This article was downloaded by: [University of Saskatchewan Library]

On: 31 December 2014, At: 21:46

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Psychology & Health

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/gpsh20

Attachment style as a piedictor of sexual attitudes and

behavior in late Adolescence
a a a a
Judith A. Feeney , Candida Peterson , Cynthia Gallois & Deborah J. Terry
a
Department of Psychology , University of Queensland , Australia
Published online: 19 Dec 2007.

To cite this article: Judith A. Feeney , Candida Peterson , Cynthia Gallois & Deborah J. Terry (2000) Attachment
style as a piedictor of sexual attitudes and behavior in late Adolescence, Psychology & Health, 14:6, 1105-1122, DOI:
10.1080/08870440008407370

To link to this article: http://dx.doi.org/10.1080/08870440008407370

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall not be
liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities
whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of
the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Psychologyond Health. 2000. Vol. 14, pp. I 105-1 122 02000 OPA (Overseas Publishers Association) N.V.
Reprints availuble directly from the Publisher Published by license under
Photocopying permitted by license only the H m o o d Academic Publishers imprint,
part o f The Gordon and Breach Publishing Group.
Printed in Malaysia.

ATTACHMENT STYLE AS A PIEDICTOR OF

SEXUAL ATTITUDES AND BEHAVIOR IN LATE
ADOLESCENCE
JUDITH A. FEENEY *, CANDIDA PETERSON,
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

CYNTHIA GALLOIS and DEBORAH J. TERRY

Department of Psychology, University of Queensland, Australia
(Received 10 August, 1998: infinalform 22 January, 1999)

This research applied attachment theory to the study of sexual attitudes and behaviors in a sample of late adoles-
cents. Four hundred and seventy heterosexual undergraduate students completed questionnaires assessingattach-
ment (discomfort with closeness; anxiety over relationships). relationship history, communication about sex.
sexual self-efficacy and locus of control, and attitudes to condoms. Eight weeks later, participants reported on
sexual behaviors occurring during the eight-week interval. and perceived risk of these activities. Both discomfon
with closeness and anxiety over relationships were associated with external locus of control for sexual outcomes,
and with use of drugs before sexual contact. Anxiety over relationships was linked to unsafe sex and to negative
attitudes to condoms, but discomfort with closeness was associated with a more cautious approach to sexual risk-
taking. Some results were qualified by gender differences. and by differences between the full sample and those
who were sexually active. The findings arc discussed in terms of attachment style and its links with communica-
tion and affect regulation.

KEY WORDS: Attachment, sexual attitudes, sexual behavior.

Issues concerning adolescent sexuality have been the focus of considerable research in
recent years, with particular emphasis on sexual attitudes and behaviors that may place
individuals at risk of infection with HIV and other sexually transmitted diseases. Studies of
adolescent sexuality have been conducted in many countries and in a range of samples (col-
lege students, high school students, and "at-risk" groups). These studies suggest that ado-
lescents are tending to engage in sexual behavior earlier than in the past; further, sexually
active adolescents typically have multiple sexual partners before establishing monogamous
relationships, and mostly engage in unprotected sex (Abbott-Chapman and Denholm,
1997; Boyer, 1990; Dunne, Donald, Lucke, Nilsson, Ballard and Raphael, 1994; Moore
and Rosenthal, 1993; Rollins. 1989; Rotheram-Borus, Mahler and Rosario, 1995; Vogels,
van der Vliet, Danz, Hopman-Rock and Visser, 1993).
Research into the sexual attitudes and behaviors of adolescents has produced consensus
on two important points. First, most young people are quite knowledgable about the trans-
mission and prevention of HIV, although they are less knowledgable about other sexually
transmitted diseases (Hillier, Wan and Haste, 1998; Lucke, Dunne, Donald and Raphael,
1993). However, knowledge levels are a poor predictor of safer sex behavior (Fisher and
Fisher, 1992; Rimberg and Lewis, 1994; Seal and Agostinelli, 1994; Vogels et al.. 1993).
Hence, to identify factors predictive of patterns of sexual activity, researchers need to
move beyond knowledge-based variables. Recent research suggests that relevant factors

* Corresponding author. E-mail: j.feeney@psy.uq.edu.au

I105
 1106 J.A. FEENEY EFAL.

include attitudes toward condom use (Gold, Karrniloff-Smith, Skinner and Morton, 1992;
Misovich, Fisher and Fisher, 1997). perceptions of risk (Gold er al., 1992; Seal and Ago-
stinelli, 1994). direct disclosure on sexual issues (Edgar, 1992). and confidence in one’s
ability to plan, negotiate, and control sexual encounters (Edgar, 1992; Misovich et af.,
1997). These findings support the emerging view that patterns of communication and
couple interaction are vital to an understanding of sexual activity (Metts and Fitzpatrick,
1992).
Second, adolescents’.experiences and beliefs about sexual matters vary widely. For
example, gender and ethnic group differences in sexual attitudes and behaviors have been
identified, and there also appear to be distinct groups defined by perceptions of risk and
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

patterns of risk-taking behavior; researchers need to take account of such sources of vari-
ation, rather than characterising adolescents as a homogeneous group (Moore and
Rosenthal, 1992; Rosenthal and Moore. 1991).
Attachment theory is a theoretical perspective which is relevant to the prediction of
sexual attitudes and behaviors, and which acknowledges the importance of individual dif-
ferences in relationship beliefs and experiences. There is already considerable evidence
that adolescentswho report weak attachment to their parents are more likely to be involved
in early sexual activity, and to have large numbers of sexual partners (e.g., Smith, 1997;
Walsh, 1993). These findings support the broader proposition that inadequate parent-child
attachment is a risk factor for adjustment problems in adolescence, including poor self-
esteem and substance abuse (Rice, 1990).
The link between attachment and sexual expression is supported not only by evidence
that inadequate parent-child attachment predicts increased sexual activity, but also by
recent arguments that intimate relationships between adult partners can be conceptualized
as attachments (Ainsworth, 1989; Weiss, 1986, 1991). Consistent with these arguments, a
rapidly growing body of research indicates that individual differences in adult “attachment
styles” (ways of relating to intimate partners) can be reliably measured, and that such
measures predict a wide range of relationship attitudes and behaviors (Feeney and Noller.
1996; Hazan and Shaver, 1987;Shaver and Hazan. 1993).
This body of research is directly relevant to issues concerning sexuality. Based on Bowl-
by’s (1969, 1973, 1980) ethological theory of attachment, Shaver, Hazan and Bradshaw
(1988) argued that bonding between adult partners can be best explained through a model
integrating three behavioral systems: attachment, caregiving. and sexual mating. Accord-
ing to these researchers, the attachment system is pivotal to the establishment of romantic
bonds; it appears early in the course of the individual’s development, and lays the founda-
tion for the other two systems. Examining human sexuality within the context of attach-
ment theory may, as Shaver (1994) suggests, enrich the study of sexuality, a field which
has been criticised as being largely atheoretical (Sprecher and McKinney, 1993).The links
between attachment and sexuality may be particularly important during adolescence, a
time at which the sexual system comes to the fore (Shaver and Norman, 1995).
Some empirical support exists for the link between adult attachment style and sexuality.
Specifically. attachment style appears to be associated with sexual attitudes and behaviors
in ways which are consistent with the defining features of the styles. Secure individuals’
sexual expression reflects an emphasis on intimacy and involvement: these individuals are
less likely than other attachment groups to get involved in one-night stands and in sexual
encounters outside of the primary relationship, and more likely to be involved in mutually
initiated sex (Brennan and Shaver, 1995; Hazan, Zeifman and Middleton, 1994).Avoidant
individuals tend to show an “unrestricted” approach to sexual attitudes and behavior,
 A'lTACHMENT AND SEXUALlTy 1 I07

marked by more casual, uncommitted, nonintimate sex (Brennan and Shaver, 1995;
Feeney, Noller and Patty, 1993). Anxious-ambivalentindividuals frequently initiate affec-
tionate encounters, but tend to report greater enjoyment of holding and caressing than of
more clearly sexual behaviors (Hazan, Zeifman and Middleton, 1994). They also report
less consistent condom use, and express concern that attempts to discuss safe sex issues
with potential sex partners will incur disapproval (Feeney. Kelly, Gallois. Peterson and
Terry, in press).
Although these findings have emerged quite consistently, research also suggests that the
implications of attachment style for sexual expression may be relatively complex, with
some effects being specific to one gender or to one type of measure. For example, anxious/
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

ambivalent females are likely to report engaging in extra-relationshipsex and in practices

such as voyeurism and exhibitionism, whereas anxiouslambivalent males appear to be
more sexually reticent (Gangestad and Thornhill, 1997; Hazan, Zeifman and Middleton,
1994).Further, using a diary methodology,Feeney et af. (1993) found that female avoidant
and male anxiouslambivalentadolescents were the least likely to report engaging in sexual
intercourse during the 6-week period of the study. This finding suggests that gender and
attachment style may jointly affect sexual behavior, further. the low level of sexual inter-
course among avoidant females is surprising, given avoidants' greater reported acceptance
of casual sex, and highlights the importance of considering both attitudes and behaviors in
assessing sexual expression.
Despite empirical support for attachment theory as a useful perspective on adolescent
sexuality, further research in this area is required. The available studies have been based
largely on the three-group attachment typology (secure, avoidant, anxiouslambivalent).
This typology was the dominant model in early studies of adult attachment. However,
researchers are increasingly adopting either a four-group typology (Bartholomew, 1990;
Bartholomew and Horowitz, 1991),or a correlational approach which focuses on the major
dimensions underlying attachment styles: discomfort (or comfort) with closeness, and anxi-
ety over relationships(Feeney, Noller and Callan, 1994;Strahan, 1995).The latter approach
does not assume that individuals can be classified into a small number of discrete groups,
and is also likely to provide more sensitive measurement (Feeney and Noller, 1996).

The present szudy

The present study was designed to assess the association between attachment style and
adolescents' sexual attitudes and behaviors, expanding on previous studies in four ways.
First, consistent with the arguments outlined above, attachment style was measured in
terms of Bartholomew's four groups (secure, preoccupied, dismissing, and fearful), and in
terms of the dimensions of discomfort with closeness and anxiety over relationships.
Second, it was considered important to tap a broad range of measures of sexuality; some
previous studies of attachment and sexuality have focused on a single construct, such as the
degree of acceptance of casual sex. The present study employed measures of sexual atti-
tudes and behavior which are particularly relevant to adolescents, including sexual com-
munication, self-efficacy, locus of control, and perceived riskiness of sexual activities.
Third, the study involved a two-stage process of data collection. At Time 1, we meas-
ured background variables, and general attitudes to sex and to romantic relationships. At
Time 2, sexual behaviors in the last 8 weeks were assessed, focusing mainly on safe vs
unsafe sex. The purpose of the two-stage process was to prevent responses to the sexual
behavior items being influenced unduly by the measures of attachment style and general
 I I08 J.A. FEENEY ETAL.

attitudes to sex, and to facilitate the accurate recall of sexual behaviors by aligning the start
of the recall period to the first testing session.
Finally, a relatively large sample of undergraduate students was recruited for the study.
This was important to ensure stable results from multivariate analyses. to allow the invest-
igation of gender differences, and also to allow comparison of results emerging from sub-
samples defined by sexual experience and recent sexual activity. Some studies of adolescent
sexual attitudes and behaviors have failed to take account of whether respondents have
ever experienced (or experienced recently) the situations described in questionnaire items.
This failure is problematic, given that individuals' responses may be influenced by the per-
sonal salience of the sexual event (Catania, Gibson, Chitwood and Coates, 1990).
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

Based on the literature outlined above, the following hypotheses were derived.

Hypothesis 1. It was hypothesized that attachment style would be related to sexual atti-
tudes, with insecure attachment (discomfort with closeness, anxiety over relationships)
being inversely related to open sexual communication and internal locus of control. It
was further hypbthesized that anxiety over relationships would be inversely related to self-
efficacy in negotiating sexual encounters.

Hypothesis 2. It was expected that attachment style would also be related to safer sex prac-
tice. Specifically. discomfort with closeness and anxiety over relationships were expected
to be associated with greater reported use of alcohol and other drugs during sexual encoun-
ters, and anxiety over relationships was expected to be associated with less consistent use
of condoms.

Given the lack of empirical data on the link between attachment and perceptions of
sexual risk-taking, no predictions were made concerning the relationship between these
variables. Further, no specific predictions were made Concerning gender differences in pat-
terns of association. although this research question was addressed.

METHOD

Participants and procedure

As part of a larger study, undergraduate students who were enrolled in first-year psycho-
logy subjects at the University of Queensland completed a series of questionnaires for
course credit.' For the present project, participants were required to have completed ques-
tionnaires on two separate occasions, as detailed below. A small number of students whose
questionnaire responses indicated that they did not self-identify as heterosexual were
excluded from the analyses reported here.
These criteria resulted in a sample of 470 participants (330 females and 140 males), with
a mean age of 18.38 years and a median of 18 years. Because of the large number of items
in the overall questionnaire package, deletion of cases with missing data produced some-
what smaller samples for some analyses. Of the full sample, 263 participants ( 17 1 females
and 92 males) had had some sexual experience at the start of the study, as defined by their

' The data reported in this paper atz from an ongoing study reported by Feeney. Kelly, Gallois, Peterson and Terry
(in prrss). The sample described in the present paper is a much larger sample; furthcr. except for the attachment
scales and one item assessing condom use. the variables discussed in this paper have not been r e p o d elsewhere.
 ATTACHMENT AND SEXUALITY I109

affirmative response to the question “Have you ever been sexually active with any
person?’. At Time 2, 184 participants ( 123 females and 6 I males) reported having been
sexually active within the eight-week interval immediately prior to that testing session, as
defined by their affirmative response to the question “Have you engaged in sexual relations
with any person in the last eight weeks?“.
The majority of the participants (N = 386) were born in Australia. Of the remainder, the
most common places of birth were Asia (20). New Zealand (1 5). South Africa (12). and
Europe (10). The most common living arrangements reported by respondents were as fol-
lows: with both parents (212), in college accommodation (75). with friends (73), with one
parent (59), and with a romantic partner (16). Of the full sample, 253 participants reported
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

having some religious affiliation; of these, 89 reported being very involved with their reli-
gion, 80 reported moderate involvement, and 45 reported little involvement.
Two separate questionnaires were administered in group sessions, eight weeks apart.
Sessions lasted approximately two hours for the first questionnaire and one hour for the
second. Retention of subjects across the two phases was ensured by informing participants
that attendance at both testing sessions was required in order to gain course credit. To pre-
serve anonymity, a unique identifying code was used by each participant to enable their
two questionnaires to be matched. Measures of demographic variables, relationship history
(including sexual risk-taking), attachment style, and general measures of sexual attitudes
(listed below) were included in the first questionnaire. The follow-up questionnaire
focused on sexual behaviors occurring in the last 8 weeks, and on a more detailed assess-
ment of the perceived riskiness of respondents’ own sexual behavior.

Measures
Demographic variables. Respondents were asked to report their gender, age, country of
birth. religious affiliation, religiosity, and current living arrangements.

Relationship history. Participants reported on several aspects of current and previous rela-
tionships: whether they were in a current dating relationship and, if so, the length and the
sexual exclusivity of this relationship( 1 =not at all exclusive, 8 =totally exclusive); number
of previous dating relationships; and number of previous sexual partners. They were also
asked five questions about sexual risk-taking: whether they had ever engaged in oral,
vaginal, or anal sex without a condom (yedno); whether they thought that their own sexual
activities put them at no risk of HIV infection or transmission (from 1 =agree strongly to
6=disagree strongly), and how risky their sexual activities would need to be before they
would change them (from 1 = any risk at all to 5 =extremely risky).

Attachment. Attachment style was assessed using two measures. First, participants were
asked to read simple descriptions of four attachment prototypes: secure, preoccupied, dis-
missing, and fearful; and to mark the one which best described their own thoughts and feel-
ings about close relationships (Bartholomew and Horowitz, 1991). Second, they completed
the 30-item Relationship Scales Questionnaire (RSQ; Griffin and Bartholomew, 1994).
The RSQ is a measure of adult attachment style which draws together items used by a
number of leading researchers in this area. The items employ a 5-point response format,
from 1 =not at all like me, to 5 = very much like me.
Factor analysis of the 30 items yielded two orthogonal factors: discomfort with closeness
and anxiety over relationships. These two factors explained 43.3%of the total variance,
 Ill0 J.A. FEENEY ETAL.

and correspond to the dimensions identified by previous researchers (Feeney el al., 1994;
Simpson, 1990; Strahan. 1995).The discomfort with closeness factor (referred to hereafter
as discomfort) contained 13 items assessing the degree of comfort with intimacy in rela-
tionships; sample items include “I find it easy to get emotionally close to others” (reverse
scored), and “Romantic partners often want me to be closer than I feel comfortable being”.
This factor essentially contrasts avoidant and secure attachment styles. The anxiety over
relationships factor (referred to as anxiety) contained 10 items assessing fear and worry
about possible loss and rejection; sample items include “I often worry that romantic partners
won’t want to stay with me”. and “I often worry that romantic partners don’t really love
me”. This factor is similar to the anxioudambivalentattachment style. Scales were formed
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

to assess each factor using unit weights, with alpha reliability coefficientsof .84 (discomfort)
and .86 (anxiety).The two scales were not significantly intercorrelated (r=-.08).

Communication about sex. Openness of communication about sex was assessed using 13
items which were answered with respect to each of four target persons: mother, father,
friends (defined as the majority of the subject’s close friends), and current or previous rela-
tionship partner (participants who had never been in a dating relationship did not answer
the questions with respect to this target).’ The items used a response scale from 1= not at
all. to 4 =a great deal. The set of items formed a highly reliable scale for each target person
(Cronbach’salpha ranged from .91 to 94).Sample items include “I can ask them for informa-
tion or advice on sexual matters”. “We have discussed many aspects of contraception”, and
‘They have talked to me about the risks of unsafe sex”.

Sexual locus of conrrol. Seventeen items were used to measuie the extent to which sexual
outcomes were attributed to internal factors, or to external factors such as significant others
or chance. The items employed a 6-point response format, from 1=strongly disagree to
7 =strongly agree. Factor analysis of the items yielded a two-factor orthogonal solution.
explaining 39.7% of the variance. Factor 1 (8 items; a= .72)was labelled Externality, and
tapped the perception that sexual outcomes are determined mainly by luck or by other
people. Sample items from this scale include “It is chiefly a matter of luck if I hit it off
sexually with another person”, “I never seem to know in advance when I’m going to have
sex”, and “If I didn’t do sexually what my partner wants, dhe would leave me”. Factor 2 (7
items; a= .69) was called Internality, and assessed the perception that sexual outcomes are
influenced by one’s own behavior. Sample items include “When it comes to sex, I’ll decide
how far I want to go”. “If I didn’t feel comfortableabout what was happening sexually then
I would stop it”, and “I accept responsibility for getting my sexual needs met”.

Sexual self-eficacy. Perceived ability to perform the behaviors needed to satisfy sexual
needs and to negotiate sexual interactions was measured using 21 items. The items employed
a 7-point response format, from 1=extremely unlikely to 7 =extremely likely. Factor ana-
lysis of the items again resulted in two orthogonal factors, explaining 38.5% of the vari-
ance. Factor 1 (1 1 items; a= .79) was labelled Sexual Expression, and tapped the ability
to initiate sexual activities and inform partners of one’s sexual needs. Sample items include
“I could ask my partner to provide the type of sexual stimulation required”, “I could initiate

’For the measures of communication about sex, sexual locus of control and sexual selfif‘ticacy, some items
were developed specifically for this study, and others were taken from earlier work by Susan Moore, D o m n
Roscnthal. Alison Turtle, Susan Kippax. and June Crawford.
 AITACHMENT AND SEXUALITY 1111

sexual activities”, and “I could teil my partner how to treat me sexually”. Factor 2 (I0
items; 0=.80) was Sexual Negotiation, and focused on the ability to discuss safe sex
and contraception with partners and doctors, and to resist being pressured into unwanted or
unprotected sex. Sample items are “I could ask a potential partner to wait if adequate pre-
cautions are not available at the time”, “I could refuse to do something which I don’t feel
comfortableabout”, and “I could discuss precautions with a doctor”.

Aftifides fo condoms. Ten items assessed attitudes to condoms. Two of the items (“I
believe condoms are boring”; “I believe that condoms reduce sexual satisfaction”) were
rated from 1 =disagree strongly to 6=agree strongly. The remaining eight items focused
on the perceived consequences of regular condom use (e.g., “Condoms protect against
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

HTV/AIDs”’, and “Condoms reduce the intimacy of sex”), and were rated from -3=extremely
unlikely to 3 =extremely likely.

Sexual behaviors. At Time 2, participants who had been sexually active in the preceding
eight weeks reported on their sexual activities during that period. First, they were asked
how many sexual partners they had had (no explicit definition of the term “sexual partner”
was provided). Second, they were asked how often they had discussed HIV/AIDS and con-
traception with their sexual partners (in each case, from 1 = never to 5 = always).Next, they
reported on their use of condoms in the eight-week period. Did you use a condom every
time you had sexual intercourse?(yes/no);How often did you use a condom when you had
sexual intercourse? (from 1 = never to 6 =always); With how many of your sexual partners
did you USE a condom regularly when having intercourse? (from 1 =none to 6=all).
Because sexual encounters do not necessarily involve penetrative sex, participants were
also asked whether they had had non-penetrative sex (defined as nor vaginal, oral, or anal
sex) every time they had sexual contact during the eight weeks. Finally, respondents indic-
ated whether or not they had used any of the following drugs before sexual contact: alco-
hol, marijuana, injecting drugs, other drugs.
Participants who had not been sexually active in the preceding eight weeks were asked
their reasons for not having sex during this time. Eleven reasons were listed: I don’t
approve of sex; I haven’t met the right person; I am still recovering from my last relation-
ship; I prefer other sexual activities (e.g., non-penetrative sex); worry about pregnancy;
wony about AIDS; wony about other STDs; I have a religious objection to pre-marital sex;
my parents would object; my partner doesn’t approve of our having sex; I have no privacy/
opportunity for sex. Respondents ticked as many reasons as applied, and responses were
coded as no =O, yes= 1.
Finally, participants (regardless of recent sexual activity)answered eight items, designed
to tap three aspects of perceived risk. Personal risk was measured using three items assess-
ing perceptions of subjects’ own chances of getting HIVlAIDS (How great are your chances
of getting HW/AIDS?; To what extent are you womed about getting HIV/AIDS?; and
How likely do you think you are personally of contracting HIV/AIDS?). These items
used a 6-point response format, with higher ratings reflecting higher perceived risk. Relat-
ive risk was measured using three items assessing perceptions of one’s own chances of get-
ting HIV/AIDS, compared with friends, with other students, and with the general public (in
each case, 1 =much lower to 5 =much higher). Group risk was assessed using two items
assessing the perception that risk of HIV/AIDS is low for those of similar sexual orienta-
tion to the respondent (1 =agree strongly to 6 =disagree strongly).The items were used to
form the three scales, with a coefficientsranging from .69 to .79.
 1112 J.A. FEENEY ETAL.

RESULTS

Attachment characteristics of the sample

In this paper, data analyses linking attachment style with sexual attitudes and behaviors
focus on the two attachment scales derived from the RSQ. This decision was made for two
reasons. First, there is ongoing debate about the number and the nature of adult attachment
styles, and about the validity of conceptualising attachment styles as discrete categories;
hence, researchers are increasingly measuring attachment in terms of dimensions. Second,
even with relatively large samples, cell sizes become small when breaking the sample
down by attachment style and gender; this is especially the case when, as in the present
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

study, some analyses require subsets of subjects based on patterns of sexual experience and
sexual activity.
However, it was considered important to demonstrate the association between the two
dimensions and the categorical measure, in order to facilitate integration of the present
results with those of studies based on attachment groups. Hence, a MANOVA was con-
ducted relating the categorical (forced-choice) measure to the attachment scales. In this
analysis, attachment style and gender were varied between subjects. In terms of the cat-
egorical measure, the sample consisted of 139 secure individuals (92 females, 47 males),
62 preoccupied (49 females, 13 males), 87dismissing (59 females. 28 males), and 1 16 fear-
ful (93 females, 23 males). (Only 404 participants completed the forced-choice measure).
As expected, the groups differed strongly on the attachment scales (there was no gender
effect, and no interaction of attachment style and gender). Secure and preoccupied parti-
cipants (Ms = 32.92 and 32.04, respectively) reported less discomfort with closeness than
dismissing and fearful participants (Ms =40.98 and 39.23). Further, secure and dismissing
respondents (Ms = 22.99 and 23.28) reported less anxiety over relationships than preoccu-
pied and fearful respondents (Ms = 3 1.14 and 30.13). These results are consistent with Bar-
tholomew’s theoretical model, which portrays dismissing and fearful groups as tending to
avoid close contact with others, and preoccupied and fearful groups as being highly
anxious about, and dependent on, the approval of others (Bartholomew. 1990).

Relationship history
Of the full sample, 200 participants reported currently being in a regular dating relation-
ship. The number of previous dating relationships reported by participants was generally
low: no previous relationships (77). one relationship (66). two relationships (74), three to
five relationships (174). and six or more relationships (60). Similarly, the number of sexual
partners was generally low: Of the respondents who reported some sexual experience, 98
reported having had only one sexual partner, 5 1 reported two partners, 53 reported three to
five partners, and 49 reported six or more.
The attachment dimensions were not reliably related to relationship status (in a current
dating relationship or not), or to relationship length. The scales were also unrelated to
number of previous dating relationships and number of sexual partners. Anxiety over rela-
tionships was correlated, however, with females’ reports of being in less sexually exclusive
relationships (r=-.21, pc.05).
Attachment was also related to females’ reports of previous sexual risk-taking, although
the correlations were not strong ( p c .05 in each case). Discomfort was inversely related to
females’ reports of having had intercourse without a condom (r=-.18 for vaginal inter-
 ATTACHMENT AND SEXUALITY 1113

Table 1 Correlations of attachment scales with Time I reports of sexual communication, locus of control and
self-efficacy

Variable Discomfort Anxiety

Full sample Sexually acrive Full sample Sexually active

Communicarion
with mother -.07 -.24 -.09 -.I7
-.I 1 -.22* -.I6 -.I4
with father -.08 -.24 -.04 -.03
-. 16
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

-.21* -.08 -.I2

with friends -.os -.03 -.07 -.08
-.I2 -.07 .03 .07
with partner -.I3 -.I0 .o I .06
-.26** -.21* -.20* -.I92
Locus of control
internal -.20* -.45*** -.06 -.06
-.06 -.06 -.IS -.28**
external .IS a*** .25* .36***
.25** .36*** .33*** .28**
Self-eficacy
sexual expression -.06 -.21 -.08 -.I I
-.I5 -.09 -.03 -.I4
sexual negotiation -.08 -.07 -.I7 -.33**
-.04 -.os -.I4 -.23*

Note: In each cell. the cop and bottom enlria show the correlations for d e s and femles. respectively. Thc sample generally
repontd m0dcr;Uely Open communication (ovcrall MIon the Cpoinl sale mnged from 1.76 for father to 2.92 for friends). an
internal locus of control (Mson the &point scale were 5. I 8 for intern1 and 2.94 for external). and moderate to high wlf-effacy
(Ms on the 7-point scale were 4.40 for expression and 5.48 for negotiation).

course; -.20 for anal intercourse). In contrast, Anxiety was positively related to females'
reports of having had oral sex without a condom ( r = .20), and to females' reports that their
activities would need to be more risky before they would change them ( r = .22).

Sexual attitudes
Correlations were calculated between the attachment dimensions and the measures of
sexual attitudes. The correlations with the multiple-item scales (communication about sex,
locus of control, and self-efficacy) appear in Table 1. Note that the correlations are reported
separately by gender, in order to evaluate gender differences in the associations. Further,
given that the associations may be affected by whether participants have personally experi-
enced the situations described in the questionnaire items, the correlations are also reported
both for the full sample and for those who reported being currently sexually a ~ t i v e . ~
Reliable links between attachment dimensions and communication about sex were
restricted to females. and were most consistent for communication with relationship part-
ners: In line with Hypothesis 1, both discomfort and anxiety were inversely related to
females' reports of open communication with their partners. By contrast, attachment was

'Correlations w e n also calculated for all those participants who had had some sexual experience (regardless of
whether they were cumntly sexually active). For ease of presentation. these correlations have not k e n tabulated,
but were generally intermediate in value between those for the full sample and those for the sexui~llyactive group.
 1114 J.A. FEENEY ETAL.

unrelated to openness of communication with friends, and links between discomfort and
reports of less open communication with parents were restricted to sexually active females.
Sexual locus of control showed moderate to strong associations with both attachment
dimensions, providing further support for Hypothesis 1. Specifically, insecure attachment
(high discomfort, high anxiety) was associated with the attribution of sexual outcomes to
external, rather than internal, factors. The discomfort dimension of attachment was much
more strongly linked to sexual locus of control for those participants who were currently
sexually active.
Attachment was also related to perceived self-efficacy of negotiating sexual encounters,
but not to self-efficacy of expressing sexual needs. As predicted, the link with self-efficacy
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

of negotiating sexual encounters was confined to the anxiety dimension of attachment,

which was related to lowered perceptions of self-efficacy. This association was significant
only for sexually active males and females.
Correlations between the attachment scales and the items assessing attitudes to condoms
are shown in Table 2. Note that these items were analysed individually, in order to investig-
ate specific barriers to condom use that may be associated with particular attachment
styles. There were relatively few significant correlations between discomfort and responses
to the attitude statements, and these were confined to the sexually active subsample. For
these participants. higher discomfort was related to males’ agreement that condoms protect
against HIVlAIDS and against other STDs, to males’ rejection of the belief that condoms
reduce intimacy, and to females’ rejection of the belief that sex with condoms is boring. By
contrast, the anxiety dimension of attachment was linked with beliefs that condoms are
boring, interrupt foreplay, destroy spontaneity, and reduce intimacy, sexual pleasure. and
sexual satisfaction. Anxiety was also associated with females’ rejection of the belief that

Table 2 Cornlalions of attachment scales with Time I reports of beliefs about condoms

Belief statement Discomfort Anxiery

Fullsample Sexuallyactive Fullsample Sexually active

condom....
Are boring -.I2 -.09 .08 .I9
-.08 -.20* .24* .26**
Reduce satisfaction .03 -.I09 .I0 .2I
.06 , -.12 .21* .23*
Protect against HlVlAlDS .I3 .39** .09 .I4
-.01 .05 -.07 -.I 1
Protect against other STDs .09 .33* .09 .I5
.01 .05 -.05 .09
Reduce pleasure .I5 .05 .03 .I5
-.o I .01 .20* .27**
Intcrmpt foreplay -.06 -. I9 .I 1 .27*
-.07 .09 .21* .24*
Reduce intimacy -.07 -.25* .I6 .29*
.05 -.01 .26** .26**
Destroy spontaneity -.I0 -.16, .26** ,3988
-.03 -.I1 .22* .30**
Show OM’S concern for partner .07 .08 .03 -.04
-.08 -.os -.07 -.24*

Note: In each cell. h e top and bottom cnoien show the cornlatimu f
ad e n and females. mpectivcly.
 ATTACHMENT AND SEXUALITY Ill5

condom use shows concern for one's partner. These links between anxiety and negative
attitudes to condoms were largely unaffected by sample (full sample vs sexually active).

Sexual behaviors
For those participants who reported having been sexually active during the period between
the two testing sessions, correlations were calculated between the attachment scales and
reports of sexual behavior (see Table 3). Note that the only variable which showed no
significant relations with the attachment dimensions was number of sexual partners; this
variable showed limited variability, with most respondents reporting only one sexual part-
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

ner in the eight-week period. The questions concerning frequency of discussion about pre-
cautions, frequency of condom use, and frequency of safe sex (top section of Table 3) were
generally unrelated to discomfort with closeness, although for males only, discomfort was
positively related to reported frequency of discussing HIV/AIDS and to reported use of
condoms on every sexual encounter.
Anxiety was more strongly related to reported sexual behaviors, being consistentlyassociated
with less safe sex practices. Specifically,anxiety was inversely related to reported frequency
of discussing contraception with sexual partners (males only), to frequency of discussing
HIV/AIDS with sexual partners (females only), to overall frequency of condom use (males
only), to regular condom use with all sexual partners (both genders), to the use of condoms
on every sexual encounter (females only), and to safer sex practice (either non-penetrative or
protected sex) on every sexual encounter (females only). These results support Hypothesis 2.

Tabk 3 Correlations of attachment scales with Time 2 reports of recent sexual behavior

Variable Discomfort Anriery

Number of sexual partners .05 .07

.07 .04
Frequency of discussingcontraception .04 -.4l***
.o1 -.I2
Frequency of discussing HIVIAIDS 34.8 -.I1
.05 -.26**
Frequency of condom use .07 -.56***
.I2 -.I5
Regular use of condoms with all partners .I8 -.25*
.I6 -.22*
Use of condoms every time .28* .02
.I 1 -.24'*
Safe sex (protected or non-penetrative) every time .05 -.I3
.02 -.20*
Drug use before sex
alcohol .08 -.I5
-.03 ,358'8
marijuana .39** .I2
.I5 .08
injecting drugs .38** .49***
.04 -.05
Other drugs .39** ,4988.
-.06 ,2858

Nore: In each cell, the top and batom entries show the cornlations for males and females,
respectively.
 1116 J.A. FEENEY ETAL.

Also consistent with Hypothesis 2, the attachment dimensions showed moderately

strong relations with reports of drug use before sexual contact (bottom section of Table 3).
In contrast to the results for condom use and precautions-related discussion, both discom-
fort and anxiety were associated with potentially unsafe behavior. That is, discomfort was
associated with males' reports of using marijuana, injecting drugs, and other drugs, before
sexual contact; anxiety was associated with males' reports of using injecting drugs and
other drugs, and females' reports of using alcohol and other drugs.

Reasons for not having sex

Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

For those participants reporting not being sexually active in the eight-week period before
Time 2, correlations were calculated between the attachment scales and the endorsement of
reasons for not having had sex during this time. Although 11 reasons were listed on the
questionnaire, only 7 of these were endorsed with sufficient frequency (N=30 or more) to
warrant investigation of their relation to the attachment dimensions. These reasons were:
I haven't met the right person; worry about pregnancy; worry about AIDS; worry about
other STDs; I have a religious objection to pre-marital sex; my parents would object; and
I have no privacy/opportunity for sex.
For males, discomfort was moderately strongly related to citing worry about AIDS and
worry about other STDs (r= .47 and .48, respectively, p c .001). For females, discomfort
was related to citing wony about AIDS ( r = .21. p < .05), and anxiety was related to citing
lack of privacy or opportunity ( r = .28, p c .01).

Perceptions of risk
Finally, correlations were calculated between the attachment dimensions and perceptions
of the risk of contracting HIVIAIDS. in terms of personal, relative, and group risk (see
Table 4). These questions assessing perceived risk were relevant to all participants with
some sexual experience, and Table 4 shows the obtained correlations for these participants.
and for the subsample reporting being currently sexually active. The strongest finding from
this set of analyses was the link between attachment and males' perceptions of relative risk.
Perceived risk (relative to friends, other students, and the general public) was associated
positively with discomfort, and inversely with anxiety. Discomfort was also associated
with males' perceptions of greater personal risk. The correlations between males' discom-
fort and perceptions of personal and relative risk were significantly higher for those who

Table 4 Correlations of attachment scales with perceptions of risk

Risk Discomfort Anxiety

Experienced Active Experienced Active

Personal .20* .38** -.02 -.01

.I I .08 -.09 -.08
Relative .24* jg*** -.a*** -.38**
.04 .09 -.I I -.09
GWP .I2 .02 -.I0 -.I9
.I5 .21' -.09 .I0
Note: In each cell, thc top and bottomentries show the correlations
for males and females, respectively.
 ATTACHMENT AND SEXUALITY 1117

were sexually active at the time of the study. The only significant correlation for females
was between discomfort and sexually active females' perceptions of greater risk of HIV/
AIDS incurred by those of their own (heterosexual)orientation.

DISCUSSION

As expected, the attachment measures used in this study were associated with a wide range
of sexual attitudes and behaviors. This finding supports adult attachment theory as a per-
spective on individual differences in intimate relationships, and suggests that the theory
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

offers a useful approach to the understanding of adolescent sexual expression.

Specifically,at Time 1, the attachment scales were related to most of the sexual attitudes
assessed in this study (communication about sexual issues with significant others; sexual
locus of control; self-efficacy in negotiating sexual encounters; attitudes to condoms),
although links with communication about sexual issues were mainly restricted to commun-
ication with sexual partners. The attachment scales were also related to relationship history
variables relevant to safer sex (exclusivity of the current relationship; past involvement in
unsafe behaviors; degree of risk required before being prepared to change behaviors). At
Time 2, the attachment scales were related to reports of sexual behavior (frequency of safe
sex discussion, condom use, safer sex, use of drugs before sexual contact),reasons for not
having sex, and perceptions of the risk associated with own sexual behaviors.
It should be acknowledged that some of the significant correlations between attachment
scales and sexual attitudes and behavior were small in size. However, several of the cor-
relations were quite strong. highlighting the utility of the attachment perspectiveon adoles-
cent sexual activity. In particular, males' discomfort was strongly correlated with sexual
locus of control, and males' anxiety was a moderately strong correlate of drug use before
sexual contact, less frequent discussion of contraception with sexual partners, and less fre-
quent condom use. Further. there was evidence that although anxiety is associated with a
tendency to engage in unsafe sexual practices, it is also linked with perceptions of lower
risk associated with such sexual practices. and less willingness to change them. Given this
pattern of inter-related cognitions and behaviors, it seems that individuals who are highly
anxious about their relationships may be quite resistant to messages about the importance
of safer sex practice.
The observed link between insecure attachment and difficulties in sexual communica-
tion and negotiation fits with other research which highlights the communication patterns
associated with the various attachment styles. Secure attachment predicts reports of
balanced patterns of self-disclosure and constructive approaches to problem-solving. In
contrast, avoidant attachment has been associated with low levels of self-disclosure and
conversational involvement, and anxioudambivalent attachment has been associated with
coercion and high levels of conflict (Feeney et al., 1994; Mikulincer and Nachshon, 1991;
Pistole, 1989). These findingscan be understood in terns of the interactionalgoals and rela-
tionship concerns that characterizeeach style. Secure individuals seek a balance of closeness
and autonomy in relationships, whereas avoidant individualsemphasize independence at the
expense of intimacy. and anxioudambivalentindividuals fear rejection and desire extreme
intimacy (Feeney and Noller. 1996).It is not surprising that these goals and concerns influ-
ence communication in the sexual context, as well as in the course of everyday interactions.
Both of the attachment dimensionswere related to the use of alcohol and other drugs before
sexual contact. Although we have no direct evidence that such drug use was associated
 Ill8 J.A. FEENEY ETAL.

with more risky sexual practices in the present sample, previous research suggests that
this is likely to be the case. In adolescent samples (in contrast to adult samples),drinking is
linked to the occurrence of unprotected intercourse (Temple and Leigh, 1992). These
researchers suggest that alcohol may have more of an influence on sexual decision-making
for adolescents, because they are not developmentally equipped to make “safe sex” deci-
sions; sex is a relatively new experience for most adolescents, and drinking before sexual
contact may result in avoidance of responsibility and loss of personal control. Further,
MacDonald, Zanna and Fong (1996) report correlational, field, and laboratory studies sup
porting the hypothesis that alcohol decreases the likelihood of condom use during casual
sex. The studies also suggest a mechanism for this relationship: specifically, it is argued
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

that alcohol decreases cognitive capacity, and causes people to attend to the most salient
cues in a situation (in this case, the perceived benefits of having sexual intercourse).
The increased use of alcohol and drugs reported by insecure participants is consistent
with a study of university students reported by Brennan and Shaver (1995). In that study,
avoidant attachment was related to the amount and frequency of drinking; in addition, both
avoidant and anxiouslambivalent attachment were related to drinking in order to cope with
womes and with negative moods. Brennan and Shaver argue that insecure individuals’ use
of alcohol to dispel1 anxiety and tension reflects the link between attachment style and the
strategies used to regulate negative affect. They further suggest that avoidant individuals’
tendency to engage in casual, nonintimate sex may serve a similar function.
Although both attachment dimensions were associated with the use of drugs before
sexual contact, in many other respects, the pattern of association differed for the two
dimensions. This finding highlights the fact that attachment style needs to be considered as
a multidimensionalconstruct; limited information is likely to be gained from studies which
define attachment in terms of simple degree of security or strength of attachment.
More specifically, it is important to note that all of the observed links between anxiety
over relationships and sexual variables were in the direction of anxiety predicting riskier
behaviors. In contrast, and despite the association between discomfort and drug use before
sexual contact, there was evidence that discomfort was related to less risky behavior in
some regards. For example, females high in discomfort were somewhat less likely to report
ever having had vaginal or anal intercourse without a condom; further, for males, discom-
fort was associated with reports of more positive beliefs about condoms, more frequent dis-
cussion of HIV/AIDS. more consistent condom use, greater perceived risk of HIVIAIDS
(both personal and relative), and abstaining from sex because of concerns about AIDS and
other STDs.
The finding that high levels of comfort with closeness may interfere with safe sex
practice fits with previous data based on a similar student sample, which suggest that
many young people see intimacy as incompatible with safe sex (Gardner, 1992). Similarly,
Misovich et al. (1997) note that individuals in relationships that are perceived to be com-
mitted and intimate generally do not use condoms, even if they do not know their own or
their partner’s HIV status. AIDS preventive behaviors are generally seen as indicative of
troubled or uncommitted relationships. Further, as a relationship becomes more invested,
partners are motivated to achieve a state of trust in one another; this may result in biased
interpretation of the partner’s behavior, both past and present (Misovich ef al., 1997).
In interpreting the observed associations between attachment and measures of sexual
attitudes and behavior, it is important to note that the attachment scales were unrelated to
relationshipstatus and relationshiplength, measured at Time 1. Hence, the associationsare
not confounded by these relationship variables. The attachment scales were also unrelated
 ATTACHMENT A N D SEXUALITY 1119

to the number of previous dating relationships and previous sexual partners. This finding
provides further evidence that the associations between attachment style and measures of
sexuality are unconfounded by such relationship variables; at the same time, these latter
null results are somewhat unexpected, in light of research linking insecure attachment with
relationship instability and with more casual sexual involvement(Brennan and Shaver, 1995;
Kirpatrick and Hazan, 1994). These null results probably reflect the early stage of sexual
experiencerepresented here, resulting in overall low numbers of dating and sexual partners.
To summarise the results discussed so far, a consistent pattern of results has emerged for
the link between attachment dimensions and the expression of sexuality. That is, anxiety
over relationships was generally linked to unsafe sexual practice. By contrast, discomfort
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

with closeness was linked to increased likelihood of using drugs before sex, but also to a
more cautious attitude to sexual risk-taking. These general assertions need to be qualified,
however, for a number of reasons.
First, there were some clear gender differences in the observed patterns of prediction.
For example, the use of drugs before sexual contact was linked to both discomfort and
anxiety for males, but only to anxiety for females. More generally, discomfort was a
stronger correlate of sexual attitudes and behaviors for males than for females, with links
between discomfort and a more cautious attitude to sexual risk-taking being evident for
males only. Although previous research has similarly suggested that females' discomfort
with intimacy is less predictive of the quality of dating relationships than is males' discom-
fort (Collins and Read, 1990; Simpson, 1990). it is not clear at this stage why females' dis-
comfort has little effect on their approach to sexual involvement.
The second point that qualifies the general pattern of results noted above is that relations
between attachment scales and the safe sex variables measured at Time 2 depended on the
exact wording of the question. In particular, the relation between attachment and condom
use depended on whether the question assessed the overall frequency of condom use, the
regular use of condoms with all sexual partners, or the use of condoms on every sexual
encounter. This finding shows the importance of including multiple measures, in order to
gain a complete picture of relations between sexual activity and other variables; further, it
supports the decision to analyse these questions separately, rather than forming a compos-
ite measure of safer sex behavior.
The third point of qualification concerns the effect of the sample studied. The associa-
tions between attachment scales and some of the variables measured at Time 1 depended
quite strongly on the sample used (full sample, sexually experienced, sexually active in the
last eight weeks). This pattern was particularly evident for the association between males'
discomfort and sexual locus of control. Without using larger samples or a fully longitudinal
methodology, it is difficult to establish the reason for this pattern of results. It is possible,
for example, that insecure individuals who are sexually inexperienced consider that they
could exert a moderate degree of control over the outcomes of sexual interactions, but that
perceptions of control decrease with the onset of sexual experience. Regardless of the
explanation, the finding has important implications for researchers, because it suggests that
adolescents may respond differently to questions tapping hypothetical vs real situations,
and questions tapping recent vs more distant experiences.
Although the present study is subject to the usual criticisms directed at self-report
research, a number of factors suggest that the data are likely to show reasonable reliability
and validity. First, reasonably high levels of stability have been demonstrated for reports of
the frequency of sexual intercourse and consistent condom use in a 6-week period (Jaccard
and Wan, 1995). Further, Catania et al. (1990) note that socially desirable responding to
 I120 J.A. FEENEY ETAL.

questions of a sexual nature is less likely under private conditions (e.g., self-administered
questionnaires). These researchers also discuss evidence that participation bias in studies
of sexuality is not severe, unless the research involves invasive procedures or substantial
effort beyond that required to answer the questions.

SUMMARY AND CONCLUSIONS

This study indicates that attachment dimensions are reliably related to many aspects of
sexual attitudes and behaviors among late adolescents. Insecure attachment was linked to
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

increased drug use before sexual contact. Anxiety over relationships was also linked to
less safe sexual practice, but discomfort with closeness was associated with a tendency
for males to be cautious about sexual risk-taking. The results support the utility of the
attachment perspective on adolescent sexuality, but also highlight the complex relations
between attachment dimensions and sexual outcomes. Future research would benefit
from more detailed investigations of the attitudes, emotions, and cognitions which may
mediate the link between attachment style and sexual behavior. For example, those who
are highly anxious about their relationships may engage in more risky sexual practices
because they fear that non-compliance with partners’ wishes will jeopardize their rela-
tionship; the tendency to engage in drug use before sexual cpntact may reflect their
attempts to boost self-confidenceor to manage negative mood states. Research into these
issues has the potential to explain how attachment insecurity is played out in patterns of
sexual activity.

Acknowledgement
We are grateful to Peni Timmins, Malcolm McCamish and John Gardner for their help
with design, data collection and data analysis.This work was supported by a grant from the
Australian Research council to J.A. Feeney, C. Gallois. D.J.Terry and M. McCamish.

References
Abbott-Chapman. J. and Dcnholm, C. (1997) Adolescent risk taking and the romantic ethic: HIVIAIDS aware-
ness among Year I I and 12 students. Australian and New Laland Journal of Sociology. 33.306-32 I .
Ainsworth. M.D.S. (1989) Attachments beyond infancy. American Psychologist, 44,709-7 16.
Bartholomew. K. (1990) Avoidance of intimacy: An attachment perspective. Journal of Social and Personal
Relationships, 7. 147-178.
Bartholomew, K. and Horowitz, L.M. (1991) Attachment styles among young adults: A test of a four-category
model. Journal of Personality andSocial Psychology. 61.226244.
Bowlby. J. (1969)Atrachmenr and loss: Vol.1.A n a c h n r . New York: Basic Books.
Bowlby. J. (1973) A n a c h n r and loss: V01.2. Separation: Anxiety and anger. New Y o k Basic Books.
Bowlby, 1. (1980) Anachmrnr and loss: Vo1.3.Loss. New York: Basic Book.
Boyer. C.B. (1990) Rychosocial. behavioral and educational factors in preventing sexually transmitted diseases.
Adolescent Medicine: Stare of the Ari Reviews, 1.597-613.
Brennan, K.A. and Shaver, P.R. (1995) Dimensions of adult attachment, affect regulation, and romantic relation-
ship functioning. Personality and Social Psychology Bulletin, 21,267-283.
Catania J.A., Gibson, D.R., Chitwood. D.D. and Coates. T.J. (1990) Methodological problems in AIDS bchavi-
oral research: Influences on measurement e m r and participation bias in studies of sexual behavior. Psycho-
logical Bulletin, 108.339-362.
Collins, N.L. and Read, S.J. (1990) Adult attachment, working models, and relationship quality in dating couples.
Journal of Personality and Social Psychology, 58.644-663.
 ATTACHMENT AND SEXUALITY I121

Dunne. M.P., Donald. M., Lucke. J.C.. Nilsson. R., Ballad. R. and Raphael. B. (1994) Age-related increase in
sexual behaviours and decnase in regular condom use among adolescents in Australia. Inrernarional Journal of
STDsandAIDS.S,4147.
Edgar. T. (1992)A compliance-based approach to the study of condom use. In: T. Edgar, M.A. Fitzpatrick and V
S . Freimuth (Eds.)AIDS: A communicarionperspective (pp.47-67). Hillsddc. NJ: Erlbaum.
Feeney. J.A., Kelly, L.. Gallois. C.. Peterson, C. and Terry. D.J. (in press). Attachment style. assertive commun-
ication and safer sex behavior. Journal of Applied Social Psychology.
Feeney, J.A. and Noller. P. ( 1996)Adulr arrachmmr. Thousand Oaks, CA: Sage.
Feeney. J.A., Noller, P. and Callan. V.J. (1994) Attachment style, communication and satisfaction in the early
years of marriage. In: K. Bartholomew and D. Perlman (Eds.) Advances in personal relarionships Vo1.5:
Arrachmenr processes in adulrhood (pp.269-308). London: Jessica Kingsley.
Feeney, J.A.. Noller, P. and Patty, J. (1993).Adolescents’ interactions with the opposite sex: Influence of attach-
ment style and gender. Journal of Adolescence, 16, 169-186.
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

Fisher. J.D. and Fisher, W.A. (1992)Changing AIDS-risk behavior. Psychological Bullerin. 111.455-474.
Gangestad. S.W. and Thornhill. R. (1997)The evolutionary psychology of extrapair sex: The role of fluctuating
asymmetry. Evolurion and Human Behavior, 18.69-88.
Gardner. J. (1992)Prediction of AIDS risk behaviours: The influence of norms and past behaviour. Unpublished
dissemtion. University of Queensland.
Gold. R.S., Kmiloff-Smith. A.. Skinner, M.J. and Morton, J. (1992)Situational factors and thought processes
associated with unprotected intercourse in heterosexual students. Aids Care, 4.305-323.
Griffin, D.W. and Banholomew, K. (1994) The metaphysics of measurement: The case of adult attachment. In:
K. Bartholomew and D. Perlman (Eds.) Advances in personal relationships VoI.5: Arrachmenr processes in
adulrhood (pp. 17-52). London: Jessica Kingsley.
H m n , C. and Shaver, P.R. (1987)Romantic love conceptualized as an attachment process. Journal of Personal-
ity and Social Psychology. 52.5 1 1-524.
Hazan. C..Zeifman, D. and Middleton. K. (1994)Adulr romrrnric arrachmenr aflecrion. and sex. Paper presented
at the 7th International Conference on Personal Relationships, Groningen, The Netherlands. July.
Hillier. L.. Warr, D. and Haste. B.(1998)Rural youth: HIV/STD knowledge levels and sources of information.
Australian Journal of Rural Health. 6, 18-26.
Jaccard, J. and Wan. C.K. (1995)A paradigm for studying the accuracy of self-reports of risk behavior relevant to
AIDS: Empirical perspectives on stability, m a l l biases, and transitory influences. Journal of Applied Social
Psychology. 25,1831-1858.
Kirkpatrick. L.E. and H m , C. (1994)Attachment styles and close relationships: A four-year prospective study.
Personal Relarionships. 1. 123-142.
Lucke, J.C., Dunne, M.P.. Donald, M. and Raphael, B. (1993)Knowledge of STDs and perceived risk: A study of
Austnlian youth. Venereology,6.57-63.
MacDonald, T.K.. Zanna, M.P. and Fong, G.T. (1996)Why common sense goes out the window: Effects of
alcohol on intentions to use condoms. Personuliry and Social Psychology Bullerin. 22.763-775.
Met& S. and Fitzpatrick. M.A. (1992)Thinking about safer sex: The risky business of “know your partner”
advice. In: T. Edgar, M.A. Fitzpatrick and V.S. Freimuth (Eds.) AIDS: A communicarionperspecrive(pp. 1-19).
Hillsdale. NJ: Erlbaum.
Mikulincer. M. and Nachshon, 0.(1991)Attachment styles and patterns of self-disclosure. Journalof Personality
and Social Psychology. 61.2 13-33 I.
Misovich. SJ..Fisher, J.D. and Fisher, W.A. (1997)Close relationships and elevated HIV risk behavior: Evid-
ence and possible underlying psychological processes. Review of General Psychology, 1.72-107.
Moore, S.M. and Rosenthd. D.A. (1992)The social context of adolescent sexuality: Safe sex implications.
Journal of Adolescence. 15.4 15-435.
Moore, S.M. and Rosenthal. D.A. ( 1993)Sexuality in adolescence. London: Routledge.
Pistole, M.C. (1989)Attachment in adult romantic relationships: Style of conflict resolution and relationship
satisfaction. Journal of Social and Personal Relarionships. 6.505-5 10.
Rice. K.G. (1990) Attachment in adolescence. A narrative and meta-analytic review. Journal of Yourh and
Adolescence, 19.5 I 1-538.
Rimberg. H.M. and Lewis. R.J.(1994) Older adolescents and AIDS: Correlates of self-reported safer sex. Journal
of Research on Adolescence, 4.453464.
Rollins, B. (1989)Sexual aftirudesand behaviours: A review of rhe lirerarure. Melbourne: Australian Institute of
Family Studies.
Rosenthal. D.A. and Moore. S.M. (1991)Risky business: Adolescents and HIVIAIDS. Yourh Srudies, 10.20-25.
Rotheram-Borus. M.J., Mahaler. K.A. and Rosario. M. (1995)AIDS prevention with adolescents. AIDS Educa-
tion and Prevenrion. 7.320-336.
Seal, D.W. and Agostinelli. G.(1994) Individual differences associated with high-risk sexual behaviour: Implica-
tions for intervention programmes. AlDS Care. 6.393-397.
Shaver. P.R. (1994.August). Arrachmenr, care giving, and sex in adulr romanric relationships. APA Invited
Address, Division 8,Los Angeles, CA.
 I122 J.A. FEENEY ETAL

Shaver, P.R. and Hazan, C. (1993) Adult romantic attachment: Theory and evidence. In: D. Perlman and W. Jones
(as.) Advances in personal relationships Vo1.4 (pp.29-70). London: Jessica Kingsley.
Shaver, P.R., Hazan. C. and Bradshaw. D. (1988) Love as attachment: The integntion of thrce behavioral
systems. In: R J . Sternberg and M.Barnes (Eds.) The Psychology ofLove (pp. 68-99). New Haven, CT:Yale
University Press.
Shaver. P.R. and Norman, AJ. (1995) Attachment theory and counseling psychology: A commentray. Counseling
Psychologist,23.491-500.
Simpson, J.A. (1990)Influence of attachment styles on romantic relationships. Journal of Personality and Sociul
P ~ ~ h o l o g59.97
y . 1-980.
Smith. C.A. (1997) Factors associated with early sexual activity among urban adolescents. Social Work. 42.
334-346.
Sprccher, S. and McKinney. K. (1993) Sexuality. Newbury Pa& CA: Sage.
Slrahan, B J. ( 1995) Redictors of depression: An attachment theoretical approach. Journal of Fumily Srudies, 1.
Downloaded by [University of Saskatchewan Library] at 21:46 31 December 2014

33-47.
Temple. M.T.and Leigh. B.C. (1992. Alcohol consumption and unsafe sexual behavior in discrete events.
Journal of Sex Research, 29,207-2 19.
Vogels. T.. van der Vliet, R.. Dan& M..Hopman-Rock. M.and Visser. A. (1993) Young people and sex: Behavi-
our and health risks in Dutch school students. International Journal of Adolescent Medicine and Healrh, 6,
137-147.
Wdsh, A. (1993) Drug use and sexual behavior. Users,experimenters, and abstainers.Jouml of Social Psycho-
logy, 132.691693.
Weiss. R.S. (1986) Continuities and transformations in social relationships from childhood to adulthood. In:
W. W. Hartup and Z . Rubin (Fds.) Relarionships and development (pp.95410). Hillsdale, NJ: Erlbaum.
Weiss. R.S. (1991) The attachment bond in childhood and adulthood. In: C. M.Parkes. J. Stevenson-Hinde and
P. M d s (Eds.)Atrachmenr ucross the lifr cycle (pp.66-76). London: TnvistocWRoutledge.