Biological Psychology/Pharmacopsychology

Editor: P. Netter (Giessen)

Original Paper
Neuropsychobiology 2004;50:102–107
DOI: 10.1159/000077947

Hostility-Aggressiveness, Sensation
Seeking, and Sex Hormones in Men:
Re-Exploring Their Relationship
Anton Aluja a Rafael Torrubia b
a University

of Leida, Lleida, and b Autonomous University of Barcelona, Barcelona, Spain

Key Words
Luteinizing hormone W Aggressiveness W Testosterone W
Sensation seeking W Sex hormone-binding globulin

Aggression and Verbal Aggression obtained significantly higher scores in SHBG and TT. These findings support
Zuckerman’s personality model for the sensation-seeking trait.
Copyright © 2004 S. Karger AG, Basel

To evaluate the relationship between sex hormones and
aggressiveness, hostility and sensation seeking we studied 30 healthy males. Using a standardised technique of
radioimmunoassay, we obtained blood values of luteinizing hormone (LH), follicle-stimulating hormone (FSH),
17ß-estradiol (E2), total testosterone (TT), sex hormonebinding globulin (SHBG) and the free androgen index
(FAI). Personality was evaluated by the Buss-Durkee
Hostility Inventory and the Sensation-Seeking Scale,
form V. The results showed a lack of significant correlations between the measures of aggressiveness-hostility
and hormones. Nevertheless, Spearman and Pearson
correlations between Sensation Seeking and testosterone were positive and significant after controlling for
age. Considerably higher correlations were obtained after controlling for LH and SHBG. A group of subjects with
high scores in a factor made up of Experience Seeking,
Disinhibition and Boredom Susceptibility obtained significantly higher scores on TT and FAI. Subjects with
high scores in a factor made up of Assault, Indirect


© 2004 S. Karger AG, Basel

Fax + 41 61 306 12 34

Accessible online at:


Historically, self-reports and direct observations have
dominated the assessment of the relationship between sex
hormones and aggressiveness and hostility in humans.
The Buss-Durkee Hostility Inventory (BDHI) is among
the best-known self-report measures [1]. Working with a
student sample, Persky et al. [2] found a high correlation
between testosterone production rate, plasma testosterone levels, and self-reported measures of aggression and
hostility derived from the BDHI. However, two studies
using the same instrument yielded contradictory results
[3, 4]. Using behavioural measures and criteria, Kreuz
and Rose [5] and Ehrenkranz et al. [6] found that a group
of inmates with a record of violent crimes showed higher
testosterone levels than a group without a violent crime
record. Rada [7] and Rada et al. [8] found that sexual
assault criminals had higher levels of testosterone than
non-sexual assault criminals. Bradford and McLean [9],

A. Aluja
Department of Pedagogy and Psychology
University of Lleida, Complex de la Caparrella, s/n
ES–25192 Lleida (Spain)
Tel. +34 973 702326, Fax +34 973 702305, E-Mail

however, did not find differences in testosterone levels
between diverse groups of sexual assault criminals classified according to their level of violence. More recent studies considering total or free testosterone, either in blood or
saliva, have shown a relationship between testosterone
and aggressiveness-dominance in normal and criminal
male and female samples [10–12]. Recently Gerra [13]
found that subjects with higher scores on the BDHI had
significantly higher basal testosterone levels and aggressive behaviour when compared to low BDHI scorers.
In the review by Archer [14], it was concluded that
comparisons between high and low aggressiveness groups
reveal higher testosterone levels in the more aggressive
groups, despite the only low positive correlations between
trait measures and testosterone. Overall, the author concludes that androgens might influence aggressive behaviour although they are only one of several influencing factors and not the determining one. Probably the opposite
relationship (i.e. the environment and behaviour influencing hormone secretion) is the stronger of the two
Daitzman et al. [15] reported correlations between
both testosterone and 17ß-estradiol and the Disinhibition
(Dis) subscale of the Sensation-Seeking Scale (SSS) in
men. In a second study, Daitzman and Zuckerman [16],
selecting subjects on the basis of extreme Dis scores, confirmed the results obtained in the first study. Male subjects with higher scores on the Dis subscale of the SSS also
had higher levels of plasma testosterone and 17ß-estradiol. Galligani et al. [17] have found that athletes with an
intake of androgens show more disinhibition, extraversion and risk behaviours [see also 18], although other
studies have found no relationship between sensationseeking scales and testosterone [19]. Using the NoveltySeeking scale from Cloninger’s Tridimensional Personality Questionnaire, a number of authors have reported positive relationships between androgens and this scale [20,
21]. Using the Monotony Avoidance scale from the Karolinska Scales of Personality in a sample of young delinquents, alcoholic subjects with high testosterone levels
obtained elevated scores on this scale [22, 23].
For males, the luteinizing hormone (LH) is related to
testosterone through the stimulation of the Leydig cells,
while the follicle-stimulating hormone (FSH) is related to
the initiation and maintenance of sperm production
through the stimulation of the Sertoli cells [24]. Both hormones are related to testicular function, and testosterone
and behaviour studies should control for the effect of
these hormones since any abnormality in the production
of LH and FSH can affect testosterone levels. There are

Relationship between Behaviour and Sex
Hormones in Men

many explanations for abnormalities in the LH and FSH
levels, such as genetic predisposition, nutritional factors
and toxic effects. In alcoholics, for instance, a significant
increase in serum LH, FSH and sex hormone-binding
globulin (SHBG) levels, and a decrease in the free androgen index (FAI) [25] have been found. The SHBG is produced in the liver and transports testosterone to the blood.
It is possible that in alcoholics and heroin addicts, as well
as in subjects with hepatic diseases, there is an increase in
total testosterone due to the effect of SHBG [26]. Stålenheim et al. [27] found significant correlations between testosterone and SHBG in alcoholics, and both compounds
correlated with the antisocial factor (factor II) of the Hare
Psychopathy Checklist-Revised (PCL-R). Taken together,
the aforementioned results suggest that in studies of trait
and behaviour correlates of testosterone, results could be
refined if hormone levels were controlled for LH, FSH
and SHBG.
Summing up, enough evidence exists with regard to the
relationships between gonadal axis activity and aggressive
and sensation-seeking traits. Nevertheless, these studies
do not consider the relationship between the two personality characteristics and sex hormone levels. Also, most
published studies did not control for the mediating role of
LH, FSH and SHBG. The major aim of the present study
was to investigate the relationship between basal levels of
testosterone – and their aromatization into estradiol –,
aggressiveness-hostility (BDHI) and sensation seeking
(SSS-V) in a sample of healthy volunteers in order to clarify the extent to which relationships were attributable to
each one of these traits. The second aim was to ascertain if
the relationships usually found between sex hormones
and these personality traits were also observed after controlling for LH, FSH and SHBG.

The sample consisted of 30 males who volunteered to participate
in the study; they were recruited among hospital sanitary staff and
final year undergraduate medical students. Age ranged between 21
and 40 years (mean age 27.9 B 4.9 SEM). All measures were
collected anonymously. Two questionnaires had missing data, so
both were excluded from statistical analyses. Furthermore, SHBG
measures from two participants were missing. Finally, one participant was also excluded because he had an abnormal FSH level
(25 mU/ml).
Personality Measures
The Spanish version of the Sensation-Seeking Scale, form V [28]
has 40 items and includes the subscales of Thrill and Adventure

Neuropsychobiology 2004;50:102–107


Seeking (TAS), Experience Seeking (ES), Dis, Boredom Susceptibility (BS), and a total Sensation-Seeking score [29]. Reliability alpha
coefficients were between 0.62 and 0.82, very similar to those
obtained in English-speaking samples [29].
The BDHI has the following subscales: Assault (Ass), Indirect
Aggression (Ind), Negativism (Neg), Resentment (Res), Suspicion
(Sus), Verbal Aggression (Ver), Irritability (Irr) and Guilt [1]. Five of
these scales measure aggression (Ass, Ind, Irr, Neg and Ver) and two
measure hostility (Res and Sus). The Guilt subscale was added subsequently. The inventory was adapted to Spanish by De Flores [30].
The split-half reliability was calculated by means of the SpearmanBrown method, obtaining an index of 0.96.
Blood Samples for Hormone Measures
In order to minimize the pulsatile effects on hormone levels,
blood samples were drawn three times (10 ml ! 3) from the antecubital vein between 08.00 and 09.00 h at 20-min intervals. They were
then pooled and transferred to heparinized tubes and centrifuged.
Supernatant plasma was withdrawn and stored in a freezer (–20 °)
before proceeding with each determination. These samples were
radioimmunoassayed for LH, FSH, E2 (Medgenix, Brussels, Belgium; note that for the calculation of E2, cases with values lower than
11 were deleted because the sensitivity of the kit only yielded measures above 10 pg/ml), TT (Immunotech, Marseille, France) and
SHBG, measured using the IRMA method (Farmos Diagnostics,
Turku, Finland). The FAI was also calculated with the following formula: [testosterone in nmol/l divided by SHBG in nmol/l] ! 100. To
change the ng/ml TT into nmol/l, it was multiplied by a constant
value of 0.0347 [31]. Kurtosis and Skewness values indicated that
hormone values met distributional assumptions, except for SHBG.
Statistical Evaluations
Firstly, a combined factor analysis of the SSS-V and the BDHI
was performed in order to reduce the number of variables. To examine the association between hormone levels and personality factors,
we used Spearman’s correlations to avoid biases from the low size of
the sample and the non-normality of the SHBG distribution. We also
performed partial correlations controlling for age, LH, FSH and
SHBG because of the changes in testosterone level due to those variables. However, non-significant correlations may obscure differences
between high and low scorers. Therefore we additionally used the
non-parametric Mann-Whitney U test (one tailed) to compare hormone levels in extreme groups. We performed all analyses using the
Statistical Package for the Social Science, edition 10.


Spearman’s correlations among the hormone measures
show a significant coefficient between LH and TT (0.46;
p ! 0.01). FSH was negatively related to FAI (–0.42; p !
0.05). Furthermore, E2 was significantly related to FAI
(0.50; p ! 0.01). SHBG was also positively related to TT
(0.51; p ! 0.01), and TT to FAI (0.70; p ! 0.01). Note that
partial correlations controlling for age remained significant.


Neuropsychobiology 2004;50:102–107

In order to reduce the variables related to aggression
and sensation seeking, a principal component analysis
was carried out for the SSS-V and BDHI subscales together. Four factors were extracted using eigenvalues 61 as
the extraction criterion. They account for 75.33% of the
variance. The first factor was composed of Guilt, Sus,
Res, Neg and Irr, the second factor of Ass, Ind and Ver,
the third of BS, Dis and ES, and the fourth of TAS. The
Neg and Irr subscales obtained high secondary loadings
on factor II. Given that Neg and Irr had high loadings on
two factors, and TAS is encapsulated in an independent
factor with regard to ES, Dis and BS from the SSS-V, a
second principal component analysis was performed excluding the Neg, Irr and TAS subscales. Three factors
were obtained with an eigenvalue 61, accounting for
73.99% of the variance. The first factor grouped BS, Dis
and ES, and was called Sensation-Seeking factor (SSF).
The second factor was composed of Sus, Res and Guilt,
and was named Hostility factor (HosF). The third factor
was represented by Ass, Ind and Ver, i.e. a factor of
Aggressiveness (AggF).
Table 1 shows Spearman’s correlations between personality factors and hormones, and partial correlations
controlling for age, LH, FSH, and SHBG in different combinations. Age has no role in the observed correlations.
However, when age and LH were controlled, correlations
between TT and FAI with SSF increased (0.47 and 0.46;
p ! 0.01), as did the correlation between TT and SSF
when age and SHBG were controlled (0.45; p ! 0.05). It
should be noted that no significant correlations were
found between hormones and the AggF and HosF. The
following partial correlations are also shown in table 1:
(a) hormones with SSF, controlling for AggF and age;
(b) hormones with AggF, controlling for SSF and age, and
(c) hormones with HosF, controlling for SSF and age.
Note that correlations between TT and SSF do not
improve substantially when the aggressiveness factor and
age were controlled.
In a second step, two groups of high and low scorers
were created for each personality factor. The criterion for
selection was to obtain a percentile equal to or higher than
70 for the high group, or equal to or lower than 30 for the
low group. Plasma determination of hormones was converted into T scores in order to compare mean differences
(fig. 1). High scorers on AggF (n = 10) obtained mean
scores significantly higher than low scorers (n = 8) on
SHBG (p ! 0.04) and TT (p ! 0.03). Figure 1 also shows
the U test between 9 high and 8 low scorers on SSF. Subjects scoring higher on SSF also obtained a higher mean
on TT (p ! 0.02), and on FAI (p ! 0.03). No significant


Fig. 1. Mean hormone values (T scores) for
high and low Aggressive factor (AggF;) Ass +
Ind + Ver for high and low Sensation-Seeking factor (SSF; ES + Dis + BS) groups, and
Mann-Whitney U test comparisons. LH =
Luteinizing hormone; FSH = follicle-stimulating hormone; SHBG = sex hormone-binding globulin; TT = total testosterone; FAI =
free androgen index.

Table 1. Spearman’s rho (#) and partial correlations between personality factors and hormones (correlations including 17ß-estradiol were
only calculated for the 17 cases with values 1 10 pg/ml)


Sensation seeking factor

Aggressivity factor

Hostility factor


Controlling for


Controlling for








Age +

Age + Age + AggF

–0.11 –0.10 –0.13

–0.01 –0.21
0.33 0.28 0.30
–0.08 –
0.47** 0.36 0.45* 0.39*
0.46** 0.42* –

AggF +

Age + Age + Age + SSF

–0.03 0.10 –
–0.23 –0.07 –0.06
0.33 0.43 0.22
0.28 0.22 0.20
0.33 0.28 0.26
0.14 0.18 0.16


0.10 0.06 0.10 0.10

–0.17 –0.06 –0.07
0.39 0.43 0.49 0.43
0.27 –
0.21 0.22
0.27 0.21 0.30 0.30
0.17 –
0.20 0.20

Controlling for

HosF age



Age + Age + Age + SSF



–0.04 0.02 –0.02 –0.03

–0.13 –0.21 –0.23
–0.39 –0.35 –0.30 –0.34
–0.20 –
–0.26 –0.28
–0.20 –0.04 –0.16 –0.16
–0.13 –
–0.01 –0.01

LH = Luteinizing hormone; FSH = follicle-stimulating hormone; E2 = 17ß-estradiol; SHBG = sex hormone-binding globulin; TT = total testosterone; FAI = free androgen index.
* p ! 0.05; ** p ! 0.01.

difference was obtained for the Hostility factor. It turned
out that partialling out age did not show much influence
in the correlational analysis, and therefore can be neglected in this group comparison.


The main results obtained in the current study were:
(1) no significant relationship was found between either
aggressiveness or hostility and gonadal hormones; nevertheless, analysis of extreme groups showed higher testosterone levels in highly aggressive subjects; (2) a positive

Relationship between Behaviour and Sex
Hormones in Men

relationship between testosterone and sensation seeking
was found when using correlations and extreme group
comparisons; (3) the above-mentioned correlations were
substantially increased after controlling for LH and
SHBG, and (4) correlations between SSF and TT or FAI
did not change after controlling for AggF and age, and correlations between AggF and TT or FAI did not change
after controlling for SSF.
Some authors have suggested that age could be a confounding factor in studies linking the gonadal axis to
behaviour since both aggression and TT decrease with
age. It should be noted that most studies which obtain significant relationships between self-reported aggressive-

Neuropsychobiology 2004;50:102–107


ness and personality have been performed with rather
young subjects. Schalling et al. [32] attributed these
results to the fact that aggressive behaviour is relatively
rare in non-clinical adults, and also to the difficulty of
using adequate instruments. In any case, our results suggest that when comparing hormone values between extreme AggF scorers, the most aggressive subjects obtain
significantly higher mean values in SHBG and TT than
the low aggressive ones, although not on the FAI. It is also
noteworthy that age, LH, FSH and SHBG did not modulate the results obtained.
The correlational results between gonadal hormones
and personality show a consistent relationship between a
Sensation-Seeking factor (ES, Dis and BS) and total and
free testosterone. Analyses of extreme groups also confirm
these results. Estradiol also presents a relationship with
the Sensation-Seeking factor, albeit this is not statistically
significant. These results confirm previous evidence [15,
16, 18, 20]. As far as we know, no previous study has controlled the effect of the pituitary hormones, LH, FSH and
of the binding globulin SHBG in the relationships between hormones and personality. In fact, our results indicate that controlling for both LH and SHBG reinforces
the relationship between testosterone and SS. It means
that the free fraction of testosterone is associated with
sensation seeking because it is not mediated by higher
release of LH, which would induce more testosterone, nor
is it mediated by higher production of the SHBG. However, as opposed to sensation seeking, aggressiveness may
be related to free as well as to bound testosterone and
could also depend on LH production, since there was no
change in correlations after partialling out LH and SHBG,
which is a valuable finding for differentiating between
sensation seeking and aggression. Our results suggest that
the relationship between SSF and TT or FAI was not
modulated by aggressiveness or hostility factors. This

would suggest that, in normal samples, the personality
variable associated with sex hormone levels is sensation
seeking rather than aggressiveness or hostility.
The present results show different hormonal profile
patterns for the extreme sensation-seeking and aggressive
groups. While the high SS subjects present high levels of
TT and FAI compared to the low SS, those subjects with
high scores on aggressiveness show higher levels in SHBG
and TT. Thus high SS and the most aggressive individuals
would share high levels of TT, but would on the other
hand differ in their SHBG and FAI profiles. Judging from
the present results, SHBG could be a biological marker of
aggressive behaviour but not of sensation seeking, whereas free testosterone could be more related to sensation
seeking than to aggression in non-clinical healthy subjects.
This study simultaneously explores the relationship
between hormones and aggressive personality traits and
sensation seeking. One question that remains to be answered is what results would have been obtained if
account had been taken of sensation seeking in studies in
which a relationship between aggressive traits and testosterone has been observed. A possible explanation for most
studies might include the notion of testosterone as a common mediator of the sensation-seeking trait and aggressive tendencies. Further specific studies would be required to test this hypothesis.

This research was supported by grants from the Comissio´ Interdepartamental de Recerca i Innovacio´ TecnolÒgica (CIRIT) of the
Autonomous Government of Catalonia, Spain. We also thank Dr.
M.J. Martı´nez de Osaba, endocrinologist at Barcelona Clinical Hospital, and Dr. L.F. Garcı´a for his help in the statistical analysis.

1 Buss A, Durkee A: An inventory for assessing
different kinds of hostility. J Consult Clin
Psych 1957;21:343–349.
2 Persky H, Smith KD, Basu GK: Relation of
psychological measures of aggression and hostility to testosterone production in man. Psychosom Med 1971;33:265–277.
3 Meyer-Bahlburg HFL, Boon DA, Sharma M,
Edwards JA: Aggressiveness and testosterone
measures in man. Psychosom Med 1974;36:


4 Doering CH, Brodie KH, Kramer HC, Moos
RH, Becker HB, Hamburg DA: Negative affect
and plasma testosterone: A longitudinal human
study. Psychosom Med 1975;37:484–491.
5 Kreuz LE, Rose RM: Assessment of aggressive
behavior and plasma testosterone in a young
criminal population. Psychosom Med 1972;34:
6 Ehrenkranz J, Bliss E, Sheard MH: Plasma testosterone: Correlation with aggressive behavior
and social dominance. Psychosom Med 1974;

Neuropsychobiology 2004;50:102–107

7 Rada R: Plasma androgens and sex offenders.
Bull Am Acad Psychiatry Law 1981;8:456–
8 Rada R, Laws D, Kellner R: Plasma testosterone levels in the rapist. Psychosom Med 1976;
9 Bradford J, McLean D: Sexual offenders, violence and testosterone: A clinical study. Can J
Psychiatry 1984;29:335–343.
10 Dabbs JM, Frady RL, Carr TS, Besch NF: Saliva testosterone and criminal violence in young
adult prison inmates. Psychosom Med 1987;


11 Harris JA: Review and methodological considerations in research on testosterone and aggression. Aggress Violent Behav 1999;4:273–291.
12 Aromäki AS, Lindman RE, Ericksson CJP:
Testosterone, aggressiveness, and antisocial
personality. Aggress Behav 1999;25:113–123.
13 Gerra G: Neurottransmitter-neuroendocrine
responses to aggression: Personality influences;
in Raine A, Brennan P (eds): Biosocial bases of
violence. NATO ASI Series. Life Sciences,
292.1997 (pp. 333–335). NY: Plenum Press.
14 Archer J: The influence of testosterone on human aggression. Br J Psychol 1991;82:1–28.
15 Daitzman R, Zuckerman M, Sammelwitz P,
Ganjam V: Sensation seeking and gonodal hormones. J Biosoc Sci 1978;10:401–408.
16 Daitzman R, Zuckerman M: Disinhibitory sensation seeking, personality and gonadal hormones. Pers Individ Differ 1980;1:103–110.
17 Galligani N, Renck A, Hansen S: Personality
profile of men using anabolic androgenic steroids. Horm Behav 1996;30:170–175.
18 Bogaert, AF, Fisher WA: Predictors of university men’s number of sexual partners. J Sex Res
19 Rosenblitt JC, Soler H, Johnson SE, Quadagno
DM: Sensation seeking and hormones in men
and women: Exploring the link. Horm Behav

Relationship between Behaviour and Sex
Hormones in Men

20 Gerra G, Avanzini P, Zaimovic A, Sartori R,
Bocchi C, Timpano M, Zambelli U, Delsignore
R, Gardini F, Talarico E, Brandilla F: Neurotransmitters, neuroendocrine correlates of sensation-seeking temperament in normal humans. Neuropsychobiology 1999;39:207–213.
21 Wang S, Mason J, Charney D, Yehuda R,
Riney S, Southwick S: Relationships between
hormonal profile and novelty seeking in combat-related posttraumatic stress disorder. Biol
Psychiatry 1997;41:145–151.
22 Mattsson A, Schalling D, Olweus D, Löw H,
Svensson J: Plasma testosterone, aggressive behavior, and personality dimensions in young
male delinquents J Am Acad Child Psychiatry
23 Virkkunen M, Kallio E, Rawlings R, Tokola R,
Poland RE, Guidotti A, Nemeroff C, Bissete G,
Kalogeras K, Karonen S, Limoila M: Personality profiles and state aggressiveness in Finnish
alcoholic, violent offenders, fire setters, and
healthy volunteers. Arch Gen Psychiatry 1994;
24 Ducharme JR: Normal puberty: Clinical manifestations and their endocrinology control; in
Collu R, Ducharme JR, Guyda H (eds): Paediatrics endocrinology. New York, Raven Press
Books, 1981, pp 294–327.
25 Vilalta J, Ballescà JL, Nicola´s MJ, Martı´nez de
Osaba, MJ, Antunez E, Pimentel C: Testicular
function in asymptomatic chronic alcoholics:
Relation to ethanol intake. Alcohol Clin Exp
Res 1997;21:128–133.

26 Vermeulen A, Verdonck L: Some studies on the
biological significance of free testosterone. J
Steroid Biochem 1972;3:421–426.
27 Stålenheim EG, Eriksson E, von Knorring L,
Wide L: Testosterone as a biological marker in
psychopathy and alcoholism. Psychiatry Res
28 Zuckerman M, Eysenck SBG, Eysenck HJ:
Sensation seeking in England and America:
Cross-cultural, age, and sex comparisons. J
Consult Clin Psychol 1978;46:139–149.
29 Pérez J, Torrubia R: Fiabilidad y validez de la
versio´n española de la escala de Bu´squeda de
Sensaciones (forma V). Rev Lat Psicol 1986;
30 De Flores T: Paradigma d’agressio´ de Buss en
relacio´ a variables fisiològiques i psicològiques.
Tesis doctoral no publicada. Universidad de
Barcelona, 1982.
31 Vankrieken L: DPC Technical Report: Testosterone and the free androgen index. Los Angeles, DPC, 1997.
32 Schalling D, Edman J, Åsberg M: Impulsive
cognitive style and inability to tolerance boredom: Psychobiological studies of temperamental vulnerability; in M. Zuckerman (ed): Biological Bases of Sensation Seeking, Impulsivity
and Anxiety. New Jersey, Erlbaum, 1983, pp

Neuropsychobiology 2004;50:102–107