- mmmmmmm. BecauseEmonsandKim
(1988) have found that individuals ted nct to act on the goals about
midntheyfeel anbivalent, itispredictedthatscoresonthe
anbivalece scale will be negatively correlated with scores on the
expressiveness scale.

nethod
m. Two hundred and ninety-nine Michigan State University
stdents (117 males and 182 fanales) coupleted a packet of
questionnaires for extra credit in varios undergraduate psychology
courses. Malesnbjectsrangedinagefronlstozswithaneanageof
19.8years. FaelesubjectsrangedinagefronlStoBZwithanean
ageof 19.1. Allbuttwoofthefanalerespodentswereundertheage
of 26. .

Subjectswereinformdinclassthattheirpacketswouldbe
checkedforconpletenessandcarefulness (inordertodeterminethe
annountof credit received) afterwhichtheirnaneswouldberacved

fronttepacket, esuringanonymity. Capletion ofthepacketwas
conpletely volunntary.

43
m
'Itesmreyswhidnverecapletedbythissanple of subjects
includedthenevlydesignedAmandEEn. 'IheRaulinIntense
Anbivalece scale (Raulin, 1984) and the Marlow-Crowns Social
Desirability Questionnaire (Marlow-Crone & , 1964) were included for
construct validity purposes.
Results and Discssion
[htaanalyses forStdylinvolved4steps: 1) theitanand

factor analysis and validation of the AEQ and Em; 2) a comparison of
Am and EEK) scores for male and fanale subjects; 3) the computation of
Psarson correlation coefficients for the Am, the RED, their
subscales, the Marlow-Crone, and the Raulin I-A scale; and 4) the
coupnntation of the Pearson correlation coefficients for the revised
Am and revised mm.
W ear—Id Easter files 21: he m

ItasontheAmweredeleted forcoceptual reasons, becausethey
seaedtopertainto inhibitionof actional expression ormaskingof
expression without mention of a conflicting desire to express or
because they referred to the quality of actional experiece itself.
Several itensweredeletedbecatsetheyseenedtopertainncreto
self-assertion than actional expression. One itan was deleted
because itreferredtothinkingaboutlifeeventsratherthan
ecpressing action. Additional items were deleted on the basis of low
item-total correlatios. ‘Ite deletion of itans left a total of 28
itas in the revised Am with a mean inter-ital correlation of .23, a
minimum inter-ital correlation of .02 and a maximum interitan
correlation of .64. Ital-total correlatios for the A150 are shown in

44
Table 1. The alpha reliability coefficient of the Am was .89. The
neanforscoresonthescalewasz.9 (onaS—pointscale) andthe

scale standard deviation was .58.

45
Table 1

Ital-total Oorrelatios for Revised AEQ

Lg Dig-.1222; correlation
1. Itishardtofindthatrightmrdstoindicate .58

to others what I am really feeling.

2. Iworrythat iercpressnegativeactiossuchas .56
fear and anger, otherpeople will nct approve of ne.

3.IwanttoexpressmyactioshonestlybutIam .56
afraidthatitmaycausemeabarrassnentorhurt.

4. IoftencannnctbringmyselftoecpresswhatIam .56
really feeling.

5. I'd like to talk about my problas with others, .55
but at tines I just can’t.

6. Iwanttotellsoneoewhenllovethen, .55
but it is difficult to find the right words.

7. Iwmldliketoenpressmydisappointnentwren .55
thingsdon'tgoaswellasplamed,butIdon't
wanttoappearvulnerable.

8. OftenI’dliketostcwothershowaeel,but .52
saethingse-stobeholdingmeback.

9. Itrytchidemynegative feelingsaroundothers, .52
eventhoughIannnctbeingfairtotrcseclcsetonne.

10. OftenIfindthatIamnctabletotellothershow .51
mudntheyreallyneantone.

11. Itrytokeepmydeepestfearsandfeelings hidden, .48
butattimesI’dliketocpenuptoothers.

12. Iwouldliketobencrespontaneosinmyactional .46
reactiosbnutljustcan'tseantodoit.

13. IcanrecallatinewhenlwishthatIhadtold .45
soneoehovmuchIreallycaredaboltthen.

14. Ifeel guilty afterlhaveexpressedangertosoneone. .44

15. Iwould lihetoexpressmyaffectionncre .42

pnysicallybutIamafraidotherswillget
ttewroginpression.

16. Itrytosuppressmyanger,b\utlwou1dlikeother .42
pecpletoknnowtcwlfeel.

46
Table 1, continued

11% Item-totalcorrelation

17. ItrytoapologizethenIhavedonesonething .41
wrogbuthorrythathillbeperceivedas
incapetent.

18. AfterIexpressangeratsoneoe,itbothersne .41
foralogtinne.

19. I trytoshcwpecple I love than, although at times .40
Iamafraidthatitntaymakemeappearweakortoo
sesitive.

20. Istrivetokeepasnileonmyfaceinorderto .40
convinceothersIamharpierthanIreallyam.

21. Mnensoneoebothersne, I trytoappear indifferent .38
eventhoughI’dliketotellthentcwaeel.

22. ItrytoavoidsnlkingevenwhenI feel like it. .36

23. WenIamreallyproudofsonethingIacconplish .36

Iwanttotellsoneoe,butIfearIwillbethought
ofascoceited.

24. Itrytorefrainfrongettingangryatmyparents .36
eventhoghIwanttoattines.

25. Itryncttoworryotherseventhough .36
sonetimestheyshouldknowthetrnuth.
26. Itrytocontrolmyjealousycocemingmyboyfried .34

girlfriedeventhoghIwanttoletthenlmowI’m
hurting.

27. Ithinnkaboutactingwtenlamangryhutltrynctto. .33

28. Imakeanefforttocontrolmytaperat all times .25
eventhoghI’dliketoactonttesefeelingsattimes.

47

An exploratory factor analysis was conducted on the interitenn
correlationnatrix fortheAm fornenandwonenseparatelyto see if
therewereinportantdifferemsbetweenthesexesinthefactor
structureoftlescale. Forbothmaleandfaale subjects, 2 large
factors aerged—one for positive enction and 1 for negative action.
Since there were only slight differeces in the factor solutios, data
were contuined across sexes in furtl'er annalyses.

An exploratory factor analytic procedure was used to investigate
the factor structure of the AH). Principle capoennts extracted two
factors with eigenvalues greaterthanone. AVARIMAXrotationwas
used because it was judged ncst likely to yield an interpretable
solution (Hunter, 1985) . Sixteen itans loaded highest on the first
factor (eigenvalue = 6.70) accounting for 17% of the variance. Itas
andloadingscanbeseeninTable 2. Itans loadingonthis factor
cocerned love and affection as well as a fear of expression of
actios that might lead to vulnerability. Ebenples of the types of
itens that loaded on the first factor include, "Often I find that I am
nctabletotellottershowmudntteyreallyneantone," "Itryto
slmpeoplel lovethennalthoghattineslamafraidthat itmaymake
neappearweak,"and"IwanttotellsoneonewhenI lovethan, buntit
is difficult to find the right words." This factor can be terned
aunbivalece over the expression of positive action and
self-disclosure.

Twelve itens loaded on the secod factor (eigenvalue== 1.35) .

This factor accounted for 12% of the variance. Itans loading on on
this factor also pertained to anbivalece over expressing primarily

negative enotios (e. g., "Afterquuressangeratsoneoneitbothers

48

neforalog tine") However, oeiten dealing with pride also
loadedontlesecodfactor,"¥henlamreallyprodofsouethingl
acconplishIwanttotellsoneoebntIfearIwillbethoghtofas
coceited." Anctherenanpleofanitenfronthatclusterisiten4,
"I try to control my jealousy cocerning my boyfried or girlfried
eventinghIwanttoletthanlocwI'mmurting." 'Iheseitansand
theirloadingsareshowninTablez. Becauseitasloadingonthe
secodfactorpertainednctonlytoangerbuttofeelingsofprideand
jealousy, the term of "entitlement" (as per Shields, 1987) was
selectedtoecolpassallofthese. Thesecodfactorwasconcltded

to tap anbivalence over expression of actios of entitlaent.

49
Table 2
Ioadings forTwoFactorSolution forRevisedAm

Factor 1 Factor 2
M Mgf-Disclosure Entitlaent
32. OftenIfindthatIamnctableto .68 .04
tell others how mudn ttey really
mean to re.
33. Iwanttotellsoneoewhenlove .66 .12
then, but it is difficult to find
the right words.
16. Often I’d like to show others how .61 .13
I feel, but sonnething seems to be
holding me back.
25. I’d like to talk about my problas .59 .20
with otters, but at times I
just can't.
23. Itrytokeepmydeepestfearsand .57 .12
feelings hidden, but at times I'd
like to open up to others.
2. I wannt to egress my actions .55 .27
honestlybnntIamafraidthat it
may cause re abarrassnent or hurt.
36. I often cannnct bring myself to .54 .30
egress what I am really feeling.
42. It is hard to find that right words .54 .34
toindicateotherswhatIamreally
feeling.
13. I would like to egress my affection .51 .09

ncrephysicallybnutIamafraid
otterswillgetthewrongingression.

34. I would like to egressmy .46 .35
disappointment when things don’t
goaswellasplannned, butIdon't
want to agear vulnerable.

39. Itrytohidemynegative feelings .45 .33
aroundothers, eventhoghIamnct
beingfairtottcseclosetome.

40.

28.

22.

38.

14.

48.

44.

41.

43.

46.

26.

50

Table 2, continued

Factor 1

minute/W

Iwouldliloetobencre
inmyenotionalreactionsbutI
justcan'tseantodoit.

ItrytoshowpecpleIlovethem,
althoghattimesIamafraidthat
itmaymakenearpearweakortoo
sesitive.

Istrivetokeepasnileonmy
faceinordertoconvinceothersI
amhamierthanIreallyam.

Icanrecallatiunewhenlwish
thatIhadtoldsoneoehowmuchI
reallycaredaboutthan.

Itryncttoworryothersevenn
thoughsonetimestheyslmldlucw
thetruth.

AfterIegressargeratsoneoe,
itbothersneforalogtinle.

Ifeelguiltyaftethave
egressedangertosoneoe.

Itrytosuppressmyanger,b\utI
muldlikeotherpetgletokncwhow
Ifeel.

Iworry that iergress negative
actiossuonasfearandannger,
otherpeople will nctanprove of ne.

Ithinkaboutactingwhenl
amangrybnutItrynctto.

Muensoneo'ebothersme, Itry
toappearindiffereuteventhough
I'dliketotellthanhowaeel.

Imakeanefforttocontrolmy
taperatalltineseventhough
I'dliketoactonthesefeelings
attiunes.

.44

.41

.39

.38

.38

.10

.16

.15

.36

.09

.20

.03

Factor 2
Midst—en;

.15

.21

.31

.15

.63

.59

.51

.49

.49

.42

.38

4.

9.

51

Table 2, continued

Factorl
Ira mew—L—Sel-Disclosum
8. ItrytoavoidsulkingevenwhenI .17
feellikeit.
27. Itrytorefrainfrongetting .19
angryatmyparentseventhoughI
wanttoat .
30. ItrytoapolcgizethenIhavedone .29

Itrytocontrolmyjealousy .18
cocerning my boyfried or girlfried
eventhoughIwanttoletthanknov

I’m hurting.

MnenIamreallyprodofsonething .24
IacconplishIwanttotellsoneone,

butIfearIwillbethoughtofas
coceited.

Factor 2

Erntitlaent

.38

.37

.35

.34

.32

52

A subsequent confirmatory factor analysis was perfornred on for
the results of the egloratory factor annalysis. A confirmatory factor
analysis was undertaken as a follow-up to the egloratory analysis for
several reasons. First, the confirmatory factor analysis allows. for a
deckoftheinnternalconsistencyoftheclusters. Iftheanalysis
aploys comrmalities in the diagonal, the itas’ loadings on the
clustersrepresentanestimate oftheircorrelationwiththetrue
score for the trait underlying the cluster. If the itans all load
highestontheclusterinwhichtheyhavebeenplacedbythe
egloratory analysis, the clusters can be coclrded to be internally
ccsistent (Hunter, 1985) . Most ingortanntly, a confirmatory factor
analysisallows forthetestingofaneasurenentncdelbasedon
classic reliability theory (Hunter).

Itae loadingoneachfactorwereenteredintoaclusteranalysis
using PACIQnGE (Hunter, 1988) . Both clusters had high reliabilities
(o<= .87 for the anbivalece over positive action cluster and d =
.77 for tie anbivalece over egressios of entitlement cluster). All
oftheitasdidloadhigtestontieclusterintowhichthe
egloratory annalysis placed then. Generally, however, itens were
highly correlated with both of the clusters. In addition, the two
clusters were thanselves highly correlated (_r: = .71). This high
correlation as well as the high correlatios of itas with both
clusterssuggestedttettleAmmightbebestdnaracterizedastaming
a single general costruct, anbivalece over actional egression.

A secod order coufirnatory factor analysis was perforuned on the
two cltsters yielding the netrix shown in Table 3. A single geeral

factor aerged with an eigenvalue of 1.39 on which both clusters

53

loaded highly.

Table 3
Am Cluster Correlation Matrix with the General Factor
Factor 1 Factor 2 Factor 3

Positive . 71 . 70 . 84
Negative .71 .84
General 1 . 00

Apathanalysiswasusedtotesttheneasuranentngdelthatisshom
inFigure4. 'Ihepnrposeofthistestwastodenonstratethatthe
correlations between the clusters could be seen as spurious—a product
ofthegenneralfactorunderlyingthenboth. ‘nnesumofthesquared
deviations of the predicted correlations fron the observed
oorrelations(ameasureof overall fit)waslessthan.02. A
dni-squaretestoftheresidualswasnotneoassarytodemomtrate
these deviations to be insignificant.

Altlnghtheseresnfltssupporttheoontentionthatthemis
unidinnennsional,wewill retainthesubclusters whidnemergedinthese
analysesinordertoinwestigatetheprqoosedsexdifferenoesforthe
different specific types of auctions, after which the subscales will
bedisregarded.

54

Figure 4. Measuranent model with path coefficients.

Anbivalenoe over mpressing Love

/

.84

General Anbivalenoe

\.

84

Anbivalenoe over Expressing
Enotions of Entitlanenrt

55

____Miabili g m m QI IE m

Initially, items on the Emotional Expressiveness Questionnaire
(Elfin) were deleted for conceptual reasons. Upon close enaminnation it
became clear that several of tie itans tapped positive
self-assertiveness rather than auctional expressiveness per se, and
theseitansweredeleted. Otteritansweredeletedbecausethey
referred to emotional experience itself rather than to emotional
expression. Additional items were deleted on the basis of low
item-total correlations and inconsistent inter-item correlations . The
deletionof items leftatctalof 16 itasintheEEQwithanrean
inrter-itan correlation of . 18, a mininunn innterhitemn correlation of . 06
and a maximmn inter-iten correlation of . 59 . Item—total correlations
fortrefimareshcwnin‘rablell. 'Ihenneanforscoresonthescale

was 4.6 (on a 7-poinnt scale) and the scale standard deviation was

.76. 'lte alpha reliability coefficient of the E151: was .78.

 

56

Table 4
Iten—total Correlations for Revised Em
Item Itsnicgtel 92mm
8. I often tell people that I love then. .54
45. Istmthatllijesoneonebynnnggirg .49
or touching that person.
36. I often touch friends during conversations. .49
23. Watching television or reading a book can make ne .49
laugh out loud.
10. I laugh a lot. .42
1. MenIamangrypeoplearcmrineusuallylom. .40
12. People can tell from my facial expressions how .37
I am feeling.
30. Whenever people do nice things for me, I feel .37

"pnt on the spot" annd have trouble expressing
my gratitude. (-)

18. mnen I really like saneone they know it. .37

25. I often laugh sohardthatmyeyeswater .34
or my sides ache.

22. When I am alone, I em make myself laugh by .31
by remembering sanething fron the past.

34. My laugh is soft annd subdued. (-) .30

29. If a friend surprised me with a gift, I numldn't .29
know how to react.

14. I apologize when I have done sanething wrong. .24

26. If some makes ne angry in a pablic place, I will .19
"cause a scee."

7. I always express disappointment when things .18

don’t go as I’d like then to.

57

Initially, factor analyses for the EEQ were conducted for males
and females separately, waver because tie solutions were virtually
idennticalthegronpwasusedasawholeinall furtheranalysos.
Becausenoapriori clusterswerehypcthesized fortheEEQ, the
inter-item correlations were entered into an exploratory factor
analysis, using comunalities to correct for attenuation due to error
of measurennent. Principle conponents extracted 3 factors with
eigenvalues greater than 1 whidn were rotated orthogonally via VARIMAX
rotation. The eigenvalue for the first factor was 3.41, for the
second factor it was 1.10 annd for the third factor it was 1.30. Seven
itens (23, 22, 15, 10, 45, 36, annd 34) loaded on the first factor.
All of these itens refer to the expression of positive ennotions such
as laughter, liking, and affection. ‘Ilms, this factor was labeled
"Ebcpression of Positive Emotion" . It accounted for 14% of the
variannce. Itens and their loadings appear in Table 5.

AscanbeseeninTableS, fonritens (1, 12, 7, 26) loaded
highest on the second factor. All but one of these itens (i. e., #12)
concern tie expression of negative auctions (e. g., anger,
disappointment). ‘Ihus, this factor was tented the "Expression of
Negative Emotion" . It accounted for 10% of the variance.

Five itens loaded highest on the third factor. All of these
items (30, 29, 8, 18, 14) pertain to liking (18), love (8), gratitude
(29, 30) and apologizing (l4) . Because these auctions center aronnd
relational concerns this factor was retained as different fron the
first factor and it was termed "Expressions of Intimacy". This third

factor accomted for 12% of tie variannce.

58
Table 5

'Ihree Factor Solution for Revised Em

a gift, I wouldn’t knonr how to react.

Ioaggg'
Factor 1 Factor 2 Factor 3

Itenn E; tive Negative Intimagy

23. Watching television or reading .72 .07 .10
abookcanmakenelaughontlond.

22. When I am aloe, I can make myself .67 -.05 -.01
laugh by rerenbernng smething
fron the past.

25. I often laugh so hard that my eyes .58 -.02 .06
water or my sides ache.

10. I laugh a lot. .50 .02 .24

45. I show that I like sonneoe by lugging .49 .29 .35
or tondnirg that person.

36. I often tonch friends during .42 .19 .39
conversations.

34. My laugh is soft and subdued. (-) .30 .20 .05

1. When I amangrypeoplearouninne .07 .67 .12
usually know.

12. People can tell fron my facial .12 .51 .15
expressions how I am feeling.

7. I always express disamoinntmennt when .00 .50 -.06
thingsdon'tgoasI’dlikethento.

26. If soneone makes ne angry in a -.03 .50 .01
public place, I will "case a scene."

30. Whenever people do nice things .07 -.05 .74
forne, I feel "ertonthesport" and
have tronble expressing my gratitude. (-)

29. If a friend surprised ne with .02 -.08 .67

59
Table 5, continued
leaning
Factor 1 Factor 2 Factor 3

8. I often tell people that I love then. .20 .41 .49
18. When I really like soneoe they knew it. .24 .33 .47

14. I apologize when I have done sonething

wrong. .19 .13 .24

A subsequent confirmatory factor analysis (using PACKAGE, Hunter,
1988) wasconductedusingtheclusterswhichenergedinthe
exploratory analysis. Each cluster had a substantial reliability
coefficient considering the enall number of itens in each; for the
expression of positive auction cluster, °< = .74; for the expression
of negative enortion cluster, 0‘ = .63; and for the expression of
inntimacy cluster, 9‘ = 67. The cluster correlation matrix is shown in
Table 6. All of the clusters were positively correlated althongh the
expression of negative enotion cluster was net correlated as strongly
with either the expression of positive enetion cluster or the
expression of intimacy cluster as tlese two clusters were correlated

with each other.

60
Table 6
Cluster Intercorrelation Matrix for Three Factor Solution

FortheRevisedEEDScale

EEP EEN INT
EEP 1.00 .29 .58
EEN 1.00 .37
M 1.00

Note. EEP = Expression of Positive Emotions; EEN = Expression of
Negative Emotions; INT = Enpressions of auctions of
Intimacy/Relationship.

S__gc_ig_l fiirabilitv annd Anbivalm

Correlations were conpnted between scores on the Am, the EEJQ,
their subscales, and the nerlov-Crwne Social Desirability scale to
eeurethatthenewscaleearenetcontaminatedbysocial
desirability. The correlation between the AH) annd the Marlow-Crowne
was significantly negative (1‘ = -.20; p < .001). For both of the AH)
subscales the correlation with the Marlow-Crowne was similarly
negative and significant. For the annbivalence over expressions of
positive enction subscale, the correlation was -.16 (p < .01) annd for
the ambivalence over expression of auctions of entitlement the
correlation was -.22 (p < .01). The correlation between the Em and
the Marlow—Crone was also negative (;= -.11; p < .03). For the three
expressiveness subscales, correlations with tie Marlow-Crone were
negative but nnonsignificent.

As a test of convergent validity, correlations between scores on

61
theAEn, its subscalesandtheRaulinInteseAnbivaleee sonlewere
conputed. As expected, the AH) was significantly positively
correlated with the Raulin I-A scale (1 = .35; p < .001). The
subscale measure of annbivalence over egressios of love was also
significantly positively correlated with the Raulin I-A sale (3 =
.32; p < .001) as was the subscale for anbivalence over expressions of
entitlenent (r = .30; p < .001).
Sex differences

Pearson correlation coefficients were conpnted between sex of
subjectandscoresontheEEn, Am, andtherespectiveclusterscoree
forthesetwoscales. Because fennalevescodedasZandmaleas 1,
positive correlations inndicate that wonen scored higher on the scale
in question. As was predicted, wonnen scored significantly higher on
theEEnthandidnen (;= .15;p< .005). Themeanforwonnenonthe
EEQ was 4.72 while the nean fornen was 4.47. Also as expected, wonen
reported thenselves as more expressive of positive auction and
intimacy (_r; = .14; p < .009 and r = .17 p < .002, respectively).
Contmry to predictions men were net nere expressive of negative
enotiosthanwonenalthonghtnecorrelationwas inthepredicted
direction (; = -.06; n.s.).

With regard to the Alan, predictios were partially supported.
Wonenscoredsignificantlyhigherthanmenonthem (r= .10; p<
.05). ‘Ihemean forwonenontheAEDwas 2.99 whilethennean formen
was 2.88. However, with regard to the subscale scores predictions
werenotsupported. Ithadbeenpredictedthatnnenwculdbemore
annbivalent than woten over the expression of positive enotions while

wonen would be more ambivalent over the expression of entitlenent

62
emotios. Contrarytopredictios, vunenscoreihigherthannnenon
the ambivalence over egression of positive auction subscale (; = .11:
p < .03) while there was re sex difference on the annbivaleee over
auctions of entitlenennt onbscale (_r = .03; n.s.).

Table 7

Correlatios for Em, Am and Cluster scores, Sannple 1

1. 1a. 1b. 2. 2a. 2b.
1. Am -
la. Positive .79 -
1b. Ehtitlenent .88 .71 -
2. E2152 -.24 -.31 -.10 -
2a. Negative -.19 -.21 -.08 -.09 -
2b. Positive -.12 -.05 .03 -.03 .81 -
2C. Inntimacy -.30 -.40 -.15 -.11 .76 .81

Me. For _N = 292, _r’s 2 .10 are significant (p < .05); _r’s 2 .13 are

significant (p < .01).

Correlation§ betwee_n me Am g E

Tanble7 showsthecorrelatiosbetweentheAEn, its subscalesannd
theEEQarditssubscales. Itwaspredictedthatthetwosalesvmld
be negatively correlated and this prediction was born ont (; = -.24) .
The anbivaleee over egressing love subscale was negatively
correlated with the Em and the expression of negative enortion and

intimacy subscales, however it was net correlated with the egression

63
of positive ennotion subsale. Tie anbivaleee over egression of
entitlenent subscale was significantly negatively correlated with the
EEQ, and the egressionn of intimacy subscale but was net significantly
correlated with either the egression of negative enetion or
egression of positive action subscales.

Several conclusioscanbedrawnformtheresults ofStndyl.
First of all while the AFT) is best characterized as unidimesional,
the BBQ taps three different subsets of egressiveness: expression of
positive enotions, egression of negative enetions and egressions of
intimacy. The factor structure of tie Em is net surprising, given
that the egerience of positive and negative auction has been
coeeprtualized as indepedent (Dieer & Firms, 1985) . Whether or
net the individuals experieeed differing degrees of positive enotions
ornegative enetionscannetbeaddressedbyStudyl, however. The
question of whether individuals wle feel more of a particular quality
ofenetionegressmoreofthatenctionrenainstobeaddressed.

PredictionsabonttheotheranalysesinStudylwereconfirned
fortnenestpartaltlnghthereweresoneinportantecceptios.
Neither the AEQ nor tne Em was positively correlated with social
desirability. These scales were, in fact, negatively correlated with
the Marlon-Crone scale. This negative correlation indicates that for
an individualtoscorehighlyontheAmreorshenustbewillingor
able to admit to negative affect. This conclusion is a sensible one
althongh it is importanttonetethat itrevealsoeweaknessofthe
Amineasure. TheAEDdoesnettapintotheannbivaleeethatmaybe

felt by individuals who do net or cannnet admit to experiencing

64
negative affect. This coelusion would indicate that the scale
cannnet, for instance, uncover uncoscions conflict over egression.
lIheAmrequiresanopennnesstotheawarenessofnegativeaffectin
oe’s life. This snertccmning may net be fatal, however. Ehnnncns and
King (1988) have denonstrated the inplicatios of coscions conflict
for psychological and physical well-being. Still, it will be
inportanttobearinmindthattheAEDdoesnetpurporttoneasure
unconscions conflict or negative affect whidn an individual will net
acknewledge.

The negative correlation between the ER) and the Marlow-Crone is
sonewhat more puzzling than the negative correlation between the AEQ
and the Marlow-Crone In order to further examine this correlation,
Pearson correlation coefficients were coguted between all of the
itensintheEEQandthemrlow-mene. Fiveofthe16 itenswere
negatively correlated with the Marlow-Crowns. Two of these iters
concerned laughing (23 and 10 inAppedixA). (be concerned facial
egressions (12 in Appedix A), one apologizing (14 in Appedix A),
and oe love (8 in Appedix A). The correlation with the apologizing
itenmakessonesenseinnthatinordertorespodpositivelytothat
itenoenustadmittoawronqioing-mthimthatmnldnetbe
egected from one scoring highly on the Marlow-Crone. However, with
regard to the laughter, facial egressios and love itens the negative
correlation is net as easily understood. Pennaps trese correlatios
can be best described in tens of self-disclosure. Emotional
egression has previonsly been disonssed in terns of self-disclosure.

ItislikelythatoewhoscoreshighlyontheMarlOw-Crcwneisnot

65
very self-disclosive. In fact, by respodirg positively to the itens
in the Marlow-Crowns individuals are net being self-disclosive and so
the negative correlation with a measure of expressiveess, the Em, is
net only reasonable—it might have been egected.

The AR) was significantly positively correlated with the Raulin
I—A scale, suggesting that it also taps annbivalece. It is inportant
toncte, however, thattheAEQismtsinplyequivalenttotheRaulin
scale. Inn the first place, the two scales were highly correlated but
net perfectly correlated. Secodly, these scales did net behave in a
perfectly parallel manner when correlated with the m2. Although the
Raulin I—A scale was negatively correlated with the Em (r = -12; p <
.05) , this correlatios was net as strong as us correlation between
theAEnandtneEEQ. WhiletleRaulinI—Ascaletapsintese
annbivalece in geeral, tie Am taps specifically annbivalece over
enotional egression.

With regard to the third prediction, cocerning sex differeces,
only the first general statenent was wtelly supported. Wonen were
nere expressive of intimacy than men and nere expressive of positive
enotionthannen. Aspredictedwonenwerenereannbivalentover
innhibition of egression, even thogh trey were more expressive.
Theseresults supporttheideathatwonnenarelessconfortablethan
hen with inegressiveess as an enetion managennent strategy. Also,
these results suggest that, as postulated earlier, althongh men are
less egressive, they may be confortable within their
inengressiveness It steuldbenetedherethattherewasnesex

differece in scores on tte Marlow-Crowns. ms, the differences in

66
expressiveess and ambivalece over expression between male and fenale
subjectsisnetlikelytobeduetoamalereluctancetoself—disclose
or to admit to negative affect. Predicted sex differeces did net
emerge with regard to the expression of negative auction.

The predictios for sex differeces in annbivalece over
egressios of negative enetios and love were also net supported.
Results indicatedthatwonenwerennereannbivalentthannenabort
egressios of love and abont inhibition, while re se: differece
energedwithregardtoenetionsofentitlenent. Theseresultssuggest
that individuals who are highly expressive of certain enotios are
aneivalent abort those very enetions. Thus, rather than being
oconpied by problens with the enetio's abont whidn oe is
inegressive, individuals seems to worry nere about those enetios
they do egress.

Aswaspredicted, tteAEDandtheEmwerenegatively
correlated. This correlation suggests that individuals who are
annbivalent about egression ted to be inegressive This correlation
may explain why nenegression has been associated with ill-being in
pastresearch. Itmaybethat individualswleseemtosufferthe
adverse cosequeces traditionally associated with nonegression are
actually suffering fron the cosegnences of annbivalece. This
statement supports the contention made earlier that ambivalece may
relptoseparatethosewhoarehealthily innexpressive fronthosewho
areacmallyinhibited. Suchastatenentnlstbetakenasnere
conjecunre, giventhescopeofStudyl. Whilethe firststudymakes
ne strong clainns about the relative inportance of egression and

67
ambivalecetowell-being, thesecodsudywilladdressthis issue
nere directly.
Study 2: Application

Thepurposeofthesecodstudywastodeteminewhetherornet
the anbivalece over actional expressiveess scale, the actional
egressiveess scaleandthefactorswhichaergedinStudylpredict
well-being. Literature reviewed previosly presents equivocal results
pertaining to the relationship between expression and well-being. The
secod study will examine the contributios of egression and
ambivalence over egression to psydelogical and physical well-being.
Inthisstudy, scoresontheAm, theEEDandtwoct‘rerneasuresof
expressiveness, peer ratings and the Affective Oammication Test
(A.C.T.; Friedman, Prince, Riggio, and DeMattea, 1980) will be used to
predict psycl'elogical and physical well—being as measured by
self-report questionnaires, daily reports of neod and physical
well-being, as well as objective neasuree of physical synptonnatology.
Because of the low nimber of nen participating, geder differences
willnetbeaddressedinthesecodstudy.

Predictios

1. m‘mflmW. TheAmispredictedtobepositively
associated with negative mood. The Am is predicted to be negatively
associated with positive neod, based on the rationale that ambivalece
may lead to inhibited egression and that egression feeds back into
actional engeriece. Thus, individuals wte inhibit egression of
positive action sholld egeriece less positive action. Scores on
the EB), the A.C.T., and peer ratings of egressiveness are net

68

predicted to be related to daily mood.
2. Qgestionnaire measures 9: mm gum. It is
predicted that Am scores will be positively correlated with measures
of ill-being and negatively correlated with measures of positive
well-being. Because expressiveess may be correlated with
extraversion, scores on the three meaenres of egressiveness are
predicted to show a positive correlation with positive affect, but, as
has been argued previonsly, scores on the egressiveness measures
shonld net be correlated with measures of negative affect. These
predictios are in keeping with the rationale that egressiveness
itself is re insulation against the adverse psycl'elogical cosequeces
which have been associated with inegressiveess
3. Measures 9: wical W. With regard to physical
symptons, scores on the AH) are predicted to be positively correlated
with daily symptom scores, scores on qnestionnaire nneasures of
synptons, themmberofI-lealthOentervisits, andthenumberof
different illnesses. Engressiveess scores are predicted to show no
relatioship with these synptonn indices.

Matted
Subjects. A sample of 48 Michigan State University undergraduates (35
fenales and 13 males) participated in a 10-week log course and
research project entitled, "Research on Goals, mod, and Health."
Subjects were from varions academic levels and in varios courses of
study. Theywererecruitedviaannunncaentspostedinthepsychology
departnnent. Participation was open to virtually anyoe and
participants received 3 honrs of conrse credit. Althongh subjects

69
weremdergraduatestheycanefronnvarionsagegronps. Amogthe
women, the age range was 19 to 47, with 5 of the subjects falling
betweenage33 and47. Themeanageforfaaleonbjectswasma
years. Amogthenen, agesrangedfronl9to27. 'nenneanagefor
male subjects was 21.6.

Procedure
m 32993.5

Dailyneodreportswerecogletedbythesnbjects for 21
conseontive days. The form were cogleted eadn day, oe at the
"middle" of the subject's day (usually in the mid to late afterreon)
andonebeforethesubjectwenttosleq). Thesenneodreports
cosisted of a list of adjectives describing varios positive and
negative actions (happy, joyful, pleased, enjoyment/fun, unnhappy,
angry, annxions, depressed, and frustrated). ‘Ihe adjectives listed on
these reports were chosen on de basis of earlier factor annalytic work
by Dieerand Runners (1984). Subjects ratedtheectenttowhichthey
had egerieced the auction listed during the part of the day prior
to completion of the form on a 6-point scale with 1 indicating "net at
all" and 6 indicating "ectraely men". The ratings on the positive
affectadjectivesweresnmlnedovertl'ezldaystoprodnceaconposite
positive affect score and the ratings on the negative affect adjective
were sunmed to produce a conposite negative affect score. These
conposite scales have been utilized in other studies (see Dieer, 1984
for a review) and their taporal reliability and innternal cosisteey
coefficients approach .90.
Egoholgical Well-Beg m

During the first weeks of the term-log study, subjects capleted

70

a several questionnaires nneasuring psychological well-being. These
questionnaires included the Hopkins Syngton Checklist (Derogatis,
Lipnan, Rickels, Uhlenhuth, & Ccvi, 1974) which neasures psychological
synpvtons on several dinnersios, inclnding Depression, Guilt, Annxiety,
Paraneia, Phobias, and Cbsessive-Cogulsive tedencies. Subjects also
connpleted the well-being scale fron the Differential Personality
Questionnaire (DPQ; T‘ellegen, 1979) , the Satisfaction with Life Scale
(SWLS; Dieer, Ehmcns, Larsen, and Griffin, 1985), Rosenberg’s
Self-Esteenn Scale (Rosenberg, 1963) and a revised version of the
Bradburnn Affect Balance scale (Bradburn, 1969; revised by Warr,
Barter, & Brcwnnbridge, 1983) , which assesses the extent to which the
respodent has recently egerieced positive and negative affect. The
Beck Depression Inventory was also cogleted by subjects as a further
measure of psychological ill-being. Also presented at this tinne was
the Affective Communication Test (A.C.T.; Friedman, Prince, Riggio,
DeMattea, 1980) a treasure of enotional egressiveness.
gnygical Symptomatoloqv

Five different measures of physical synprtons were obtained for
these subjects. First, subjects cogleted tte Pennebaker Inventory of
Linnbic Ianguidness (PILL; Pennebaker, 1982) a questionnaire neasure of
cannon physiological synnprtons. Secod, subjects capleted the
Sonatization scale fron the Hopkins Synpton Checklist (Derogatis,
Lipton, Rickels, Uhlenhuth, & Ccvi, 1974). Third, the daily ncod
reports ineludedasynprtonchecklist. Subjectsdeckedthesynptons
whichtleyhadexperiecedonthegivenday. Thesynptonslistedon
the daily reports included, readaches, stonach ad's/pain, chest pain,

71

runny/cogested nese, conghing/sore throat, fainntnness/dizziness, out
of breath, acre/pimples, and stiff/sore Insoles. The nine categories
werechosenbaseduponfactoranalyticworkonalarge listof
physical synptcms (Pennnebaker, 1982). Friedmn n. Bocth-Kewley (1987)
state that stomach ache/pain, chest/heart pain, ont of breath, and
sore/stiff muscles can be cosidered preonrsors to serions illnesses
which have also been related to action: ulcers, heart disease,
asthma, andarthritis. Thesesynnptonnreportsweresnmuedcverall
daystocreateaglobalsynptonscoreforeadnsnbject.

Subjectswereaskedtosignaconsent formgivingthe
experimenters permissiontoaeeeestheirl'ealthrecords atthe
University health center. Subjects were assured that their records
would be kept conpletely confidential. Given these records, a rater,
blind to the subjects’ anbivalece scores, connted the number of
visitsmadetothehealthoenteraswellasthenmberofdifferent
illnessesdiagcsedinthepastyear. ‘nnefolrthandfifthnneasures
of synetonatologywereprovidedbytlesetwecomts, ofhealthcenter
visits and of different illnesses. Tne number of visits ranged from 0
to 18, with a nean of 2.89 and a standard deviation of 4.35. The
nnmberof illnessesrangedfronOtolo, withanneanandstandard
deviation of 1.81 and 2.19 respectively.
Beer m

Lastly, subjects provided the names and addresses of individuals
wneknewthenwell. Theseindividualswerecontactedandaskedto
complete a a brief packet of questionnaires. Included in this packet
was a 6-itan expressiveness scale ( = .59) abort the subjects.

These "peer ratings" of egressiveess provided an additional neasure

72
of egressiveness. All subjects were asked to provide 10 addresses.
onlythosesubjects forwhonat least2peersrespodedtothemailer
wereusedinpeerratinganalyses. 'netotalnemberofpeers
respodingwa5207. Thenmberofpeerspersubjectreturnningthe
packetrangedfromnenetolo. Tremeannnunberofpeerspersubject
was 4, with a standard deviation of 1.9. Of the individuals who
conpleted and returned the peer questionnaire 58% were frieds, 14%
were parents, 13% were otter relatives, 4% were boyfrieds or
girlfrieds, and 11% shared sole otter relationship with the subject.
Three subjects were eliminated fronn annalyses involving peer data
because of lack of peer respodents.
Results and Discssion

Threesets ofannalyseswereperformedonthedatacollected in
Study 2: 1) Correlatios for measures of egressiveness and
annbivalece of egressiveness were coguted to see if the correlatios
fronn Study 1 replicate. 2) Pearson correlatios well-being measures
andAEnandegressivenessneasureswerecapnted. 3) Partial
correlations of Am and E512 scores as predictors of symptons,
ill-being, negative affect, positive affect and positive well-being
were conputed.

TteneansfortleAEDandtheEEpforS‘udwaerecogarableto

those forStudyl, 2.84 fortheAmand4.33 fortheEEn. TteAEDhad
a standard deviation of .60 and the E112 standard deviation was .63.
Correlatios between the BBQ, the A.C.T. and tie Am are presented in
Table 8. The BBQ and the A.C.T. were signifiantly positively
correlated. Pearson correlatios were also coguted within peer

73

ratings so that the each subjects' self reported expressiveness could
be corpared with each of his or her peers' estimates of expressiveness.
Peer ratings of expressiveness were significantly positively correlated
withscoresontheEEn (;'= .43; p< .001, y=207) andscoresonthe
A.C.T. (_r; = .21; p < .01)

FortheAEQandtheEEn, resnfltsweresimilartoSb.dy1,(;=
-.25; p < .05) . However, the A.C.T. was not significantly correlated
with the AEQ. Peer ratings of expressiveness were significantly

negatively correlated with the Am (r = -.31; p < .001) .

Table 8
Correlations for Am, Cluster Scores

and Ebcpressiveness Measures, Sannple 2

1. 2. 3.
1. Am -
2. BBQ -.25 —
3. Peer Rating -.31 .43 -
4. A.C.T. -.08 .36 .21

Note. For all but the peer ratings of expressiveness, _Ij = 48. For
g: 48, £2 .24 are significant (p < .05); _rz .33 is
significant (p < .01). For peer ratings, 11 = 207, r 2 .10 is

significant, p < .05 and I _>_ .16 is significant, p < .01:

74

.ImmoMHOAOAnm u .uo:ouum newoouudm u .uudm newnosm u .aonm uhHOMxe< u .x:<

”newnaoheon u .909 “abwnaonloo obfiuaomnno u .o.o "massacol sewage: you .oadom acouu<
o>mudwoz shonoaum u moz .m "huouaobam newnumueon Moon u Ham “Hummus o>wuumue huwdo u <2
“acouu< o>uawoom ehsnodum u .nom .m meadow an“; new: newscauowudm u mazm noumd::o«auuoa
humadoonuom unwasououuwa on» loan canon masonndaox u on: “Hummus o>wuwnon mama: u

em ..ee. v a. eauoeeeauea use an. A e “Ame. v a. eaaoeeeamea use as. A s .me u 2 has .oeoz

 

va. mm. «m. mm. no. No. we. on. on. Na. a~.n v¢.n m¢.n ¢¢.I .bom .6 .mu
In so. an. an. no. no. Hm. av. ow. no. mN.I wv.l N¢.I H¢.I .HOAOhum .wn

on. an. >5. am. he. an. ow. on. Hc.n mN.I ¢N.I HN.I Hawsw .m~
II «n. no. em. Hm. «V. an. «N. «N.n vv.l cv.l ov.n .dhdm .NH
In on. av. NM. ov. on. «N. N~.I «m.n mn.n ¢~.I .noam .nm
In am. mm. mm. NM. ha. no.l NN.I O¢.I «v.1 ..xu< .Oa
1: mm. am. me. ha. o~.n nm.n mv.l av.l .non .m
In «a. on. em. cN.I ON.I cN.I Nm.l .U.O .m
assume:
I: am. am. no.1 c~.n an.l an.l .moz.m .5
nu ma. mo.n vm.n va.n mH.n Ham .m
In ac. NN.I wN.I oN.I <2 .m
In an. on. vH. .uom.m .v
an be. no. mutm .n
II «n. can .N
In <m .H
«a ma Na aw cg m m h w m e m N a

wowoanmaoz Accumoaonohmm no «nuance: nucleon occaudaouuoo

a wands

'75
Amb__xa_1___i ence oer Mien. W and W
wil-m

Correlations for the measures of psychological well-being are
shown in Table 9. Generally all of the measures of positive well-being
were positively correlated and the neaanres of ill-being were
positively correlated with each other. As would be expected, neasures
of positive well-being were negatively correlated with measures of
ill-being.

Pearson correlation coefficients were conpnted for neasures of
psychological well-being, the Alan and the neasures of expressiveness.
The results are shown in Table 10. Generally, the AM was negatively
associated with neasures of psychological well-being and positively
associated with treasures of psychological ill—being, as predicted. 'Ihe
AEQ correlated significantly negatively with life satisfaction as
neasuredbytheSWlSandwithself—esteemasmeasuredbytheRosenberg
Scale.

Positive correlations were found for the Am and daily negative
affect, obsessive/compulsive tendencies, depression, paranoia, and
phobias. Annbivalennce over enotional expression was also positively
associated with depression as neasured by the Beck Depression Inventory
and with negative affect as neasured by the Bradburn Affect Balance
Scale. In sum, the AH) behaved largely as predicted, correlating
positively with neasures of poor functioning and negatively with
neasures of healthy functioning.

With regard to treasures of expressiveness, correlations also

emerged as predicted. 'nne Em correlated significantly positively with

76
the DPQ neasure of well-being (I = .27; p < .04) and positive affect as
neasuredbytneBradhlrnAffectBalarcescale (r= .24, p< .05). Peg-
ratings of expressiveness correlated significantly positively with the
DPQ as well. In terns of ill-being, the E120 was positively associated
with daily reported negative mood (I = .23, p = .06), anxiety (r = .25;
p < .05) and guilt (_r = .24, p < .05). Peer ratings of expressiveness
did not correlate with neasures of ill-being. The A.C.T. did not
correlate significantly with any of the measures psychological
functioning. To summarize, the neasures of expressiveness were
generally unncorrelated with well-being neasures, although sore
correlations did ennerge for positive functioning and expressiveness .
Those correlations that existed between expression and ill-being were
positive nnot negative as would be predicted by a perspective

orphasizing expression per se as healthy.

77
Table 10
Correlations Between Measures of Egressiveness,

Annbivalence over Ebcpression, and Psyclnologionl Well-Being

Well-Begg' Scale ED 5&1; 3 5m
PA .05 -.08 -.02 -.06
DPQ .27 .05 .30 -.19
SWIS .19 .03 .03 -.25
Self-Esteem -.03 -.O3 .09 -.40
Bradburn PbSitive .24 -.02 -.07 .05
NA .23 .19 -.05 .28
Beck Depression -.02 .09 -.05 .39
Bradburn Negative .10 .01 -.05 .24
H .

Obs/Conn .11 -.03 -.lo .42
mobia -.O4 .02 -.10 .31
Parannoia -.06 .08 -.16 .40
Depression .16 .14 .04 .26
Anxiety .25 .07 .05 .15
Psychoticism . 05 . 19 - . 03 . 26
Guilt .24 -.13 -.03 .30

N_ot_e. Em = Emotional Expressivenness Qestiomaire; A.C.T. =

Affective Conmunication Test: H2 = peer rated expressiveness:
Am = Anbivalence over Egression Questionnaire; PA = Daily
positive affect; DPQ = Well-being scale fron the Differential
Personality Questionnaire; SWIS = Satisfaction with Life
Scale; B. Fee. = Bradburnn Positive Affect; NA = Daily negative
affect. For H = 48, _r' 2 .24 is significant (p < .05); _r: 2 .33
is significant (p < .01). For peer ratings, lg = 207, r z .11

is significant (p < .05) and r 2 .16 is significant (p < .01).

78

_____Anbiva1enoe oer expression. W. as; Ensign; gel-sins-

Correlations between the measures of physical conplaints are
reported in Table 11. Although the questionnaire measures of physical
symptonatology were all positively correlated, only the PILL was
related to the number of illnesses with marginal significance (p <
.06) . None of the questionnaire measures of physical ill-being were
significant predictors of health center visits.

Table 11

Correlations for the Measures of Physical Syntptcmatology

1. 2. 3. 4.
1. Average Daily Syn'ptons -
2. Health Center Visits .07 -
3. Nnmber of illnesses -.06 .81 -
4. PILL .23 .11 .22 -
5 . Sonatization . 36 -. O1 . 11 . 26

Note. PILL = Pennnebaker Inventory of Linnbic Ianguidness;
Sonatization = Sonatization Scale fron the Hopkins Synpton
Checklist: Inventory. For g = 48, r 2 .24 are significant (p <

.05); r 2 .33 are significant (p < .01).

79

Pearson product-mnent correlations were also conpnrted for the
measures of physical synptonatology and the measures of psychological
well-being. These correlations are presented in Table 12. Generally,
the measures of negative affect and psychological ill-being were
positively correlated with questionnaire neasures of physical
synptonatology. Daily synptons were positively correlated with daily
negative affect, the 801, the cbsessive-conpnlsive, anxiety, and
depression subscales fron the Hopkins Synpton Gecklist. Health center
visits were positively correlated with the Bradburn negative affect
scale. The nunnber of differennt illnesses was positively correlated
with the BDI, the Bradburnn negative affect scale and the
obsessive-compulsive and depression subscales fron the Hopkins Symptom
Checklist.

Questionnaire neasures of positive affect and well-being were
negatively associated with the General Severity Index and the
Somatization subscale fron the Hopldns. [hily positive affect was
significantly negatively correlated with the PILL. Daily synptons were
not correlated with any of the neasures of positive affect. Health
Center visits and the number of different illnesses were negatively

correlated with daily positive affect.

80
Table 12
Correlations Between Measures of Physical and Psychological Well-being

grass» rain. Lsslr' BQE' Illnesses

PA -.10 -.32 .07 -.24 -.24
DPQ -.35 .02 -.03 .08 .01
SW18 -.22 -.O7 .07 -.01 -.03
Self-esteem -.30 -.10 -.04 -.13 -.16
Bradburn Positive -. 13 . 04 . 09 . 10 . 09
NA .32 .32 .39 .16 .18
BeCk Depression .31 .21 .31 .14 .26
Bradburn Negative . 22 . 22 . 10 . 24 . 37
G. Sev. .65 .43 .23 .07 .16
ngkins

Obs/com .57 .35 .30 .06 .26
PhObia .46 .32 .10 -.17 -.05
Paranoia .40 .20 .09 -.08 .03
DepreSSion .40 .37 .16 .19 .32
Anxiety .62 .36 {.26 .04 .19
PsyChoticism .29 .36 .08 .02 .16
Guilt .29 .40 .29 -.10 .07

H913. Sonat. = Sonnatization Scale fronn the Hopkins Sylnpton Checklist;
PILL = Pennebaker Inventory of Linbic Ianguidness. PA = Daily
positive affect; DPQ = Well-being scale fron the Differential
Personality Qlestionnaire; SWIS = Satisfaction with Life Scale:

NA = Daily negative affect; G. Sev. = Index from the Brief Synpton
Inventory. For H = 48, _r: 2 .24 are significant (p < .05); _r; 2 .33

are significannt (p < .01) .

81
Table 13

Correlations between Measures of Expressiveness,
Anbivalence Over Expression, and Physical Synptons

was scales
gm Measures 2‘.) Log-.1. EB fl
Sonatization .19 -.01 -.0‘7 .20
PILL .05 -.05 .12 .18
Mean Daily Symptons .09 .oo .oo .06
Health Ctr. Visits .21 .07 .07 -.08
Number of Illnesses .12 -.06 -.Ol .09
No_t_e. Em = Emotional EcpresSion Qiestionnaire; A.C.T. = Affective

Communication Test; H? = Pser ratings of expressiveness; AEQ =
Ambivalence Over Ebcpression Qestionnnaire; Sonatization =
Sonatization Scale from the Hopkins Synpton Checklist; PILL =
Pennebaker Inventory of Linnbic Ianguidness. For all variables
exceptthepeerratings, N=48. For_N=48, 3:2 .24 is
significant (p < .05); r 2 .33 is significant (p < .01). For
peer ratings, N = 207, r > .11 is significant (p < .05) and r

2 .16 is significant (p < .01).

Pearson correlation coefficients were counted for the

expressiveness neasures and the 5 neasures of physical conplaints.

These correlations are presented in Table 13. mile the questionnaire

82
measure of synptonatology was correlated with the Am, the objective
neasures of physical distress (i.e., number of health center visits and
nunber of illnesses) were not correlated with either expressiveness or
ambivalence over expression. Daily synpton reports were nnot associated
with either expressiveness or anbivalence. One surprising, though only
marginnally significant result is that the Em was positively correlated
with health center visits (; = .21; p < .1).

In order to further test the predicted relationships between
ambivalence over expression and expression, partial correlatios were
conpltedbetweenthevariousneasuresofpnysical illnessandtheAEQ
partialling out the influennce of first the Elm then the A.C.T. and both
of these neasures of expressiveess. The results of these partial
correlations were not as expected. AlthoLgh the correlations between
the AEQ and the questionnaire neasures of synptonatology remained
positive, nno correlatios energed between daily synptons, health center
visits or number of illnesses.

Next, anediansplitwasperformedonthescoresontheAEnand
EEQinordertotesttherelationshipsprtposedinFigure 3. Two (high
vs low ambivalence over expression) by two (high vs low expressiveness)
annalyses of variance were conducted for each measure of synptonatology
as a depedent variable. A min effect for anbivalence over expression
was expected, however, no significant effects were fond. Iast, a
sinnilarprocedurewasused, thistinneusingonlytheupperandloer
thirdsofscoresontheAmandEm. Subsequentanalysesofvariance
revealed nno significant effects. Analyses were repeated using only
fenale subjects and only those subjects within a traditional college

83
age range (i.e., < 26 years). Once again, neither anbivalence over
expression nnor expression predicted health status.

Tosunmnarize, theresultsofthesecodsundycanbedividedinto
twoparts, theresultspertainingtotherelationshipsbetween
expressiveness, anbivalence over expression and psychological
well-being and the results pertaining to expression, anbivalence over
expression and physical well-being. With newect to the former,
results were basically as expected. Ecpressiveness correlated with
sons measures of positive well-being but did nnot correlate negatively
with neasures of psychological ill-being. The positive correlation
betweentheEEnandneasuresofnegativeaffectmnsdirectlycontrary
to the thesis that expression is a positive, advantageous behavior in
and of itself. These positive correlations are sonewhat surprising and
three explannationns will be suggested for this finding. The first two
explanations assume tint the self-reports of negative affect do in fact
reflect differences in affective experience between expressive and
inexpressive individuals. The positive correlations between the
neasures of lowered psychological functioning and expressiveess may be
explained via the feedback relationship that has been postulated as
existing between expression and enotional experience (e. g., Iaird,
1974) . Individuals who are very expressive of emotions in general
should experience more of all of these auctions—positive and
negative. An alternative explanation would be that individuals who are
expressive actually feel more negative affect because of their
expressive behaviors. That is, given the previous discussion of the
cultural disdain for enotional display, it is nno woder that these

84

individuals may experience enbarrasenent, shame, or guilt over their
expression. A f innal explanation for tress positive correlations does
not assunne that the correlations reflect real differences in affective
experience between expressive and inexpressive individuals. The
positive correlation between measures of ill-being and expression may
be due to the fact that reporting negative affect is, itself, an
expressive behavior. Thus, individuals who are expressive of their
enotional states are more likely to report their negative affect than
innexpressive individuals—regardless of whether there is any difference
in the affective experience of these he groups. Tne data collected in
thetwostndiespresentedrereismtsufficienttoconfirmordiscomt
these three possibilities. In any case, the positive correlations
thenselvesdoat leastcastsonedoubtontheideathatexpressive
individuals are more enotionally healthy than innexpressive individuals.

Ambivalence over expression, on the other hand, was consistently
correlated with neasures of poor psychological funnctioning. It also
correlated negatively with sons measures of psychological well-being.
These correlations are in keeping with initial predictions. A
connparisonoftheAEpandtheneasuresof expressiveness illustrates
the accuracy of the contention that expressiveness aloe can reveal
little about an individual's psychological health status.
Expressiveness did not show a negative relationship with measures of
ill-being. on the contrary, annbivalence over expression did correlate
with ill-being. Thus, these results support the thesis that expression
of sanction provide: nno innsulation against the kinds of adverse
psychological consequences that have been traditionally associated with

85
innexpressiveness.

Tiesecodsetofresultspertainstotherelationshipsbetween
expression, antbivalence over expression and physical health. As
predicted, expressiveness was generally nnot associated with neasures of
health status, supporting further the contention that expression of
enotionisnotthekeytol'ealthyendtionmanagenent. Indeed,
expressiveness as measured by the Em was positively correlated with
health center visits. This rather startling finding will be discussed
further below. Ambivalence over expression was positively correlated
with questionnaire neasures of synptone, as predicted. However,
contrary to predictions, anbivalence over expression showed no
relationship with health center visits or umber of different
illnesses.

The finding that the BBQ was positively associated with health
center visits while the AR} was not deserves special consideration
here. Thismnexpectedresultcanbeecplainedthronghreseardnthat
hasbeencoductedintheareaofhealthpsychology, focusingon
that there is considerable variation in synpton perception, experience
and interpretation (Roseetock & Kirsdnt, 1979). Zola (1972, cited in
Rosenstock & Kirscht) reported differences in Italian Americans and
Irish Americans in their synpvtonn reportage. Italian Annericanns teded
to exaggerate synptons while Irish Americans teded to "suffer in
silence". Although there are nany obvious cultural differences between
thesetwogroups, withinthecontectofthisstndyoeofthese

differences—emotional expressiveess—warrants consideration. If

86
synptonreportingisthoghtofasanecpressivebehaviortientlelack
of correlation between synpton reports and the AEQ is understandable.
This conceptualization of synpton reporting also offers further
explanation for the positive correlation between health center visits
and expressiveness.

Researdninl‘ealthpsydnologyalsocastsadifferentlightonthe
variable of health center visits as a measure of psychological
well-being rather than physical ill-being. Within the field of health
psychology, the concept of "readiness to use services" or "orientation
tocare" (Rosenstocl<&l<irsont, 1979) hasbeenusedtodescribethe
seeking out of treatment for synptons. Psychological readiness
involves the belief that one's health is within one’s control and that
health professionals and services are a means of maintaining oe’s good
health. RosenstockandKirsdntdescribethisconceptastheinNerseof
helplessness (p. 178) . In light of the concept of "orientation to
care," reexamination of the reenlts of Study 2 with regard to
annbivalence over expression the predicted positive relationship between
theAEnandhealthcentervisitsnakesnudnlesssensethanthelackof
relationship that energed.

87
Geeral Discussion

Itmaybehelpfultobeginthisdisonssionwitharecapofthe
major findings from both stndies. The first inportant finding for
Study 1 was the unidinnensionality of the Am. The positive correlation
betweentheAEDandtheRaulinI—Ascaleaswellasthenegative
correlationbetweentheAEDandtheEEDledsupporttotheideathat
the Alan taps ambivalence over enotional expression. Also, in Stndy 1,
ithadbeenpredictedthatwonennwouldbemreenpressiveandmre
annbivalent over expression than men. This prediction was supported,
with the added consideration that wonen were nore expressive of and
ambivalent over positive ewtions. Amg the major findings of Study 2
weretherelationshipsbetweentieAmandquestionnaireneasuresof
psychological ill-being and physical synptons and the absence of a
relationship between expressiveness and these measures of ill-being.
Alsoof inportanceinStudwaastheladcofrelationshipbetween
health center visits, number of illnesses and expressiveness or
ambivalence over expression.

Thefirstofthesefindingstobeaddressedhereisthesex
differences fund in Study 1. Given the previous disonssion of geder
differences in expression it sees appropriate at this time to atterprt
to integrate the present provontive findings with past research.
Whilereseardnersintheareacfsexdifferencesinemotional
expression have focused on anger and negative auctions as a stumbling
block for wonnen (e. g., Hochsdnild, 1975; Bernardez-Bonesatti, 1978:

Sheilds, 1987) , the onrrent reenlts indicate that it was positive

88

enortions about which wanen were more expressive and more anbivalent.
Perhapsthebestpointofentryforthisdisonesimistocmsiderthe
conflict that may underlie the anbivalenoe felt by wunen with regard to
positive elation. 'Ihe inrportance of the feninine role of mirturer' has
been discussed previously. Hodnschild (1975) has suggested that wanen
are placed in a conflictful situation routinely as a result of cultural
expectations of nurturance and situatienal denands for neutrality or
enotional detachment. The present results would suppon her
contention. Oursubjectsmayhavebeenreactingtothiskindof
conflict—the desire to function successfully in a mntraditional role
mileatthesametimedealwithdenendstobeeipressiveofcareand
support forothers. Itisinnterestingtonoteaswellthatthe
subjects participating in the first study were all college wenen—those
whocanbesaidtobetryhgatleasttoscneexterrttobalancethat
traditional female role with individual ambitions. It may be that for
these young mnen, it is not negative auction that causes problens but
positive enotions—those that may seen inarprcpriate to their career
concernsbltwhidlarepartandparceloftraditia'nal femininity.

A second possible explanation for the sex difference in ambivalence
over expression is fenale ignorance of display rules for
nonexpression. Previously presented research an the development of
facial expressions and display rule kn'nowledge suggests that fenale
infants areecposedtoawidervarietyof expressionswhilemale
infants are given more explicit training in the inhibition of
expression. It may be that, because they are trained as children that

nonexpression is an appropriate behavior, men are more canfortable than

89

wanen in not expressing any enotion at all. Renenber that previously
mentionedreseardnhasshownthatwunen, notmen, tedtovalue
expressiveness (Shields, 1987) . Because lack of expression is the
status quo within a male daninnated clltnlre (Balswick, 1988), it is not
surprisingthatwanen, whoaremreetpressiveumldalso feelmore
ambivalent. Interestingly, sane of the items in the factor on which
wunenscoredhigherthanmenpertainedtotheinnhibitionof
expression. It wcnuld be fascinating to examine more closely the
mderpimingsofthesexdifferencesthatenergedinflnisshdy.
Particularly, experimental nanipulatia'n studies are recanmeded so that
causal inferencescanbedrawn. Futurereeeardlinthisareamight
involve experimental manipllation of qportunity to express enotions to
see if wenen do feel more uncanfortable than men in situations which
prohibit expression. Also, research might investigate whether lack of
knowledge of situational demands for expression in a situation is
related to heightened annxiety for lien and wunen.

'IheresultsofStldyZprovidesurprisirgevidenceabouttheroles
of expression and an'bivalence over expressim to well-being. Although
it had been predicted that expressim would not be associated with
measures of well-being, this reallt cantradicts many of our assunptions
aboutexpressionandhealth. Itmaybethatthemstinportant result
of this innvestigation is that expressiveness provided no insulation
fmn the adverse consequenms traditienally associated with
inecpressiveness. These results underscore the need for more
sophisticated views of enotienal expressim.

Perhapsthemoststraightforwardresultofthesecadstndywasthe

90
relationship between anbivalence over expression and psychological
distress. 'IheAEnwaspositively correlatedwithammnberofmeasures
of poor psychological functioning. Although the correlations could be
taken as a whole to indicate that anbivalence over expression is
associated, geerally, with psychological pain. War, at least one
of the correlations that energed is qnecifically interesting, since it
relates to research reviewed earlier by Bentler, Engle, Ore-Beutler,
Daldrup, and Meredith (1986) . Recall that in review it was concluded
thatthecanflictovertreexpressionofintenseenctionwasacamm
link between depression and chronic pain. 'Ihe Am correlated very
strongly with measures of depressian. It would be profitable to
utilize this scale in studies like those revieded by Beutler, Bgle,
arc-Beutler, Daldrup, and Meredith to further investigate the link
between depression and conflict over expression. This also suggests a
subject population for when the AH) could be a valuable tool.

While the AH) correlated with measures of psychological distress
and questionnaire neasures of synptane, it did not correlate with the
numberof illnesses sufferedorwithdailyreportedsynptans. Several
explanations seen possible. 'Ihe first explanation is, of course, that
there is no relationship between anbivalence over expression and
physical distress. This explanaticn would hold that arbivalence over
expression is a purely psychological [ileum that does not inpact on
physical well-being. mile the existence of any such phenanenon would
seen highly unlikely given the previous literature review on the
relationship between expressiveness and health, it is still possible

that this particular form of emotional turmil is not physiologically

91
pathogenic. The second explanatian is that anbivalence over expression
isanemotional experiencethatrequixestilneto inpactonphysical
health. longitudinal data on the individuals who were highly
ambivalent over expression would denanstrate whether over tine this
ambivalence inpacts on physical health.

A final explanation for the lack of correlation between the AH) and
physical synprtans is that while anbivalence over expression is
associated with physical health, the current measures of ambivalence do
not measure the construct adeqnately to uncover this relationship.
Other measures of anbivalence might be preferable to the AEQ. For
example, Elmore and King (1988) examined ambivalence as it related to
idiographic personal strivings and found a relationship between
ambivalence and health outcanes. A similar study stressing strivings
about emotional expression might better tap the anbivalence over
expression-symptanatology connectian.

Severalproblensareapparentinthepresentstndiesandthese
deservetobeaddressedhere. aneproblenisthatneeofthehealth
measureswastr'ulyobjective. Oneissuethatisbroughttothe
forefront by this study is the qnestian of how mlch self-reports of
symptomatology are in theneelvee expressive behaviors. It has already
beensuggestedthatthecorrelatianbetweenhealthcentervisitsand
expressiveness may be due to differences between expressive and
inexpressive individuals in synptan reporting. 'Ihus, even realth
certervisitsarenereslbjectivethanmightoriginallyseentobethe
case. What is called for here is a more objective nethodology for

assessing health status. Such a methodology might involve mandatory

92

physicals for all subjects as well as anbulatory heart and blood
pressure monitoring and tests of immune finctionirg. Also galvanic
skin response monitoring might be necessary, as Pennebaker (1985) had
found this to be the best indicator of inhibition.

Ifthereiscnethingthatismadeobviousgiventheresultsofthe
present studies it is the unique challenge that lies before creators of
neasuresofenpressiveness. Twoinportantproblenccanbeseento
plague the neasures of expressiveess used in this investigation and
they are problems that are typical of neasures of expressiveness .
First of all, the EEQ, A.C.T. and peer ratings of expressiveness all
confound expressiveness with emotional experiece itself. Because
these scales cover a variety of different enctions they assume that all
subjects actually have felt the auctions mentioned. 'Ihis is a
dangerous assnmlptien in that there my well be individual differeces
in frequency and duration of various affective states (Ennmons & Dieer,
1985) . Recent investigations by Balswick (1988) have sought to avoid
this pitfall by utilizing scales which include itens which are worded
insudnawayastoseparatetypesofenotimsexperiencedfran
expression. Thus, an item like "I often tell pecple that I love then"
fmntheEEnvmldbereworded'WnenI feel love forsaneoneItell
them often." Itens about other emotiee wonld also be prefaced with
the phrase "When I do feel..." Inthisway, the items are less likely
to confound enotional experience with expression of that experience.
Balswick’s scales provide good models for future measures of
expressiveness.

'Ihesecondproblenwiththemeasuresof expressicnusedinthis
investigation is that they ignore social context of expression. In
noreofthesescaleswastheindividualtowhananenotionwastobe

93
egressed nentioned clearly. 'Ihis disregard for social content
presents an inportant problem, for it my be that individuals are more
likely to confide in or egress auctions to sane individuals in their
lives (e. g., spouse, therapist, parent, fried) than others (e. g.,
boss). Once again, Balswick (1988) provides a neans by which social
contextofegressioncanbeincludedandenamined. Inarecentstudy,
Balswick had subjects fill out egressiveness questionnaires several
times—for different individuals in their social environment. For
instance, subjectswould filloutthequestiamaireoncewithregardto
theirspouse, againwithregardtoafriend, etc. Suchan'ethodology
would offer several benefits to researchers in the area of enctional
egression. Not only would our locwledge of the inportanee of
egressiveness benefit, but our knowledge of the adaptiveness of
individual styles of egression would increase as well.

The current investigations suggest several avenues for future
research, internsofbothinprovingthemethodologyinworkonthe
ambivalence over egression censtruct and in defining the conceptual
franework of enotion research. It is necessary that we begin to focus
our research on enctional egressien on the individual goals which
underlie that egression. Also, ennnasis on egression for its own
sake is no longer justifiable. In addition, the use of measures of
egressiveness alone in health psychology research will not accanplish
the task of differentiating the healthy fran the potentially
unhealthy. 'Itesetwostatenentsarestruglysnnpportedbytlnepresent
results.

'IhererenainsthereedtoworkbeyadtheAEfltowardmeasuresthat

94

ecanpass social contextandthatwill predictphysical
synptanatology. 'Ihese net methods may include different
questionnaires. Projectivetestsmyalsobeneceesaryinordertotap
the underlying anbivalence of those who are nnot able to consciously
state their feelings of anbivalence over egression. Also, more
idiographic, individualized methods may be helpful in assessing
individual egressiveness styles and the adaptiveness of these styles.

Amore idiograpnicwayinwhidnreseardncanbegintoaddressthe
meaning and importance of annbivalece and conflict over egressiveness
goalstrivings istoaskindividualswhytheyfeel anbivalentabout
certain personal strivings. It may be that it is not the ambivalence
itself that is pathogenic but the situation that underlies the
ambivalece. Perhaps, it is the case that only certain mnderlying
constellations of enotionc prancte ill health or adverse enetional
consequences. Imingortantendtionsthatmightbeconsideredare
share and guilt. Additionally, anbivalence that is fostered by goal
conflictmaybethekindthatismostprtbablypathogenic. Wand
King (1988) found that goal conflict is associated with poor health and
psychological well-being. 'Ihey surmised that conflict and annbivalence
may be related—that individuals who egerience chronic conflict my be
prone to becane anrbivalent over certain goals. 'Ihe underlying reasons
for ambivalence may help determine if this is the case. Of particular
interestrerevmldbeanecaminationofthereasensforanbivalece
over strivings dealing with enotional egression.

Afinalpointthatisfittinglyrepeatedlnereasasnmmaryofthe
findings of this investigation is that egression in and of itself is

95
notthekeytowell-being. Rather, itmaybehowme feelsaboutone’s
egressive behaviors that determines the potentially harmful dnaracter
of lackofegression. Theremaybeothercagonents inthe
relationship between enotion and health. We nust continue to look for
those factors that separate innegressiveness fran inhibition—to
distinguishbetweenindividualswhoareegressiveandhealthyand
thosewhoareegressiveandatriskforserious illness. Thisisa
continuing challege for researdners of enctienal egression and health
psychology. The construct of annbivalence over egression and one
measureof it, them, arepresentedasafirststepinneetingthat
challenge.

IIS‘I'OFREFEMNCES

96
LISI‘OFREFERENCIS

Allport, F. H. (1924). gig], m. Boston: Houghton-Mifflin.
Anderson, C. D. (1981). Engression of affect and physiological

response In psydnosanatic patierrts- genre]. 2: Wis
Research _2_5, 143-149.

Balswick. (1988). Essie-29mm. Iaxingtm, MA:
IexingtonBooks.

Banmeister, R. F. & Cooper, J. (1981). Can the public egectation of
enotion cause that enctian? m 9f mm, 52, 49—59.

 

Bauneister, R. F. & Tice, D. M. (1987). Ehetian and self
presentation. mile in my. 2. 181-199-

Bernardez-Bonnesatti, T. (1978). Wanen and anger: Conflicts with
aggression in contenporary wanen. M, 3;, 215-219.

Beutler, L. E., Engle, D., Oro-Bertler, M. E., Daldrup, R., Meredith,
K. (1986) . Innability to egress intese affect: A ccmncnn linnk
between depression and pain. m; 9f W _a_n__nd Clinical
mm, .53., 752-759-

Bell, P. A. (1978). Repression-sesitization. In H. London and J. E.

Ether (Eds), Dev-seize 2f mm (139- 449-485). mm
Wiley.

Bradburn. M- N. (1969). 11s ermine 2f Medal hell-M-
Chicago: Aldine.

Buck, R. W. (1984). file mm 2f m. N.Y.: (Milford.

Chodorow, N. (1974). Family structure and feminine personality. In M.

Rosaldo and L. Ianpere (Eds.) m, N, am society.
Stanford, an: Stanford Press

Cole, P. M. (1985) . Display rules and the socialization of affective

displays. In G. Zivin (Ed.) 1m M 9f expressive
behavior. (pp. 185-220) New York: Acadenic Press.

Cox, T'. & McCay, C. (1982). Psychosocial factors and
psychophysiological mechaniens in the aetiology and developnent
of Gamers. $29.13;; £91m end raisins. lé. 381-396-

Darwin. C- R- (1872). mmemmmmm
animals. London: John unrray.

Derogatis, L. R., Abeloff, M. D., Melisaratos, N. (1979).
Psychological coping mechanisns and survival time in metastatic
breast cancer. meremicanmmeie. gig.
1504-1508.

97

Derogatis, L. R., Lipaan, R. 8., Rickels, J., Innleinuth, E. H., &
Covi, L. (1974). The Hopkins Synptonn checklist: A measure of
primary sympton dimensio's. In Pichot (Ed. ) W
rrssigrsrrrssinnisaszrrzrumrrrussun.ltrksrlinsrflt__.inn
W. Basel, Switzerland. 8. Krager. 77-110

Derogatis, L. R. 8 Spencer, P. M. (1982). The brief syuptonn inventory
(BSI). Baltimore: Clinical Psyohonetric Research.

Diener. E- (1984)- SUbiective‘wellébeing-.Esrshglegieal.flulletinn 2§.
542-575.

Dieer, E. & Emma's R. A. (1985) . The independence of postive and

negative affect. grazing.gutEsrsgnalityuand.§ggial.Esyshglegy.
51, 1105-1117.

Dieer, E., Bones, R. A., Larsen, R. J., & Griffin, G. (1985).
Satisfaction with life. M 9; mil-12 Assessment, fl,
71-75.

Hanan, P. & Friesen, W. V. (1971). Constants across olltures in the

face and emotionm agurnal.2f.£er§9nality and fiesial Rsrshelegy.
11, 124-129.

Elmons, R. A. (1985). Personal strivings: An amroach to personality
and subjective well-being. gm 9f Pesonalig _o_ard_ Social
W: 5_1. 1058-1068.

Ennmons, R. A. & King, L. A. (1988). Personal striving conflict:
Immediate and long-term inplicetions for psychological and
physical wellébeingz degree;.Qf.Eersgnality'and.§g§ial
w-

Eyseck, H. J. & Eyseck, S. B. G. (1964). mgmm
9mm inventog. San Diego: Educational Testing Service.

Field, T. (1982) . Individual differeces in the egression of
neonates andyourg infannts. InR. S. Feldnan (Ed.) W
91; nonvgml M in m. (pp. 279-298) New York:
Springer-Verlag.

Freud, 8- (1917/1977). WM 91 W. (J-
Strachey, Trans.) NewYork, NY: W. W. Norton & Co.

Fridlund, A. J., Newman, J. B., 8: Gibson, E. L. (1984). Putting
emotion in behavioral medicine: Discrete-enotion
psychophysiology and its relevance for research and therapy. In
C. Van Dyke, L. Tenoshok, & L. S. Zegans (Eds.) mg
Health fl I_lln_n_e_§_s (pp. 117-134). N. Y.: Grune & Stratton, Inc

Friedman, H. S. & Booth-Kewley, S. (1987). The "disease-prone
personality:" A meta-analytic vial of the costruct. American
.2§yshelggisr. $2. 539-555.

98

Friedman, H. 3., Hall, J. A., & Harris, 1!. J. (1984). Nonverbal
expression of auction: Healthy diarisma or coronary-prone
behavior? In C. Van Dyke, L. W, & L. S. Zegans (Eds.)

Wigwam (pp- 151-165)- N- Y-= Grune &
Stratton, Inc.

Friedman, H. 5., Prince, L. 11., Riggio, R. E., & DeMattea, M. R.
(1980) . Understanng and assessing nmverbal communication of
emotion: The Affective Oammicatim Test. m _qf

mm am Essie]. mm. 12. 333-351.

Greer, S. & Morris, T. (1975). Psychological attributes of wanen who
develop breast cancer: A controlled sudy. m 91‘
mm: m. 12. 147- -

Hinds, R. A. (1985). Expression and negotiatim. In G. Zivin (Ed.)

E W of mine m. (m- 105-116) New York:
Academic Press.

Hochschild, A. R. (1975). The sociology of feeling and auction. In
M. Millman & R. Kanrter (Eds.) m m (pp. 280—307). Garden
City, N. Y.: Andwr Books.

Hochschild, A. R. (1983). :13 m 1333;. Berkeley: University of
Chlifornia Press.

Hokanson, J. E. & Burgess, M. (1962). The effects of 3 types of
aggressionmvascmlarprooesses. mgmmlagdwcial
m: .650 446‘4490

Hokanson, J. E. & Shetler, S. (1961). The effect of overt aggression
on mysiological arousal level- Jsanrnal of am and 5292's;
W. Q: 446-448.

Hunter, J. (1984). Factor analysis. In P. Merge & J. Cappella,

(Eds.) . W W in m Wig m
(pp.229-257). New York: Academic Press.

Hunter. 51- (1988). PM: Am forms mm.
Unpublished manuscript, Midnigan State University.

James, W. (1884). What is an engtim? Ming, 2, 188-205.
Jones, H. E. (1935) . The galvanic skin response as related to overt

emotional expmssim- m learns]. of mm. 91.
241-251

Keller. E. F- (1985)- W on we: and m. New Haven:
Yale University Press.

99

Kluckhohn, C. & mrray, H. A. (1953). Outline of a oonoqntion of

personality. InH. A. unrrayandc. nudmdm, Wig
_Natlne. sosim. and gem. 219 m. (139- 3-32)- New York:
Alfred anaf.

Laird. J. D. (1974). Self-attributim of auction: The effects of
expressive behavior on the cpality of anotional experience.
Jesse] of Permit! and seen]. Missy. 22. 475-486.

Lanzetta, J. T., Cartwright-Simth, J., & Kleck, R. E. (1976). Effects
of nonverbal dissimlatim on eugtiaial experience and autonanic

m-mmmmmm. 12.

Larsen, R. J. & Diener, E. (1987). Affect intensity as an individual
difference characteristic: A reviav. m of m _i_n

Personality, 3;, 1-39.

Lewis, M. Michalson, L. (1985). Faces as signs and symbols. In G.

Zivin (BIL) Tee designers of mire mien. (pp-
155-182) New York: Acadanic Press.

Wane D P (1985) m intimnendthelifessg=
Wmmm Wood IL=Dorsey

Mocanne, T. R. 8 Anderson, J. H. (1987). Engticnal responding
following experimental manipulatim of facial electranyographic
activity- gm of my and insist; Perm. .52..
759-768

Malatesta, C. Z. (1985) . Developnental course of auction expression
inthehumaninfant. InG. Zivin(Ed.):nggg_vel_;_m1th
expressive M. (pp.184-220) New York: Acadanic Press.

Malatesta, C. 2. & Haviland, J. M. (1982). Learning display rules.
_Glild MI .13! 991-10030

Marlowe. D. & Creme. D- D- (1964)- mmmms
evaluative W; N.Y.: Wiley.

Neehl, P- E- (1964)- Maire]. forgmi Mgwfiz io
siJn_s. Minneapolis, W: University of Minnesota Medical School,
Psychiatric Research Unit.

Monntgcmery, W. (1985). Charles Darwin' 3 ttnght an expressive
mechanisms in evolution. In G. Zivin (Ed.) fig m 915

mile ML:- (m. 29-48) Nev York: Acadanic Press.

100

Morris, T., Greer, 8., Pettingale, K. W., 5. Watson, M. (1981).
Patterns of expression of anger and their psychological
correlates invnnenwithbreastwger. 51%,],ng
w. .25. 111-117.

Murray, E. J. (1985). Coping and anger. In T. Fields, P. MoCabe, and

N. Sdnneiderman (Eds.) m and M (pp. 243-261).
Hillsdale: Erlbaum Associates.

Notarius, C. I. & Levenson, R. W. (1979). Enpressive tendencies and
physiological response to stress. lease]. of Lemuel and
.Sgiel mm. 31. 1204-1210

Pelletier, K. R. (1985). mm as m, m as slayer. N.Y.:
Delaoorte Press.

Pemebaker, J. W. (1985). Traumatic experience and psychosonatic
disease: anloring the roles of behavioral inhibition,
obsession, and oonfiding. m m, g_6_, 82-95.

Pennebaker. J- w. (1982)- The measles! of. Miss]. swans; N.Y.:
Springer-Verlag.

Raulin, M. L. (1984). Development of a scale to measure intense

ambivalence WW :2
53-72. 987

Roger, D. & Nesshoever, W. (1987). The oostruction and preliminary
validation of a scale for measuring onctional control.
Personalig and Individual mm, 3, 527-534.

Rogers, C. (1961). EEOC—(11.93:! m. Boston: Hoghton-Mifflin.

Rosenberg, M. (1963) . The association between self-esteem and
anxiety- ifs-1.1m]. of We m. l. 135-152-

Rosenstock, I. M. & Kirsdnt, J. P. (1979). Mny people seek
healthcere. In G. C. Stone, N. F. Cohen, 8 N. E. Adler (Eds.),

_Health mag—a. ham. (pp- 161-188) - San Francisco,
CA: Jossey-Bass.

Sackheim, H .A. (1983). Self-deception, self-esteem and depression:
The adaptive value of lying to oneself. In J. Masling (Ed.)

Luisal flee of We neonate. V-1 (pp- 101-157) .
Hillsdale, NJ: Erlbaunn.

Sartre, J. P. (1948). The m. New York: The Wisdon Library.

Shennum, W. A. & agental D. B. (1982). The developnont of control
over affective expression in nonverbal behavior. In R. S.

Feldman (Ed-L W of maze]. Eerie: in Loren.
(pp. 100-121) New York: Springer-Verlag.

101

Shields, S. A. (1987). Wonen, men, and the dilouna of emotion. In P.
Shaver & C- Hendrick (Ede) mswmm
_P_sydn_gl_ggy, 1, (pp. 229-250) Beverly Hills, (.1: Sage
Publicatios.

SnowH. (1883). mamm. Iodon:J&Acnmrdnill.

Tavris, C. (1984). m thewisdon of counting to ten: Personal and
social dangers of anger expression. In P. Shaver (Ed.) Review 9f
mm and $95411 Melee» (pp- 170-191) - Beverly H1115:
Sage.

Tarlau, M. & Smalneiser, I. (1951). Pmsonality patterns of patients
withmalignanttrmnoursoftnebreastandcervix. Mic
Medicine, 1;, 117.

Udelman, H. D. & Udelman, D. L. (1981). Enctios and rheumatologic
disorders. germ Jena]. of W. 2.5. 576-587-

Warr, P., Barter, J. & Brombridge, G. (1983). On the independence of
positive and negative affect. Ml 91 W E Social
ngcholm, 55, 644-651.

APPE‘IDICFS

AppendixA
Em
Usingthe7-poinnt scale below, indicatetheextonttowhichyon

believe tne statement characterizes you, by placing the appropriate
number on the line preceding the itan.

1 2 3 4 5 6 7
not at slightly sonewhat moderately very extranely
all characteristic
characteristic

8. I often tell people that I love them.

45. IstmthatIliJesoneonebymggingortonchingthat
person.

36. I often toxzh friends during conversations.

23. Watching television or reading a book can make me laugh out
loud.

10. I laugh a lot.
1. WhenIamangrypeoplearondneusuallyknm.
12. People can tell fron my facial expressions how I am feeling.

30. Wheneverpeopledonicethinm forne, I feel "pntonthe
spot" and have trouble expressing my gratitude. (-)

18. WtenIreallylikesoneonetheyknmit.
25. Ioftenlaughsohardthatmyeyeswaterormysidesade.

22. WhenIamaloe, Icennekemyselflaughbyremanbering
sonethingfronthepast.

34. My laugh is soft and subdued. (-)

29. If a friend surprised me with a gift, I Wildn’t know how
to react.

14. IapologizewhenIhavedonesonethingwrong.

26. If soneone makes ne angry in a public place, I will "cause a
scene."

7. Ialwaysexpr‘essdisarpoinrtmentwtenthingsdon'tgoasl’d
likethemto.

102

AmendixB
Am

Belowandontlenextfewpagesares‘tatonentsthatreferto
typically oconrring emotional reactios. We wannt you to decide how
frequenntlyyonhavefeltliketl'ethonghtsbeingexpressedinthe
statenent, andttenindicatethisnsingtheSpointsoalebelov.

1 ............... 2 ............... 3 ............... 4 ..............
5
I have never I frequently
felt like this feel like this

The statanent may cosist of two thonghts. Carefully read the
statenentasawholebeforedecidingonhowdnaracteristic it isof
yon. For exannple, cosider tne iten:

"I trytohonestly criticize others for theirowngood, but I worry
thattheymaygetangrywithneifldoso"

Yonmnldgivethisitemahighrating ifandonly ifyonbothtryto
honestly criticize others am worry abort their getting angry. If
yon simply try to criticize others and yon don’t worry abort their
anger, or if yon don’t try to criticize others at all, then yon would
rate this item lower. It is inportannt to cosider the colplete
thonghts being expressed before yon respod.

1. IwanttoexpressmyanotioshonestlyhntIamafraidthat
itmayoauseneonbarrassnentorhurt.

_2. Itnytocontrolmyjealonsyconcerningmyboyfriendor
girlfriendeventhonghIwantto letthemknowI’m lmrting.

3. Imakeanefforttocontrolmytenperatalltimeseven
thonghI’d liketoactonthesefeelings at times.

4. Itrytoavoidsulkingevenmenlfeellikeit.

5. menIamreallyprodofsone‘thingIaccotplishedeanrtto
tellsoneone,butIfearIwillbetlnnghtofasconceited.

6. I would liketoexpressmyaffection more physicallybut I
amafraidthatotnerswillgettlewroginpression.

7. Itrynottomrryotlerseventlnghsonetimestheyshonld
logwtletruth.

8. OftenI'dliketoshovotnershowaeel,b\ntsonethingseems
tobeholdingneback.

9. Istrivetokeepasnnileonmyfaceinordertoconvince
others I’m happierthanI really am.

103

1....0.0.000...020.000.000.000003 OOOOOOOOOOOOOOO 4.0.0.0... ...... 5
Ihavenever Ifrequently
felt likethis feel likethis

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Itrytokeepmydeepestfearsandfeelingshidden,butat
timesI'dliketoopenuptoothers.

I’d liketotalkabontmyprcblemswithothers, butat times
Ijust can’t.

Whensoneonebothersne, Itrytoarpearindifferenrteven
thonghI’dliketotellthanhowaeel.

Itrytorefrainfrongettingangryatmyparenntseventhongh
Iwanttoattimes.

ItrytoshowpeoplethatIlovethon, althoughattintes
I am afraid that it mayneke ne appearweak or too sensitive.

ItrytoapologizewhenIhavedonesonethingwrong,bntI
worrythathillbeper'ceivedasinnconpetennt.

OftenIfindthatIamnotabletotellotlershowmnch
theyreallyneantone.

IwannttotellsoneonevtenIlovethem,butitisdiffionlt
tofindtherightwords.

Iwonldliketoexpressmydisamointmentwnenthingsdon't
goaswellasplannned,!mtIdon'twannttoanpearvulnerable.

IcanrecellatinewheanishthatIhadtoldsoneonehow
mnanreallycaredabontthenn.

Itrytohidemynegativefeelingsaronndotners,eventhongh
Iamnotbeingfairtothoseclosetome.

Iwonldliketobemorespontaneosinmyonotionalreactions
buthustcan'tseenntodoit.

Itrytosupressmyanger,hnthonldlikeotherpeopleto
knowhowaeel.

Itishardtofindtlerightwordstoindicetetootherswhat
Iamreallyfeeling.

Imrrythatiernanessnegativeowtionssmflnasfearand
anger,otherpeoplewillnotanproveofme.

Ifeelguiltyaftethaveexpressedangertosoneone.
IthinkabontactingwhenIamangrybutItrynotto.

104

1.0.0.0000000000200000 ...... 0.0.3 ...... O ........ 4.0.0.00000000005

I have never I frequently
felt like this feel like this

_27. TtemnevebeentimswlnonItoldsoneoeIlovedthen

onlytolaterregrethavingdoneso.

_28. AfterIexpressangeratsoneoe, itbothersneforalorg

time.

105