AGGRESSIVEBEHAVIOR
Volume20, pages17-26 (1994\
Intakeand Aggressiveness:
Testosterone
RealEffector Anticipation?
Kaj Bj6rkqvist,Tor Nygren,Ann-ChristinBi6rklund,and
Stig-EyrikBj6rkqvist
Depaftmentof SociatSciences, AboAkademi university,Vasa,Fintand
(40
In a double-blindexperiment,human males(n = 27\ weregiveneither testosterone
mg/day), placebo,or no treatment, over a one week period. Subjectiveand observer
mood estimationswere conductedbeforeand after treatment. Testosteronelevels
assessed
in saliva were measuredwith radioimmunoassay.The results revealeda signilicant placebo effect [c.f. Medicine and Sciencein Sports 42124-126]:After treatment, the placebo group scored higher than both the testosteroneand the control group on self-estimated anger, irritation, impulsivity, and frustration. Observer-estimatedmood yielded
similar results. The lack ofa placebo effect in the testosteronegroup is intriguing, and
may be due to secondary effectscausedby suppressionof the body's own testosterone
production, since recorded non-protein bound testosteronedid not significantly rise due
to treatment. The results suggestthat androgen usagecausesexpectations,rather than
Inc.
an actual increaseofaggressiveness. tj 1994wiley-Liss,
Key words: testosterone,aggression,radioimmunoassay,mood, steroids
INTRODUCTION
by sportsmen
abusing
anabolic
of homicidecommitted
steroids
Reports
are,every
now and then, presentedin the newsmedia. Such casesare typically describedas having
no apparentmotive and as accompaniedby uncontrolled fits of rage. In 1984, a Finnish
athletewho killed his girlfriend blamed the usageof steroidsfor his irrational behavior.
That same year,a steroid-usingSwedishathletekilled his best friend, and on December
6, 1992, a TV news report on TV announcedthat a Swedish steroid abuser had shot
and injured six people. These are but examples,and similar instancescan easily be
found in countriesoutsideScandinavia.
June9. 1993.
for publicationMay 9, 1993:accepted
Received
of SocialSciences,Abo ekademiUnivenity.Virrigatan
to Kaj Bjirrkqvist.Department
Addressreprintrequests
9. SF-65100Vasa.Finland.
@ 1994Wiley-Liss' Inc.
18
Bjdrkqvisret al.
The mediapresentthe connectionbetweenthe useof steroidsandaggressiveness
as
if it werea well established
fact.The pressandTV, naturally,do not reportabusersnot
killing or not becomingaggressive.
Likewise,whennon-abusers
commithomicideor
someotherviolentcrime,theprefix "non-abuser"is not containedin thereport.Negativecasesand falsepositivesneverpassthe newsthresholdand thus,everyreported
incidentearnsdisproportionate
coverage.
It hasneverbeenshownthatathletesabusing
steroidscommithomicidesignificantlymoreoftenthannonabusing
athletes,andeven
if this weretrue, it wouldnot be proofof a cause-effect
relationship.
overthe useof anabolicsteroidshasunquestionably
Controversy
createda greatdilemmain the world of sports.Alreadytwo decadesago,Coopertl972l reportedthat
80Vaof weightlifters,javelin throwersand shot-putters
regularlyusedillegalsteroids.
Recentcases,suchas Ben JohnsonandKristinaKrabbe,indicatethat steroidusageis
now commonamongshort distancerunners,too. Whethersteroidintakereally improvesperformance
hasnot beenscientificallyestablished,
althoughsportsmenseeking fastrecoveryafterintensivetrainingandrapidmuscledevelopment
seemconvinced
that it does[e.g., Lucking, 1982;Perryet al., 1990].In a double-blindexperiment,
Freedet al. [975] found a slight improvementin the performance
of weight lifters.
The only changein physicalperformance
Hannanet al. [991] couldestablishwasan
improvementin the first trial of a pegboardtask.Ariel and Saville11972)foundplaceboto be moreeffectivethansteroids.
The relationshipbetweentestosterone
levelsand aggressionin subhumanspecies
hasbeenstudiedextensively,and findingssuggesta connectionbetweenthe two [for
reviewssee,for instance,Bouissou,1983,and Brain, 19771.Castrationis known to
reducephysicalaggressionamonganimals,includingsubhumanprimates[Bernstein
et al., 1983].Thereare,however,goodreasonsto believethatlevelsoftestosterone
are
primarilyrelatedto socialstatusand dominance.Rejeskiet al. [988] gaveinjections
of testosterone
to an experimentgroupof malecynomolgusmonkeys,while a control
groupreceivedshaminjections.Althoughtheadministration
of testosterone
resultedin
a significantincreasein aggression,more importantwas the finding that changesin
behaviorare mediatedby socialstatus;the incidenceof both contactand noncontact
aggression
in dominantmonkeyswasfar greaterthanthe frequencyof thesebehaviors
in subordinate
monkeys[Rejeskiet al., I 988].
As far ashumansareconcerned,the questionof whetherthe levelof testosterone
in
malesis relatedto aggressionis moreopento debate.While somestudiesindicatea
relationship,othersdo not. For example,Olweuset al. [980] founda correlationbetweenunboundplasmatestosterone
andcertainaspectsof aggressiveness
andimpulsivity.
Ehrenkranzet al. [974] usingprisonersas subjects,foundplasmatestosterone
to be
correlatednot only with aggression,
but with socialdominance,too. Radaet al. |9741
investigated
plasmatestosterone
levelsin rapistsbut werecautiousaboutinterpreting
theirresultsasindicatinga causalrelationship
with rape.Lindmanet al. |9921did not
find a correlationbetweenserumtestosterone
levelandaggression
in malesarrested
for
spouseabuse,neitherat thetime of arrestnor in a sobercontrolstate.
Theseexamplesshowthat somestudiesdo indicatea relationship
betweentestosterone and aggressiveness,
othershavefailedto find a correlation.It is well knownthat
positiveresultsaremorereadilypublishedthannull findings,andmanyresearchers
do
not evenbotherto submit manuscriptsreportingnonconfirmedhypotheses.
Accordingly,manystudiesinvestigating
thetestosterone-aggression
yieldinga negarelationship
tive resultarelikely neverto havebeensubmittedand/orpublished.
Intal<e
Testosterone
19
In view of the inconsistentfindings of researchon this topic, the conclusionsdrawn
by reviewersalso tend to be in disagreement.While some claim that a connection can
be established[e.g., Donovan, 1985], others [such as Benton, 1992]are of the opinion
that, on the basis ofexisting data, there is no reasonto suggestthat human aggression
is related to level of testosterone.Benton [983a, 1983b] emphasizedthat the claim of
a relationship between testosteroneand aggressionis based primarily on animal data.
Extrapolationsfrom animal to human behaviorare questionable,since, among humans,
social and cognitive mechanismsplay a much greater role than physiological factors.
The closer animal is to man on the phylogenetic scale, the smaller the influence of
testosteroneon aggression.Benton's |9921conclusion is that in man, aggressivebehavior is a reflection of psychosocialhistory, and differencesin aggressivenesscan be
attributedto the level of testosteroneonly to a very limited extent, if at all.
If Benton is correct-and the presentauthors think he is-then the main source of
variation in aggressiveness
among humans is really to be found in the psychohistoryof
the individual. What, then, is the explanationfor the immenseinterestin the testosteroneaggressionconnection? Is it primarily a reflection of hardy attitudes about the aggressive male and the submissivefemale, still prevalentwithin our society?
If males within subhumanvertebratespeciesare more aggressivethan females(which
is not true for all species) [Adams, 1992], among humans, this seemsto be reflected
only in greaterphysical aggression.In regard to direct verbal aggression,human females are as aggressiveas males, and with respectto indirect aggression,evenmore so
[Bjokqvist etal.,1992a; Bjorkqvist etal.,1992b; Lagerspetzetal., 1988].The greater
amount of physicalaggressionamong males is more likely to be a consequenceof learning mechanismsand greater physical strength, rather than of testosterone[Bjorkqvist
and Niemelii, 19921.
Most researchon aggressionis conductedby men, with male subjects,perhapsre"male" perspective
on the subject. For instance,Frodi et al. |977)
sulting in a biased,
found that of 314 experimentalstudieson aggression,54Vowere concernedexclusively
with men, and only 87ocould be found specifically investigatingaggressionin women.
Another revealing example is the different interest arousedby two experimentsconducted by Edwards [Edwards, 1969, and Edwards and Herndon, 19701.In Edwards
[969] , it was reported that female mice treated with androgens at birth were more
aggressiveas adults. This result is well-known and well-publicized. The secondstudy
[Edwards and Herndon, 1970] showedthat newbom female mice treated with estrogen
at birth also fought more as adults! This finding has gone almost unnoticed, since it is
not in accord with the generally accepted.
The aim of the present study was to investigate whether there is any truth in the
reports of increasedinitability and aggressivenessin males after steroid intake [Perry
etal., l990l.Thestudywasadouble-blindexperiment.Thesubjects,universitysportsmen who volunteeredfor the experiment, were divided into three groups: testosterone,
placebo, and no treatment. Testosteronewas administeredperorally for one week. The
dose (40 mg/day) was well within the levels administeredfor therapeuticreasonsand,
taken for such a limited period, the treatment could not possibly cause the subjects
physical harm. The subjects were, however, under strict medical supervision for the
duration of the experiment.
There are findings suggestingthat it is the level of non-protein-boundtestosterone
which is decisive in behavioral manifestations[cf. Vermeulen and Verdonck, 19'12].
Unbound testosteroneconstitutes l-4%oof the total level [Sannikka et al. , 1983]. In the
20
Bjdrkqvisteral.
presentresearch,unboundtestosteronelevelswere measuredwith radioimmunoassay
of the saliva,beforeand after the treatmentperiod [Furugyamaet al., 1970;deGomez
et al., l98l; Ismailet al., 19721.This methodoffersa morebenignwayof measuring
in salivacorrelatehighly with the
than blood sampling.Measurements
testosterone
level: .75 accordingto Sannikkaet al. [983],
serumtestosterone
non-protein-bound
to Wanget al. [1981].
and.94according
In the presentexperiment,pre- and-posttreatmentbehavior,mood, and physicalfitandobserver-evaluated
includingself-estimated
by multiplemeasures,
nesswereassessed
behavior,and interviews.Each subjectwas observedin a
mood, observer-estimated
werevideotaped.In this papel
groupsituationwith two other subjects.All sessions
only theeffectof treatmenton moodwill be reportedin detail.
MATERIALSAND METHODS
Subjects
universitystudents,all males,andall but oneactivein sports,particiTwenty-seven
patedin the experiment.Most of them weremembersof the sportsassociationof Abo
for the experiment,and reAkademiUniversity,Turku, Finland.All had volunteered
of 100,-FIM for their participation.Their meanagewas23.9
ceiveda remuneration
years(agerange2 I -3 I years). The meanweight of the subjectswas 73.6 kgs (range
62-88 kgs), and the meanheightwas 180.3cm (range170-197cm). None of them
or other steroids.Subjects
admittedto havinghad pesonalexperienceof androgens
groups:
placebo,
and control, on the
testosterone,
were divided into three treatment
in the saliva.The main criterionfor selection
basisof matchedlevelsof testosterone
wasthat the testosteronelevel of the threegroupsshouldcorrespondascloselyas posthat therewereno
sible.The physicianconductinga medicalexaminationto ascertain
for the selection
participation
experiment
was
responsible
in the
medicalobjectionsfor
interview
andperwere
also
checked
by
counterindications
of subjects.Psychological
possible
precaution
were
rejected.
Every
volunteers,
three
Of 30
sonalityassessment.
wastaken,bothprior to andduringtheexperiment,to ensurethatsubjectswouldcome
wereignorantas to which groupeachsubject
to no harm. The actualexperimenters
designof thisexperiment.
with
the
double-blind
accordance
belonged,in
Sessions
The subjectsattendedthreesessionsin the laboratory:Thesewill be referredto as
I and II wereseparatedby2 to 3 weeks,
Sessions
I, II, and II[, respectively.
sessions
periodbetweensessions
II andIII wasl0 days.Testosterone/placebo
andthe intermediate
data,
intakebeganexactly7 daysbeforesessionIII. SessionII deliveredpre-treatment
datawereobtainedfrom sessionIII.
andpost-treatment
Testosteroneand Placebo
producedby
consistedof 40 mg pills of Panteston@,
testosterone
The administered
perorally,i.e., it hasto be kept underthe tongue
is administered
Oregon.Panteston
until dissolved,it is not to be swallowed.Administeredperorally,it worksthroughthe
lymphaticcirculationand is not brokendown by the liver.The placebopills werecalin thesameway.
pills which wereadministered
cium carbonate
TestosteroneIntakc
2I
Radioimmunoassay (RlA) of the Saliva
Subjectsdelivered 5 ml saliva into a test tube after first rinsing their mouth with pure
water and then chewing on a parafilm in order to stimulate the secretionof saliva. The
delivery alwaystook place at the sametime ( 19.00) in order to avoid differencesdue to
circadian rhythm. The samples were immediately frozen and stored until the time of
actualanalysis.Analysiswasconductedaccordingto the methoddescribedby Sannikka
etal. [ 19831.
The Self-Estimated Mood Check List
The mood check list utilized successivemagnitudescales[Lindman, 1985]. Subjects were requestedto indicate the degree to which they confirmed experiencing certain moods (e.g. relaxed)on a straight 100 mm line, with zero confirmation at one
end. and 1007oconfirmationat the other.
Twelve items (expressingmood) were included:Relaxed,irritated, excited, angry,
indiff'erent.
interested,frustrated,elated,nervous,impulsive,self-confident,andactive.
Observer-Evaluated Mood
All sessions,except for the medical examination, were videotapedfor the subsequent assessmentof subjects' mood during the sessionby two independentobservers.
The observerswere unawareof the group to which subjects belonged. A similar sucmood check list
cessivemagnitudescaleand the same l2 items as in the self-evaluated
wereused.The inter-raterreliability was .86 (P < .01).
Session I
The sole purposeof the first sessionwas to investigatethe psychologicaland medical suitability of those volunteeringfbr the experiment. SessionI consistedof four
parts: l) delivery of saliva fbr the RIA test, 2) a medicalexamination,3) an interview,
4) personalityassessment.
datafor alloThe purposeof the salivasamplewas to establishbaselinetestosterone
cation into groups.The purposeof the medicalexaminationwas to investigatemedical
were conductedin order
suitability.while the interviewand the personalityassessment
to test fbr psychologicalcounterindicationsto participation in the study. Three subjects
were excluded on basis of medical and psychologicalgrounds. The whole session,
exceptfbr the medicalexamination,was videotaped.
Session ll
The subjects were taken to the laboratory in groups of three. each from a different
no treatment)group.The sessionconsistedof two parts:
treatment(testosterone/placeboi
l) completion of the self-estimatedmood check list, to obtain pre-treatmentdata on
mood, 2) a physicalperfbrmancetest. The whole sessonwas videotaped
self-estimated
for subsequentobserverassessmentof subjects' mood. Physical performancewas evaluated during sessionII, fbr comparisonwith post-treatmentperformanceduring session il1. by exercisesdemandingphysicalstrengthas well as fine motor skills. Subjects
3 daysafter sessionII, and 7 daysbeforesessionIII.
startedtaking testosteroneiplacebo
Session lll
SubjectsattendedsessionIll in a different group of three to that of sessionII. The
) d e l i v e r y o f s a l i v a t b r R l A ( a t e x a c t l yl 9 : 0 0 ) , 2 ) c o m p l e t i o n o tf h e
s e s s i o n i n c l u d el d
22
Bjdrkqvistet al.
self-estimatedmood check list (post-treatmentdata), 3) a physical performance test
(post-treatmentdata), 4) completion of the mood check list a secondtime (post-treatment,
after-exercisedata), 5) an interview. The whole sessionwas again videotaped, and the
mood of subjects subsequently evaluated by observers.
RESULTS
Effects on Self-Evaluated Mood
The results were somewhat surprising: A significant placebo effect was found on
four (one third) of the twelve items of the self-evaluatedmood check list, but no similar effect was obtained for the testosteronegroup! Subjectsbelonging to the placebo
group experiencedthemselves as more angry [F(2,23):
6.01, p < .01], frustrated
p
<
=
5.85,
P
.011,
impulsive
IF(2,23):
[F(2,23) 4.30, < .02],andirritated[F(2,23)
: 6.99, P < .011. Pre and post data of two of these items (angry
and irritated) are
presentedas examplesin Figures I and2.
Effectson Observer-Eval
uatedMood
Observer-evaluations were consistent with the self-estimated data, thus supporting
the validity of the findings. For example, observer-evaluatedfrustration revealed a sis-
ANGRY
100
O = testosterone
Q = ptacebo
Q = control
* =p-.O5
**=p<,O1
BEFORE
BEFORE & AFTER
TREATMENT PHYSICAL EXERCTSE
A F T ER
TREATMENT
Fig. I .
Self-evaluatedangerbeforeand after treatmentin the threeexperimentalgroups
TestosteroneIntalcc
23
IRRITATED
100
O = testosterone
Q = placebo
O = control
r =p<.O5
*r= p< .O1
BEFORE
T R E A T ME N T
BEFORE & AFTER
PHYSICAL EXERCISE
r*lIIfil"'
Fig. 2.
Self-evaluatedirritation before and after treatment in the three experimental groups.
nificant placebo effectlF(2,23) : 7.05, P < .011. The remaining items also revealed
group differences consistent with self-estimatedmood, but these were, however, not
significant.
When observer-evaluatedanger, irritation, and frustration were added together to create a summed variable (the internal consistencyof which was ct : .90), the summed
variablerevealeda significant placebo effect[F(2,23) : 4.05, P < .05], too. Why were
group differences based on observer-evaluationscores smaller than group differences
basedon self-estimations?Subjectswere aware of the fact that they were being videotaped:they may have attemptedto hide their emotions from the camera. In other words,
they may havebeen camera-shy.
lnterviews
Data obtained in the interviews revealedthe same trends. For instance,one individual belonging to the placebo group expressedhimself with the following typical words:
"l
have, during this week, felt touchy and irascible, I got irritated almost for no reason.
I was very tired at the beginning of the week. I got mad at others, and it showed. I'm
not usually like that. My physical training has been more successfulthan ever. I felt
self-confidentin competition, which is unusual for me. ' '
24
Bjdrkqvist et al.
RIA Data
level was found after I
No significantbetweengroupsdifferencesin testosterone
levels,but not more
increased
grouphad slightly
week'streatment.The testosterone
thanmightbe explainedby chancefluctuation.
Physical Performance
in physicalperformance.
Therewereno betweengroupsdifferences
DISCUSSION
The resultsdemonstratea clear placebo effect: Subjects,although they denied this in
the interview, obviously expectedto become more irritable, frustratedand angry. What
is surprising and calls for an explanation, is the absenceof a placebo effect in the
group receiving testosterone.This finding is so consistentthat it is difficult to conceive
of it as a coincidence. Although the presentauthors do not claim to have an explanation for this finding, the following suggestionis offered: Contrary to expectation. testosterone levels did not differ significantly between the groups after I week's treatment.
What may have happened is that the artificially induced testosteronesuppressedthe
body's own production of the steroid: If the treatmenthad been carried out for a longer
period, or larger doseshad been administered,differencesin testosteronewould probably have appeared.The suppressionof the body's own production may have had secondary psychological effects, depressing excitement and activation which, in turn,
balancedout the activating placebo effect.
It is worth noting that our resultsare consistentwith thoseof Ariel and Saville ll972l,
who also found a significant placebo effect, but no effect of steroid intake. However,
unlike the present study, they examined the effect of steroidsonly on physical performance. and not on mood.
One has to remember that steroid dosage taken by athletes is considerably larger
than that administeredin the presentexperiment, or that which could be prescribedfor
therapeutic or experimental reasons.Athletes also continue steroid treatment for long
periods and without medical supervision. The question whether steroid abuse causes
uncontrollablemood changeand aggressionis still not settled:Hormonal variation may,
or may not, be related to significant mood change. A parallel may be drawn to the
debateabout the premenstrualsyndrome [Parlee, 1973], or mood changedue to steroid
based contraceptives.It is possible that sensitivity to intake of artificial steroids leading to unpleasantmood changesis subject to individual variation.
Two recent studies investigating the relationship between anabolic-androgenicsteroid use and mood change [Anderson et al., 1992; Bahrke et aI., 1992] yielded the
in
same result as the presentexperiment: Steroid intake did not increaseaggressiveness
human males. A third double-blind study [Hannan et al., l99l] did find increased
"resentment
"hostility"
and
evenmore so in high-dosegroups,with the
aggressiveness,
and aggression" subscalesof the MMPI as dependentmeasures.Two field studies [Brower
et al., l99l; Perry et al., 19901found that sportsmenabusing steroidsreported increasedaggressiveness.It should be borne in mind, however,that the last two studies
were not double-blind experiments, and the reports may be explained as placebo
effects. A review by Uzych ll992l does not exclude the possibility of increased
aggressivenessafter steroid abuse, and another by Bahrke et al. [1990] suggeststhat
Intakc
Testosterone
25
psychological effects are variable, and related to type (17 alpha-alkalatedratherthan
l7 beta-esterified), but not dose, of anabolic-androgenicsteroids administered. The
resultsof the presentand other double-blind experimentssuggesthoweverthat much of
the effect of steroid intake is really placebo. To claim that aggressionis causedby steroids, as in the news media, is misleading and dangerousfor severalreasons: l) There
rs still no scientific evidencefor such a relationship. 2) Even if a correlation were found
this would not prove a cause-effectrelationbetweensteroid abuseand aggressiveness,
ship. Factors, such as unstablepersonality may be the source of willingness to abuse
steroidsas well as aggressiveness.Steroid abuse may, for some, be a mediating factor
enhancingaggressivetendenciesby producing statesof elated emotionality. A parallel
may be drawn to the complexity of problems encounteredin establishinga causal link
betweenalcohol intoxicationand aggression[Brain, 1986; Evans, 1986]. 3) Dissemiconnection may lead to anticipation (a
nation of the myth of the steroid-aggressiveness
placebo effect) of aggressivenessamong steroid abusersand, in turn, to actual acts of
violence. [t may, in fact, work as an excusefor aggression.
Abuseof anabolicsteroidsmay be addictive[Broweret al., l99l]. Abuse is dangerous for medical reasons,and increasedvulnerability forcardiovasculardiseasehas been
"increased aggressiveness"is
observed [Cheever and House, 1992]. The problem of
really of minor importance. The disproportionatemedia coverageof the phenomenon
appearsto be a reflection of attitudes within our society, the cherished myth of the
aggressivemale, stuffed with androgens,and the submissivefemale.
REFERENCES
Adams D ( 1992): Biology does not make men more
aggressivethan women. In Bjorkqvist K, Niemelii
"Of
P (eds):
Mice and Women: Aspects of Female Aggression." San Diego. CA: Academic
P r e s s .p p l 7 - 2 5 .
Anderson RA. Bancroft J. Wu FC (1992): The ef'tbcts ofexogenous testosteroneon sexuality and
mood of normal men. Joumal of Clinical Endocrinological Metabolism 75: I 503- I 507.
A r i e l C , S a v i l l e W ( 1 9 7 2 ) : A n a b o l i c s t e r o i d s :T h e
physiological effects of placebos. Medicine and
Science in Soorts 4:124-126.
Bahrke MS. YesalisCE, Wright JE ( 1990): Psychological and behavioural eftbcts ofendogenous testosterone levels and anabolic-andrcgenic steroids
among males. A review. Sports Medicine l0:
303-337.
RH, CatlinDH
BahrkeMS, WrightJE, Strauss
(1992): Psychological moods and subjectively
perceived behavioral and somatic changesaccompanying anabolic-androgenicsteroid use. Ameri c a n J o u m a l o f S p o r t sM e d i c i n e 2 O : 71 7- 7 2 4 .
Benton D ( | 983a): The extrapolationsfrom animals
to man: The example of testosteroneand aggres"Multidiscision. In Brain PF. BentonD (eds):
pliniary Approaches to Aggression Research."
Amsterdam: ElsevierlNorth-Holland, pp 4O2-4 I 8.
BentonD ( 1983b):Do animalstell usanythingahut
andhuman
betweentertostemne
therelationship
'Animal ModIn DaveyGCL (ed):
aggression?
els of HumanBehavior."New York. NY: Wiley,pp 281-298.
andhumanaggression.
BentonD ( 1992):Hormones
"Of Mice and
In BjorkqvistK. NiemeliiP (eds):
Women:Aspectsof FemaleAggression."San
Diego,CA: AcademicPress.pp 37-48.
Bernsteinl, GordonTP.RoseRM ( 1983):The interactionof hormones.behaviorandsocialcontext
"Horin non-humanprimates.In SvareBB (ed):
monesand AggressiveBehavior." New York:
Plenum.pp 535-561.
BjorkqvistK, NiemeliiP (1992):New trendsin the
study of femaleaggression.ln Bjtirkqvist K,
NiemeliiP (eds):"Of Mice andWomen:Aspects
of FemaleAggressoin."San Diego,CA: AcademicPress,pp 3- 16.
A ( 1992a):
KMJ. Kaukiainen
BjbrkqvistK, Lagerspetz
Do girls manipulateandboysfight'l Aggressive
.
Behaviorl8:l17-12"1
A (1992b):
K, Kaukiainen
BjdrkqvistK, Osterman
of directand indirectaggresThe development
in malesandfemales.ln Bjdrkqvist
sivestrategies
K. NiemeliiP (eds):"Of Mice andWomen:Aspectsof FemaleAggression."SanDiego,CA:
AcademicPress.pp 5 I -64.
26
Bjdrkqvist et al.
behavBouissouMF ( 1983):Androgens,aggressive
ior and socialdominancein higher mammals.
| 8:43-61.
HormoneResearch
BrainPF (1977):"HormonesandAggression,Vol.
1." Monlreal:EdenPress.
Brain PF (1986):Alcohol and aggression-thenatureof a presumedrelationship.In Brain PF (ed):
'Alcohol and Aggression." London: Croom
Helm,pp 212-223.
BrowerKJ, Blow FCX, YoungYP, Hill EM (1991):
Symptomsandcorrelatesof anabolic-andmgenic
British Journalof Addiction
steroiddependence.
86:759-768.
imCheeverK, HouseMA (1992):Cardiovascular
plicationsof anabolicsteroidabuse.Joumalof
Nursing6: I 9-30.
Cardiovascular
CooperDL (1972):Drugs and the athlete.JAMA
2l:1007-l0ll.
DonovanBT (1985):"Hormonesand HumanBehavior." Cambridge,UK: CambridgeUniversity
Press.
EdwardsDA (1969):Early androgenstimulationand
aggressive
behaviorin male and femalemice.
PhysiologyandBehavior4:333-338.
EdwardsDA, Herndon,J (1970):Neonatalestrogen stimulationand aggressive
behaviorin feandBehavior5:993-995.
malemice.Physiology
Ehrenkranz
J. Bliss E, SheardMH (1974):Plasma
testosterone:
Correlationwith behaviorand soMedicine
cial dominancein man.Psychosomatic
36:467-468.
EvansCM ( 1986):Alcoholandviolence:Problems
relatingto methodology,statisticsandcausation.
In Brain PF (ed): 'Alcohol and Aggression."
London:CroomHelm,pp 138-160.
Freed DLJ, Banks AJ, Longson D, Burley DM
( 1975):Anabolicsteroidsin athletics:Crossover
double-blindtrial on weightliften. BritishMedical Journal2:47| 473 .
FrodiA, MacaulayJ, ThomePR(1977):Are women
thanmen?Psychological
alwayslessaggressive
Bulletin84:634-660.
FurugyamaS, MayersDM, NugentCA (1970):A
Stefor plasmatestosterone.
radioimmunoassay
r o i d sl 6 : 4 1 5 .
deGomezMS, Frei A, Vitins P, CekanS (1981):
and
Influenceof antisera,purificationprocedures
radioimmutracersof validity of a testosterone
noassay.
Annalsof Clinical BiochemistryI 8:28I .
HannanCJ Jr, Friedl KE, Zold A, Kettler TM,
PlymateSR (1991): Psychologicaland serum
homovanillicacidchangesin men administercd
l6:
androgenicsteroids.Psychoendocrinology
335-343.
IsmailAAA, NiswenderGD, MidgleyAR (1972):
Radioimmunoassay
withoutchroof testosterone
matography.Joumal of Clinical Endocrinology
34:177.
Lagerspetz
KMJ, BjorkqvistK, PeltonenT (1988):
Is indirect aggressiontypical of females?Gender differencesin aggressiveness
in I l- to 12year-old children. AggressiveBehavior l4:
403-414.
Lindman R (1985): On the direct estimationof
mood change. Perceptionand Psychophysics
37:170-174.
LindmanR, von derPahlenB, Ost B, ErikssonCJP
(1992):Serumtestosterone,
cortisol,glucose,and
ethanolin malesarrestedfor spouseabuse.AggressiveBehaviorI 8:393-400.
Lucking MT (1982):Steroidhormonesin sports.
Specialreference:Sex hormonesand their derivatives.InternationalJoumalof SponsMedicine3:65-67.
OlweusD, MattsonA, ScallningD, tiiii'w H (1971):
Aggression,physicalandpersonalitydimensions
in normaladolescent
males.Psychosomatic
Medicine 2:253-269.
ParleeMB (1973):The premenstrual
syndrome.PsychologicalBulletin 8O:454-465.
FerryPJ,AndenenKH, YatesWR (1990):Illicit anabolicsteroidusein athletes.A caseseriesanalysis. AmericanJournalof Sportsand Medicine
18:422-428.
RadaRT, LawsDR. KellnerR (1976):Plasmatestostercnelwels in the rapist. PsychosomaticMedicine 38:25'7-267.
RejeskiWJ, BrubakerPH, Herb RA, Kaplan JR,
Koritnik D ( 1988):Anabolicsteroidsandaggressive behaviorin cynomolgusmonkeys.Journal
of BehavioralMedicinel l:95-105.
SannikkaE, TerhoB SuominenJ, SanttiR (1983):
Testosterone
concentrations
in humansalivaand
its correlationwith non-proteinboundand total
testosterone
levelsin serum. IntemationalJournalof Andrology6:3l9-330.
UzychL (1992):Anabolic-androgenic
steroidsand
psychiatric-relatedeffects:A rcview. Canadian
Journafof Psychiatry37:23-28.
VermeulenA, VerdonckL (1972\: Somestudies
on the biological significanceof free testosterone.Joumalof SteroidBiochemistry3:421426.
WangC, PlymateS, NieschlagE, Paulsen
CA (1981):
Salivary testosteronein men: Further evidence
of a dircct correlationwith frce serumtestosterone.JournalofEndocrinoloevandMetabolism
53:1021.