Responses to unprotected extradyadic sex by one's partner: testing predictions from interdependence and equity theory. Journal of Sex Research v34, n4 (Fall, 1997):387
Buunk, Bram P.; Bakker, Arnold
B. Responses to unprotected extradyadic sex by
one's partner: testing predictions from interdependence and equity
theory. Journal of Sex Research v34, n4 (Fall, 1997):387
(11 pages). [Abstract][Long
Display]
COPYRIGHT 1997 Society for the Scientific Study of Sex Inc.
Sexual involvement of one's partner in an extradyadic sexual
relationship is usually interpreted as a serious threat to the
intimacy and exclusivity of the relationship and evokes strong
negative emotional reactions in most individuals (Buunk, 1995).
Numerous studies have illustrated that offended partners respond
with anger, which generates strained interactions, arguments,
threats, and, not infrequently, violence (e.g., Bringle & Buunk,
1991; Buunk, 1995; Daly, Wilson, & Weghorst, 1982). Evolutionary
theorists have suggested that these strong emotional responses are
due to mechanisms that are the endproduct of human evolution.
According to such theorists, human males have an evolved tendency to
be upset upon learning of sexual intercourse between their partner
and another male because of the risk of investing in another male's
offspring. Females in such a situation would especially be concerned
with the possibility that their partner might invest resources in
the children of another woman (Buss, Larsen, Westen, &
Semmelroth, 1992).
However, extradyadic sex of one's partner may also be threatening
for other reasons. Such behavior, particularly when it occurs
unprotected, may involve the risk of the transmission of sexually
transmitted diseases (STDs). The threat of Acquired Immune
Deficiency Syndrome (AIDS) especially has added a new dimension to
the impact of extradyadic sexual relationships upon the primary
relationship. The current study is focused upon the responses to
unprotected extradyadic sex of one's partner in a Dutch sample.
Despite the fact that people have become more aware of the risks of
unprotected sexual contacts with multiple partners, there is no
evidence that the frequency of non-monogamous heterosexual
relationships in the Netherlands has decreased since the discovery
of AIDS. For instance, in a Dutch sample of convenience, Prins,
Buunk, and VanYperen (1993) found that the fear of AIDS did not have
any impact upon the intention to engage in extradyadic sexual
relationships. More importantly, a representative study on sexuality
in The Netherlands showed that 5% of all individuals with a steady
relationship had in the previous year entered into casual
extradyadic sex (VanZessen & Sandfort, 1991). No fewer than
three quarters of all participants with extradyadic sexual
experience reported having had unprotected vaginal intercourse in
these encounters, and they were also having unprotected intercourse
with the steady partner.
We assessed a number of behavioral responses that might occur
when individuals find out that their partner has had unprotected
extradyadic sex. First, as suggested by research on responses to the
partner's extradyadic sex (Buunk, 1995), it seems rather obvious
that individuals may be so angry and upset that they turn away from
their partner and may even consider ending the relationship. We
refer to this response as angry retreat. Second, individuals may
respond with accommodation, i.e., adapting to the partner, by
expressing loyalty and trying to understand the partner's behavior
(cf. Rusbult, Verette, Whitney, Slovik, & Lipkus, 1991). Third,
an assertive response may occur (e.g., by taking precautionary
measures such as demanding condom use within the relationship,
requiring that the partner take an HIV-antibody test, and demanding
that the partner in the future refrain from unprotected extradyadic
sex). The responses may differ, depending on the unfaithful
partner's attitude (e.g., whether the partner shows regret, whether
the partner wants to end the relationship). Moreover, these
responses are probably not mutually exclusive, and individuals may
exhibit several responses simultaneously. Nevertheless, the current
research is focused upon three variables in the offended individual
that may make a particular response relatively more dominant:
commitment to the relationship, extradyadic sexual willingness, and
the intention to use condoms with new sexual partners. These
variables were formulated on the basis of two closely related
theories, interdependence theory (Kelley & Thibaut, 1978;
Rusbult, 1980, 1983) and equity theory (Walster, Walster, &
Berscheid, 1978).
According to interdependence theory, individuals form and
continue relationships in the light of the benefits these
relationships offer. Individuals will be satisfied with their
relationships when the rewards, such as sexual pleasure, intimate
interactions, and satisfying joint activities, outweigh the costs,
such as distress, pain, and embarrassment (Kelley, 1979; Rusbult,
1983; Thibaut & Kelley, 1959). However, interdependence theory
goes beyond the simple notion that individuals pursue relationships
merely for hedonistic reasons. The theory is based on the assumption
that in the course of a relationship individuals become attached to
and concerned with the welfare of the other, leading to a motivation
to continue the relationship despite certain dissatisfactions.
According to Rusbult (1980, 1983), commitment is the subjective
experience of dependence in a close relationship and is
characterized by a feeling of psychological attachment to the
partner, accompanied by the desire to maintain the relationship.
Rusbult proposed that commitment to the relationship is directly
based upon three factors: high relationship satisfaction (e.g., when
the relationship has high rewards and low costs and exceeds what one
expects from the relationship and perceives that others are getting
in their relationship), low quality of alternatives to the current
relationship (e.g., the perception of few attractive alternative
partners, a low need for independence, or a perceived decline in the
state of living when ending the relationship), and high investment
size (e.g., having invested time and energy in the relationship;
having common friends, shared memories, a joint identity as a
couple, and joint property). Although these factors are usually
correlated, numerous crosssectional, longitudinal, and scenario
studies have supported the hypothesis that these three factors have
independent and additive effects upon commitment in various types of
interpersonal bonds, accounting for 50-90% of the variance in
commitment (for a review, see Rusbult & Buunk, 1993).
Particularly relevant to the current issue are the predictions of
interdependence theory for the way individuals respond to problems
that arise in their relationship. High commitment implies a
long-term orientation, including a motivation to maintain the
relationship "for better or worse," a readiness to take into account
the interests of the partner, and a stable tendency to engage in
pro-relationship behaviors (Rusbult & Buunk, 1993; Rusbult et
al., 1991). More specifically, Rusbult et al. reasoned and showed
that high commitment promotes a high willingness to accommodate, for
example, by inhibiting impulses to react destructively and
increasing constructive responses when the other partner engages in
a behavior that is potentially destructive to the relationship. This
leads to the hypothesis that when the other partner has engaged in
unprotected extradyadic sex, accommodation would be particularly
predicted by a high commitment to the relationship (Hypothesis 1).
According to interdependence theory, although satisfaction, quality
of alternatives, and investment size influence commitment,
commitment is the main predictor of behaviors related to the
maintenance of the relationship. Because commitment primarily
promotes the willingness to accommodate, interdependence theory
would predict that the determinants of commitment will not have an
effect upon the responses to unprotected extradyadic sex of the
partner in addition to commitment.
From the perspective of equity theorists, hypotheses can be
formulated about the most important determinants of the two other
responses to a partner's unprotected extradyadic sex: angry retreat
and assertiveness. Equity theory is employed here, not as a
competing, but rather as a complementary theory to interdependence
theory. In fact, equity theory is closely related to interdependence
theory but emphasizes reciprocity rather than dependency (Buunk
& VanYperen, 1991). The notion that the most satisfying and
stable forms of social interaction and relationships are
characterized by reciprocity has been a cornerstone of many
approaches in the social and behavioral sciences. For instance,
anthropologists and sociologists have analyzed the role of
reciprocal exchanges in enhancing group solidarity (Emerson, 1981),
and marital therapists have developed programs aimed at increasing
the awareness of reciprocity and establishing reciprocal exchanges
of rewarding behaviors (Liberman, Wheeler, deVisser, Kuehnel, &
Kuehnel, 1980). Reciprocity exists mainly in the eye of the
beholder: It refers to the perception that the other is willing to
do, and actually does, what one is willing to do, and actually does
for the other. According to equity theory (Walster et al., 1978),
individuals prefer equitable, reciprocal relationships in which the
balance between inputs and outcomes is the same for both partners,
and a lack of reciprocity will be accompanied by negative feelings.
Many studies guided by equity theory have indeed shown that a lack
of reciprocity is related to dissatisfaction in romantic
relationships (e.g., Hatfield, Traupmann, Sprecher, Utne, & Hay,
1985; Van Yperen & Buunk, 1990).
Thus, according to equity theory, to obtain rewards, individuals
have to provide rewards themselves (Walster et al., 1978; VanYperen
& Buunk, 1994). Relationships are in general more satisfying and
stable when there is a reciprocal exchange of rewards and the
partners mutually accept from the other what they allow for
themselves. Reciprocity seems especially an important concern when
it involves potentially high rewards and costs, such as those
associated with involvement in extradyadic sex. As shown by Buunk
(1982, 1995), individuals who reported they would not engage in
extradyadic sex and those who actually had never done so have
relatively strong negative emotional responses to extradyadic sexual
behavior of their partners (whether it occurs unprotected). Because
feeling inequitably treated usually results in anger (Walster et
al., 1978), we expect that angry retreat is especially likely to
occur among those who feel they would never engage in extradyadic
sex, and who therefore feel that their partner violates a rule of
reciprocity. Thus, angry retreat as a response to unprotected
extradyadic sex on the part of one's partner will be particularly
predicted by a low extradyadic sexual willingness (Hypothesis 2).
It is assumed that the fact that one's partner engages in
unprotected extradyadic sex will also violate the rule of
reciprocity, in particular for those who have a high intention to
use condoms with a new partner. For these individuals, the partner
engages in a type of behavior in which they feel they themselves
would not engage. Although this perception might evoke angry
retreat, we suggest on the basis of equity theory that condom use
intention will be a particularly important predictor of
assertiveness. The reason for this prediction is that assertiveness,
as defined previously, refers explicitly to dealing with the fact
that the partner had unprotected sex, i.e., by asking the partner to
reduce in one way or the other the potential risks that the
encounter entailed. According to equity theory, such requests would
be a way to restore the violation of reciprocity that the partner's
behavior entailed. Indeed, individuals with a strong intention to
use condoms would probably focus relatively strongly upon the unsafe
sex aspect of their partner's behavior. It is therefore hypothesized
here that assertiveness will be predicted better by condom use
intention than by commitment and extradyadic sexual willingness.
Thus, Hypothesis 3 states that an assertive response will be
especially characteristic of those who feel they would use condoms
with new sexual partners.
We employed as criterion variables the way individuals felt they
would respond if their partner would have unprotected extradyadic
sex (see also Figure 1). There were a number of reasons to focus
upon expected responses to unprotected extradyadic sex rather than
upon the way individuals had responded to such a situation in the
past. First, past experiences are subject to various memory
distortions. For example, individuals might have a different
recollection of the same behavior when the relationship ended in
divorce than when the partner felt guilty and had done a lot to
regain the other's trust. Second, unprotected extradyadic sex of
one's partner might have occurred in rather diverse contexts. For
instance, as noted previously, the unfaithful partner's attitude
(e.g., whether he or she shows regret, whether he or she wants to
end the relationship), might affect the nature of responses. By
asking expected responses in a well-described situation, the
immediate context of the responses is held constant for all
participants. Third, it would be, difficult to relate past responses
to variables assessed at present. For instance, how would one
unequivocably interpret the fact that assertiveness in the past is
more common among individuals who are now inclined to practice safer
sex? By using the ways individuals would respond as criterion
variables, it is theoretically and methodologically more appropriate
to relate responses to variables assessed at the same point in time
and to consider these variables as predictors of the responses. For
similar reasons, in research on jealousy, anticipated responses to
jealousy-evoking situations rather than actual responses often are
assessed (e.g., Buunk, 1995).
[Figure 1 ILLUSTRATION OMITTED]
In Figure 1, we present an overview of the variables and their
hypothesized interrelationships. Because the criterion variables
concerned intended, and not actual, behavior, we also examined
primarily the willingness to become involved in extradyadic sex (in
general, with or without using condoms) and the intention to use
condoms with new sexual partners as the primary predictors of the
responses to the partner's unprotected extradyadic sex. In this way,
criterion and primary predictor variables are of the same conceptual
level. We thus supposed that the responses were primarily affected
by intentions and that behaviors in the past possibly affecting
these intentions--i.e., extradyadic sex and condom use in the
past--had no independent effect on the responses. We refer to this
last group of variables as "secondary predictor variables." The
third primary predictor variable, commitment, may be considered as a
variable of the same level as the other two primary predictor
variables, because it refers to an important degree to the intention
to continue the relationship and to the expectation that the
relationship will last. The determinants of
commitment--satisfaction, alternative quality, and investment
size--are considered secondary predictor variables.
Method
Participants
In the current study we used a convenience sample of 251
heterosexual participants, including 179 women (71%) and 72 men
(29%) (Buunk & Bakker, 1997). The age distribution was as
follows: 30 years or younger, 68%; between 30 and 40 years of age,
15%; between 40 and 50 years of age, 13%; and older than 50 years of
age, 4% (M = 28.54, SD = 10.08). At the time of the study, all
participants were in a close relationship. The mean length of the
relationship was M = 5.44 years (SD = 6.25), with a range of 30
years, a minimum of 1 month and a maximum of 30 years. Forty-two
percent had a relationship of 5 or fewer years, 18% a relationship
of more than 5 but fewer than 10 years, and 17% a relationship of
more than 10 years. We distinguished among four levels of
relationship status (1 = involved in a more or less steady
relationship, 2 = involved in steady relationship, 3 = cohabiting,
and 4 = married). Twenty-six percent of the sample were married, and
23% co-habited. According to their own report, 37% did have a
committed relationship although they were not married and did not
cohabit, and 14% had a more or less committed relationship. The term
committed relationship is a translation of the Dutch expression
vaste relatie, a term that is widely used and unequivocably refers
to a committed intimate relationship.
Measures
Secondary Predictor Variables
Determinants of commitment. Because the way the data were
collected imposed certain restrictions on the length of the
questionnaire, we decided to use brief scales, each consisting of
two items to assess each determinant of commitment. In research in
this area it is not uncommon to employ only a few items (Buunk &
Bakker, 1997), and the determinants of commitment in the current
study were not major, but only secondary predictor variables.
Satisfaction with the relationship was measured by two items from
the Relational Interaction Satisfaction Scale (Buunk, 1990) that in
a pretest had the highest correlations with items used by Rusbult
(1980, 1983) to assess satisfaction. The two items were "Things go
well between us" and "I feel happy with my partner." Response
options ranged from (1) "never" to (5) "very often." The correlation
between the two items was r = .68, p [is less than] .001. The whole
Relational Interaction Satisfaction Scale was not used because this
scale contains a number of items that may be viewed as assessing
commitment (i.e., "I consider leaving my partner"). The measures for
alternative quality and investment size were based upon scales
developed by Rusbult (1980, 1983) that were adapted and translated
in Dutch. Alternative quality was measured by summing the items "Can
you imagine that someone else would take your partner's place?" (1 =
absolutely not, 5 = possibly) and "How important is it for you to
have a steady relationship?" (1 = not important, 5 = extremely
important). The correlation between the two items was r = .45, p [is
less than] .001. Investment size was measured by summing the
following items: "At this moment, how much would you lose should
your relationship be terminated?" and "How much have you emotionally
invested in your relationship (in the sense of time, energy,
self-disclosure, and joint experiences)?" (For both questions, 1 =
little, 5 = everything). The correlation between the two items was r
= .46, p [is less than] .001.
Past extradyadic sex. This variable was assessed with two
questions. It was considered important to make a distinction between
casual extradyadic sex and relational extradyadic sex in the sense
of long-term affairs. To assess the frequency of past extradyadic
casual sex, participants were asked to indicate on a five-point
scale, ranging from "never" (1) to "more than 10 times" (5), how
often during the preceding five years they had had a one-time sexual
contact with someone else than their steady partner. In addition,
past extradyadic relational sex was assessed by asking with the same
five-point scale how often during the preceding five years the
participants had had a long-term sexual relationship with someone
else than their steady partner. In line with earlier research (e.g.,
Buunk, 1980, 1982), the definition of long-term was left to the
participants. For both questions, the term steady partner referred
to a steady partner at the time they engaged in extradyadic sex with
someone else. The time frame of five years was chosen more or less
arbitrarily. However, the assumption was that extradyadic sex that
had occurred longer ago might not be salient any longer and might
therefore have little impact on current responses. A shorter period
might lower the chance of the occurrence of extradyadic sex, which
would have resulted in little variance on this variable.
Condom use in the past. This variable was assessed with three
questions. Participants were asked how often they "had used a condom
when having sexual intercourse with a new partner," "had abstained
from sexual intercourse with a new partner when condom use was
impossible," and had insisted on condom use with a new partner, even
when that new partner did not want to use condoms" (1 = never, 5 =
often). Cronbach's [Alpha] for this scale was .56.
Primary Predictor Variables
Commitment. The measure for commitment was a translation of a
scale developed by Rusbult (1980, 1983) and contained four items
that assessed the extent to which participants felt committed to
their relationship and to their partner: "To what extent do you feel
attached to your relationship with your partner?" (1 = not at all, 5
= extremely), "Do you feel committed to maintaining your
relationship with your partner?" (1 = not committed, 5 = completely
committed), "Do you want your relationship to last forever?" (1 =
definitely not, 5 = definitely), and "How likely is it that your
relationship will end in the near future?" (1 = very unlikely, 5 =
very likely; reversely scored). The four items were combined in one
index for commitment and represented a reliable scale; Cronbach's a
in this study was .82.
Extradyadic sexual willingness. This construct was measured by
providing participants with an item from the Extramarital Behavioral
Intentions Scale (Buunk, 1990), "Would you engage in sexual
intercourse with someone else than your steady partner if an
occasion were to present itself?" (1 = absolutely not, 5 =
absolutely yes). The formulation of this item was based upon the
theory of reasoned action (Ajzen, 1991). The reason to use this item
was that it has been used as a focal variable in earlier research
(Buunk, 1995) and it is a clear and obvious indicator of the
willingness to have sex outside the primary relationship when the
opportunity would arise. When one would ask simply if participants
would be willing to engage in extradyadic sex, some participants
might answer no because they are not sure if they would ever have
the opportunity.
Condom use intention. Guided by the way behavioral intentions
need to be assessed according to the theory of reasoned action
(Ajzen, 1991), to assess this variable, we used three items asking
what people planned to do when they would have sex with a new
partner: "Use a condom when you have sexual intercourse with a new
partner," "Abstain from sexual intercourse with a new partner when
condom use is impossible," and "Insist on condom use with a new
partner, even when that new partner does not want to use condoms" (1
= absolutely not, 5 = absolutely). These items referred to precisely
the same behaviors as the items in the scale for condom use in the
past. Cronbach's a = .86. This variable thus refers to the use of
condoms with new sexual partners, including, but not necessarily
limited to, extradyadic partners.
Criterion Variables
Responses to partner's unprotected extradyadic sex. Because
individuals can learn in various ways about the fact that their
partner has had unprotected extradyadic sex, the situation to which
the participants were asked to respond was formulated as broadly as
possible. The participants were presented with the following
introduction: "When you would find out (for example, because your
partner would tell you) that your partner has had unprotected sex
with someone else (by not using condoms during sexual intercourse),
how would you respond?" Next, a number of responses were presented,
and for each of these, participants were asked to indicate on a
five-point scale how likely they thought they would respond in that
way (1 = absolutely not, 5 = absolutely). These items were used to
construct three scales. The scale for accommodation contained three
items referring to an accommodating attitude. The items were "I
would try to discuss things in an open and honest way and together
try to find a solution," "I would try to find out why my partner had
done this and try to limit together the negative consequences," and
"I would, despite everything, keep loving my partner." Cronbach's
[Alpha] for this scale is .77. The second scale, angry retreat,
assessed the degree of anger, upset, and the tendency to avoid the
partner and end the relationship. This scale consisted of the items
"I would be angry," "I would feel powerless," "I would worry," "I
would be inclined to end the relationship," and "I would not want to
see my partner for a while," Cronbach's [Alpha] = .82. The third
scale, measuring assertiveness, contained the items "I would demand
that we would in the future always use condoms when making love," "I
would demand that this would never happen again," and "I would
demand that my partner take an HIV-antibody test," Cronbach's
[Alpha] = .68. Items were presented in a fixed order, with the items
for angry retreat first, the items for accommodation next, and the
items for assertiveness last.
A factor analysis with varimax rotation was conducted on all 12
items in the 3 scales assessing the responses to unprotected
extradyadic, sex. In general, for factor analysis a ratio of 5 to 10
participants per item is suggested (DeVellis, 1991). Thus, the N in
the current study, providing a ratio of more than 20 participants
per item, was sufficiently large. The factor analysis provided
support for the conceptual independence of the three scales. A
solution with three factors was produced, explaining, respectively,
34%, 23%, and 9% of the variance. The eigenvalues were,
respectively, 3.75, 2.57, and 1.01. On the first rotated factor, all
items in the angry retreat scale, and none of the other items,
loaded higher than .70. On the second rotated factor, all items in
the accommodation scale, and none of the other items, loaded higher
than .70. On the third factor, two items in the assertiveness scale,
and none of the other items, loaded higher. than .70. The third item
in this scale N would demand that this would never happen again")
loaded .58 on the first factor and .48 on the third factor. This is
not surprising, given the content of this item. Although the item
was not very discriminating, it was decided to keep it in the
assertiveness scale to prevent that the number of items in the three
scales would be too discrepant.
Procedure
The study was part of a larger study on safer sex in heterosexual
relationships conducted in 1992. The sample was selected from a
larger sample (see also Buunk & Bakker, 1995, 1997) who were
recruited by announcements in national newspapers and magazines and
on television. The announcements explained that the research
addressed opinions about safer sex in heterosexual relations.
Furthermore, a number of people who had placed a personal ad in a
daily newspaper were contacted by a letter. The letter underlined
the importance of the study for future AIDS-prevention activities.
Because it seems likely that individuals who have engaged in risky
and promiscuous sexual behavior would know others who had done the
same, a "snowball" procedure was used to obtain as many participants
as possible. Therefore, everyone who signed up for the study
received two copies of the same questionnaire: one for her- or
himself and one for a friend, partner, or acquaintance. Both
questionnaires were accompanied by a letter that explained the goal
of the study once again. After one month, non-participants received
a reminder. In the accompanying letter, participants were promised a
small present (a ballpoint pen that was sent through the mail) when
they participated in the study. Of the total sample, 24% were
recruited through television, 34% through national magazines and
newspapers, 11% through personal ads, and 31% through others.
Eight hundred twenty-one persons were originally recruited, 87%
of whom indicated that they were "exclusively heterosexual" or
"Predominantly heterosexual." The definition of heterosexual was
left to the participants. Although the study was explicitly
announced as directed toward heterosexual individuals, 13% reported
to be "bisexual," "predominantly homosexual," or "exclusively
homosexual." We decided to drop these participants, because this
group was small and might be very unrepresentative of the bisexual
and gay population, given the fact that they responded to an
announcement for a study on heterosexual relationships. Of the 711
heterosexual participants, half were selected at random to fill out
a questionnaire on features of their current or past committed
intimate relationship. This questionnaire contained the variables
relevant for the current research. Of those who filled out the
relationship questionnaire, for this article the 251 respondents who
were, according to their report, currently married, cohabiting, or
involved in a committed, or more or less committed, relationship
were selected. The number of participants in the following analyses
varies because of missing data.
Results
Descriptive Information
Incidence of extradyadic sexual behavior. In the preceding 5
years, 15% of the sample had had 1, 19% had had 2 to 5, and 10% had
had more than 5 casual extradyadic sexual partners. Moreover, in the
same period, 17% had had 1, 10% had had 2 to 5, and 5% had had more
long-term (relational) extradyadic sexual partners. (Not all of
these extradyadic relationships had occurred necessarily during the
present relationship.) More than half the sample (56%) reported
having had sexual intercourse with a new partner without using a
condom during this 5-year period. This was not necessarily an
extradyadic partner; for the 50% with a relationship of 5 or fewer
years, this could be a partner before the current relationship, or
the current partner, who sometime in the past 5 years was, of
course, a "new" partner.
Correlations between the variables. In Table 1 the correlations
between all variables in the study are presented for descriptive
reasons. Because of the large number of correlations, and the
potential for Type I error, we only considered correlations
significant for which p [is less than] .000. Because the
correlations are not presented here to test theoretical predictions,
we will not discuss here all correlations but only point to a few
noteworthy associations. First, accommodation was negatively
correlated with angry retreat but not with assertiveness, whereas
assertiveness and angry retreat were highly correlated with each
other. Second, extradyadic sexual willingness was highly correlated
with extradyadic sex in the past, but condom use intention had no
correlation with condom use in the past. Third, remarkably, those
who had engaged more often in casual extradyadic sex and those with
a high willingness to engage in extradyadic sex were relatively less
likely to use condoms with new sexual partners. Fourth, the
variables based upon Rusbult's (1980) interdependence
theory--satisfaction, alternative quality, and investment
size--correlated, as could be expected, highly with commitment.
Fifth, not surprisingly, commitment correlated negatively with the
frequency of past extradyadic relational sex and extradyadic sexual
willingness. Sixth, the frequency of relational extradyadic sex was
positively correlated with the frequency of casual extradyadic sex,
indicating that people engaging in both types of behaviors were in
part the same people.
Table 1
Correlations among the Variables
1 2 3 4
1 Accommodation - -.28(***) .09 .28(***)
2 Angry retreat - .49(***) .14(*)
3 Assertiveness - .21(***)
4 Commitment -
5 Extradyadic sexual willingness
6 Condom use intention
7 Satisfaction
8 Alternative quality
9 Investment size
10 Condom use in the past
11 Past extradyadic casual sex
12 Past extradyadic: relational sex
5 6 7
1 Accommodation .05 .19(***) .24(***)
2 Angry retreat -.53(***) .18(**) .18(**)
3 Assertiveness -.35(***) .44(***) .18(**)
4 Commitment -.41(***) .28(***) .70(***)
5 Extradyadic sexual willingness - -.28(***) .42(***)
6 Condom use intention - .19(***)
7 Satisfaction -
8 Alternative quality
9 Investment size
10 Condom use in the past
11 Past extradyadic casual sex
12 Past extradyadic: relational sex
8 9 10
1 Accommodation .18(**) .22(***) .12(*)
2 Angry retreat -.14(*) .04 -.09
3 Assertiveness -.10(*) .06 .12(*)
4 Commitment -.73(***) .61(***) -.05
5 Extradyadic sexual willingness .30(***) -.23(***) .12(*)
6 Condom use intention .19(***) .18(**) .11(*)
7 Satisfaction .45(***) .40(***) -.02
8 Alternative quality - .59(***) .15(*)
9 Investment size - -.06
10 Condom use in the past -
11 Past extradyadic casual sex
12 Past extradyadic: relational sex
11 12
1 Accommodation .02 -.09
2 Angry retreat -.28(***) -.20(***)
3 Assertiveness -.27(***) -.21(***)
4 Commitment -.20(***) -.23(***)
5 Extradyadic sexual willingness .56(***) .42(***)
6 Condom use intention -.19(***)
7 Satisfaction -.21(***) .20(***)
8 Alternative quality .17(**) .20(***)
9 Investment size -.06 -.08
10 Condom use in the past .24(***) .13(*)
11 Past extradyadic casual sex - .51(***)
12 Past extradyadic: relational sex -
(*) p <.05
(**) p <.01
(***) p <.001
Sex differences. A MANOVA showed that women scored significantly
higher on two of the three scales assessing the responses to
unprotected extradyadic sex of the partner, multivariate F(3,241) =
7.49, p [is less than] .001. Although men (M =, 12.12, SD = 2.99)
and women (M = 11.65, SD = 2.55) did not differ significantly in
accommodation, F(1,243) = 1.63, ns, compared to men, women indicated
that they would respond with - more angry retreat, F(1, 243) =
21.54, p [is less than] .001 (for women, M = 18.78, SD = 4.12; for
men, M = 16.00, SD = 5.03), and with more assertiveness, F(1, 243) =
9.69, p [is less than] .001 (for women, M = 10.55, SD = 3.19; for
men, M = 9.28, SD = 3.27).
Testing the Hypotheses on the Predictors of Accommodation, Angry
Retreat, and Assertiveness
There were three primary predictors in the study (commitment,
willingness to engage in extradyadic sex, and intention to use
condoms with new sexual partners) and three criterion variables
(angry retreat, accommodation and assertiveness) as responses to the
partner's unprotected extradyadic sex (see Figure 1). For each
response, a specific predictor was hypothesized to be particularly
characteristic. Accommodation was supposed to be most characteristic
for those with a high commitment to the relationship, angry retreat
for those with a low willingness to engage in extradyadic sex, and
assertiveness for those with a high intention to use condoms with
new partners. To test these hypotheses, three hierarchical
regression analyses were executed, one with each response
(accommodation, angry retreat, and assertiveness) as a dependent
variable. In the first step, a number of relevant demographic
variables (age, gender, length of relationship, educational level,
and relationship status) were entered as control variables. Because
the three responses share common variance, it was necessary to
eliminate the variance because of the two other responses to be able
to assess if there was indeed, as we hypothesized, a unique
predictor for each response. Therefore, in each regression the two
other responses were entered in the second step. Thus, for instance,
in the analysis with accommodation as the dependent variable, in the
second step angry retreat and assertiveness were entered. In this
way, the variance in a given response (in this case, accommodation)
that overlapped that of the other responses (in this case, angry
retreat and assertiveness) was eliminated. In the third step, the
three primary predictors (commitment, willingness to engage in
extradyadic sex, and intention to use condoms with new sexual
partners) were entered simultaneously. The results of these
analyses, as well as the regression coefficients ([Beta]s) in the
final regression equation, are presented in Tables 2, 3, and 4. With
1 exception (out of 15), none of the demographic variables was an
independent predictor of any of the three responses. However, in all
three final regression equations, the two other responses had
significant [Beta]s.
Table 2
Summary of Hierarchical Regressions Analyses
with Accommodation as the Dependent Variable
R [R.sup.2]ch Fch [Beta]
Step 1 .20 .04 1.91
Gender (1 = female, 2 = male) -.03
Educational level .01
Relationship status .01
Age .10
Length of relationship -.14
Step 2 .40 .12 16.94(***)
Assertiveness .19(*)
Angry retreat -.41(***)
Step 3 .50 .08 8.22(***)
Commitment .31(***)
Extradyadic sexual willingness .08
Condom use intention .13
Note: [R.sup.2]ch refers to increase in [R.sup.2] by entering
step. Fch refers to F value of this increase. [Beta] refers to
regression weights for the final equation after all variables were
entered.
(*) p <.05
(**) p <.01
(***) p <.001
Table 3
Summary of Hierarchical Regressions Analyses
with Angry Retreat as the Dependent Variable
R [R.sup.2] Fch [Beta]
ch
Step 1 .47 .22 13.22(***)
Gender (1 = female, 2 = male) -.06
Educational level -.04
Relationship status -.10
Age -.15(*)
Length of relationship -.04
Step 2 .66 .21 41.61(**)
Accommodation -.26(***)
Assertiveness .33(***)
Step 3 .72 .09 13.70(***)
Commitment .04
Extradyadic sexual willingness -.31(***)
Condom use intention -.01
Note: [R.sup.2]ch refers to increase in [R.sup.2] by entering
step. Fch refers to F value of this increase. [Beta] refers to
regression weights for the final equation after all variables were
entered.
(*) p <.05
(**) p <.01
(***) p <.001
Table 4
Summary of Hierarchical Regressions Analyses
with Assertiveness as the Dependent Variable
R [R.sup.2]ch Fch [Beta]
Step 1 .29 .09 4.31(***)
Gender (1 = female, 2 = male) .05
Educational level .01
Relationship status -.10
Age -.08
Length of relationship -.00
Step 2 .55 .21 35.08(***)
Accommodation .15(*)
Angry retreat .40(***)
Step 3 .64 .11 13.52(***)
Commitment .02
Extradyadic sexual willingness -.02
Condom use intention .35(***)
Note: [R.sup.2]ch refers to increase in [R.sup.2] by entering
step. Fch refers to F value of this increase. [Beta] refers to
regression weights for the final equation after all variables were
entered.
(*) p <.05
(**) p <.01
(***) p <.001
Accommodation. As Table 2 shows, the regression with
accommodation as a dependent variable did not result in a
significant amount of explained variance when the demographic
variables were entered in the first step. Entering the two other
responses (angry retreat and assertiveness) in the second step
produced a significant increase in variance, and in the third step
the three primary predictors also generated an increase in variance.
However, in line with Hypothesis 1, of the three predictors,
commitment was the only significant predictor, [Beta] = .31, p [is
less than] .001. Thus, individuals who would respond especially with
understanding and open communication aimed at preserving the
relationship if their partner engaged in unprotected extradyadic sex
were particularly found among those who felt highly committed to
their relationship.
Angry retreat. Table 3 shows that the regression with angry
retreat as a dependent variable resulted in a significant amount of
explained variance when the demographic variables were entered in
the first step. Entering the two other responses (accommodation and
assertiveness) in the second step produced a significant additional
increase in variance, and in the third step the three primary
predictors also generated a significant increase in variance.
However, in line with Hypothesis 2, of the three primary predictors,
extradyadic sexual willingness was the only significant predictor,
[Beta] = -.31, p [is less than] .001. Thus, individuals who
indicated that they would especially react with anger, upset, and
retreat when their partner would engage in unprotected extradyadic
sex were found particularly among those who reported they would not
engage in extradyadic sex.
Assertiveness. The regression with assertiveness as a dependent
variable resulted in a significant amount of explained variance when
the demographic variables were entered in the first step (Table 4).
In the second step, the two other responses (accommodation and angry
retreat) produced a significant increase in variance, and in the
third step the three primary predictors also generated an increase
in variance,. As predicted in Hypothesis 3, only condom use
intention was a significant predictor, [Beta] = .35, p [is less
than] .001. Thus, individuals who said they would be assertive by
demanding precautionary measures when their partner engaged in
unprotected extradyadic sex were particularly those who reported
they would practice safer sex with new sexual partners.
Analyses for men and women separately. To examine the robustness
of the findings, we conducted the same analyses as described
previously for men and women separately. These produced virtually
identical results for both sexes, and these results were the same as
in the sample as a whole. We will present here only the [Beta]s of
the primary predictors after they had been entered in the third
step. With accommodation as a criterion variable, commitment was the
only significant predictor, among men, [Beta] = .27, p [is less
than] .05; among women, [Beta] = .33, p [is less than] .001. Condom
use intention (for men, [Beta] = .21, p = .07; for women, [Beta] =
12, p = .07), and extradyadic sexual willingness (for men, [Beta] =
.18, ns; for women, [Beta] = 04, ns) did not have significant
effects upon accommodation. With angry retreat as a criterion
variable, extradyadic sexual willingness was the only significant
predictor, among men, [Beta] = -.32, p [is less than] .001; among
women, [Beta] = -.31, p [is less than] .001. Condom use intention
(for men, [Beta] = -.14, ns; for women, [Beta] = .03, ns), and
commitment (for men, [Beta] = .04, ns; for women, [Beta] = .08, ns)
did not have significant effects upon angry retreat.' With
assertiveness as a criterion variable, condom use intention was the
only significant predictor, among men, [Beta] = .35, p [is less
than] .01; among women, [Beta] = .35, p [is less than] .05.
Commitment (for men, [Beta] = -.18, p = .09; for women, [Beta] =
.09, ns), and extradyadic sexual willingness (for men, [Beta] =
-.06, ns; for women, [Beta] = -02, ns) did not have significant
effects upon accommodation.
Additional analyses. Finally, we examined the total sample
through a stepwise regression if the "secondary predictor
variables," i.e., condom use in the past and extradyadic sex in the
past, and the three predictors of commitment--satisfaction,
alternative quality, and investment size (see Figure 1)--did explain
additional variance in each of the three responses after the
demographic variables, the two other responses, and the three major
predictors (commitment, willingness to engage in extradyadic sex,
and intention to use condoms with new sexual partners) had been
entered in the hierarchical regressions described earlier. These
analyses produced only one significant effect. For accommodation and
assertiveness, the secondary predictor variables did not explain any
additional variance. However, alternative quality made a small
additional contribution to angry retreat, [R.sup.2] change = .02,
Fchange = 7.97, p [is less than] .01, [Beta] = .19, p [is less than]
.01. Thus, when all other variables were controlled, those
respondents with a more positive perception of alternatives for the
current relationship would respond with more angry retreat when
their partner engaged in unprotected extradyadic sex.
Discussion
We examined the predictors of the ways individuals reported they
would respond when their partner engaged in unprotected extradyadic
sex. In line with the hypothesis formulated on the basis of
interdependence theory (Rusbult et al., 1991), the results showed
that the typical response for those high in commitment was
accommodation. Moreover, as predicted on the basis of equity theory,
for those with a low inclination to engage in extradyadic sex, angry
retreat was the most characteristic response, and for those with a
strong inclination to use condoms with a new partner,
assertiveness--demanding precautionary measures from the
partner--was the most typical response. These results are
particularly noteworthy, because they were established while
controlling for demographic variables and were found among men as
well as women.
The results with respect to accommodation are in line with
interdependence theory. As argued by Rusbult et al. (1991), high
commitment implies a readiness to take into account the interests of
the partner, which expresses itself, among others, in a high
willingness to accommodate, for example, by responding
constructively when the other partner engages in a behavior that is
potentially destructive to the relationship. Thus, the current study
expands the work of Rusbult et al. by showing that even with respect
to a rather extreme form of destructive behavior--engaging in
unprotected extradyadic sex--individuals with high commitment are
still more likely than individuals with low commitment to "stick to
their partner" by accommodation, and thus to adapt to the partner,
by expressing loyalty and trying to understand the partner's
behavior. In addition, the fact that satisfaction, alternative
quality, or investment size had no direct effects upon this response
is in line with interdependence theory. Indeed, this theory assumes
that primarily commitment, and not its determinants, promotes the
willingness to accommodate.
Support was also found for the predictions derived from equity
theory. First, angry retreat (i.e., anger, upset, and a tendency to
leave the partner) was particularly characteristic of those with a
low inclination to engage in extradyadic sex. Of the three
responses, this is the only one that indicates a sense of betrayal,
and equity theory would predict precisely that those who would not
engage in a destructive behavior would feel betrayed and angry when
the partner engaged in such behavior. These findings are in line
with a number of studies by Buunk (1982, 1995) that suggest that
negative emotional responses to extradyadic sex by the partner are
strongly determined by whether one would engage in extradyadic sex.
Further in line with equity theory is the finding that assertiveness
as a response to the partner's risky behavior was particularly found
among those who reported they would use condoms with a new partner.
Apparently, when the partner is perceived as not doing what one
would do, a sense of equity is violated, which one aims to restore
by demanding that the other "undoes" in some sense the wrongdoing.
In sum, these findings suggest that the responses to the partner's
unprotected extradyadic sex are strongly affected by what one feels
one would or would not do.
Despite the theoretically relevant findings, the current study
has a number of potential limitations. First, in line with a number
of other studies based upon interdependence theory (see also Buunk
& Bakker, 1997), we used two-item measures to assess the
determinants of commitment. Although these variables were not the
major predictors, such measures may not most adequately sample the
domain to be measured. Second, the reliabilities of the investment
and alternative quality scales were rather low, cautioning against
drawing too strong conclusions on their limited predictive power.
Third, we assessed how individuals would in general respond to the
fact that their partner had had unprotected sex with someone else.
It is of course possible that the way in which individuals find out
about the partner's behavior (e.g., by learning through a friend, by
walking in on the act) will make a difference in the way they
respond. Fourth, we asked participants about their extradyadic
sexual experience in the past five years, and participants with a
relationship shorter than five years may have referred to
extradyadic sex that occurred during a previous relationship.
However, we find it hard to imagine that our findings would have
been different if we had applied a different criterion. Moreover,
past extradyadic experience was not considered a theoretically
relevant major predictor. An additional limitation is that the
correlational nature of the study cautions against drawing causal
conclusions, and it is unknown to what extent the responses to the
hypothetical situation of a partner's unprotected extradyadic sex
are valid predictors of responses to actual unprotected extradyadic
sex engaged in by one's partner. Finally, in a number of cases,
partners of participants participated in the study, resulting in
"yoked" pairs. Because anonymity was guaranteed, it was not possible
to assess who these pairs were. This is a problem with the current
research, and in future research this problem should be remedied.
Our study suggests that, as interdependence theory predicts, high
commitment is by no means a guarantee that partners will take
adequate precautionary measures when one of them has engaged in
unprotected extradyadic sex. Even more so, the strong tendency to
accommodate might cause one to overlook the risk involved in the
partner's behavior, such as pregnancy and STDs. The AIDS crisis may
have made it more necessary, but also more difficult, to communicate
with one's partner about one's extradyadic sexual involvements. It
is important that in future health educational messages the problems
of communicating in steady relationships about unprotected
extradyadic sex be given more attention. Messages could explicitly
stress that individuals who engage in unsafe sexual relationships
not only put themselves, but also their steady partners, at risk for
STDs, including HIV infection. In addition, it is important that
individuals who engage in unprotected, extradyadic sex be persuaded
to inform their partners about this behavior or to take preventive
action when they have intercourse with the steady partner after
unsafe sex with someone else. Health educational messages may point
at mutual responsibilities with regard to each other's health. Such
educational messages could play an important role in the prevention
of the further spread of HIV in the heterosexual population.
References
Ajzen, I. (1991). The theory of planned behavior. Organizational
Behavior and Human Decision Processes, 50, 179-211.
Bringle, R. G., & Buunk, B. P. (1991). Extradyadic
relationships and sexual jealousy. In K. McKinney & S. Sprecher
(Eds.), Sexuality in close relationships (pp. 135-154). Hillsdale,
NJ: Erlbaum.
Buss, D. M., Larsen, R. J., Westen, D., & Semmelroth, J.
(1992). Sex differences in jealousy: Evolution, physiology, and
psychology. Psychological Science, 3, 251-255.
Buunk, B. P. (1980). Extramarital sex in the Netherlands:
Motivations in social and marital context. Alternative Lifestyles,
3, 11-39.
Buunk, B. P. (1982). Anticipated sexual jealousy: Its
relationship to self-esteem, dependency and reciprocity. Personality
and Social Psychology Bulletin, 8, 310-316.
Buunk, B. P. (1990). Relational interaction satisfaction scale.
In J. Touliatos, B. F. Perlmutter, & M. A. Straus (Eds.),
Handbook of family measurement techniques (pp. 106-107). Newbury
Park, CA: Sage.
Buunk, B. P. (1995). Sex, self-esteem, dependency, and
extradyadic experience as related to jealousy responses. Journal of
Social and Personal Relationships, 12, 147-153.
Buunk, B. P., & Bakker, A. B. (1995). Extradyadic sex: The
role of descriptive and injunctive norms. The Journal of Sex
Research, 32, 313-318.
Buunk, B. P., & Bakker, A. B. (1997). Commitment to the
relationship, extradyadic sex, and AIDS-preventive behavior. Journal
of Applied Social Psychology, 23, 1241-1257.
Buunk, B., & Yzer, M. (in press). Predictors of
AIDS-preventive intentions among adult heterosexuals at risk for HIV
infection: Extending current models and measures. AIDS Education and
Prevention.
Buunk, B. P., & VanYperen, N. W. (1991). Referential
comparisons, relational comparisons, and exchange orientation: Their
relation to marital satisfaction. Personality and Social Psychology
Bulletin, 17, 709-717.
Daly, M., Wilson, M., & Weghorst, S. J. (1982). Male sexual
jealousy. Ethology and Sociobiology, 3, 11-27.
DeVellis, R. F. (1991). Scale development. Theory and
applications. Newbury Park, CA: Sage.
Emerson, R. M. (1981). Social exchange theory. In M. Rosenberg
& R. H. Turner (Eds.), Social psychology. Sociological
perspectives (pp. 30-65). New York: Basic Books.
Hatfield, E., Traupmann, J., Sprecher, S., Utne, M., & Hay,
J. (1985). Equity and intimate relations: Recent research. In W.
Ickes (Ed.), Compatible and incompatible relationships (pp. 1-27).
New York: Springer Verlag.
Kelley, H. H. (1979). Personal relationships. Their structures
and processes. Hillsdale, NJ: Erlbaum.
Kelley, H. H., & Thibaut, J. W (1978). Interpersonal
relations: A theory of interdependence. New York: Wiley.
Liberman, R. R, Wheeler, E. G., deVisser, L. A. J. M., Kuehnel,
J., & Kuehnel, T. (1980). Handbook of marital therapy. New York:
Plenum.
Prins, K S., Buunk, B. R, & VanYperen, N. W (1993). Equity,
normative disapproval and extramarital relationships. Journal of
Social and Personal Relationships, 10, 39-53.
Rusbult, C. E. (1980). Commitment and satisfaction in romantic
associations: A test of the investment model. Journal of
Experimental Social Psychology, 16, 172-186.
Rusbult, C. E. (1988). A longitudinal study test of the
investment model: The development (and deterioration) of
satisfaction and commitment in heterosexual involvements. Journal of
Personality and Social Psychology, 45, 101-117.
Rusbult, C. E., & Buunk, B. P. (1993). Commitment processes
in close relationships: An interdependence analysis. Journal of
Social and Personal Relationships, 10, 175-204.
Rusbult, C. E., Verette, J., Whitney, G. A., Slovik, L. E, &
Lipkus, I. (1991). Accommodation processes in close relationships:
Theory and preliminary empirical evidence. Journal of Personality
and Social Psychology, 60, 53-78.
Thibaut, J. W, & Kelley, H. H. (1959). The social Psychology
of groups. New York: Wiley.
VanYperen, N. W, & Buunk, B. P. (1994). Social comparison and
social exchange m marital relationships. In M. J. Lerner & G.
Mikula (Eds.), Entitlement and the affectional bond: Justice in
close relationships (pp. 89-115). New York: Plenum.
VanZessen, G., & Sandfort, T. (1991). Seksualiteit in
Nederland: Seksueel gedrag, risico en preventie van AIDS [Sexuality
in The Netherlands: Sexual behavior, risks and AIDS prevention].
Amsterdam: Swets & Zeitlinger.
Walster, L, Walster, G. W, & Berscheid, E. (1978). Equity:
Theory and research. Boston: Allyn and Bacon.
Portions of these data were presented at the 7th International
Conference on Personal Relationships, Groningen, The Netherlands,
July 4-8, 1994.
This research was financially supported by grant no. 91033 from
the Dutch Ministry of Welfare, Health and Cultural Affairs (WVC).
We thank Regina J. J. M. van den Eijnden and Frans W. Siero for
their contributions to this research project.
Address correspondence to Bram P. Buunk, Ph.D., Department of
Psychology, University of Groningen, Grote Kruisstraat 2/1, 9712 TS
Groningen, The Netherlands. Phone: 1-50-363-6360. Fax: 1-50-3636304.
E-mail: a.p.buunk@ppswrug.nl.