Arch Sex Behav DOI 10.1007/s10508-013-0200-3

ORIGINAL PAPER

Rebound Sex: Sexual Motives and Behaviors Following a Relationship Breakup Lindsay L. Barber • M. Lynne Cooper

Received: 21 July 2011 / Revised: 4 June 2013 / Accepted: 15 August 2013  Springer Science+Business Media New York 2013

Abstract The present study used a longitudinal, online diary method to examine trajectories of psychological recovery and sexual experience following a romantic relationship breakup among 170 undergraduate students. Consistent with popular beliefs about rebound and revenge sex, having sex to cope with distress and to get over or get back at the ex-partner were elevated immediately following the breakup and then declined over time, as did the probability of having sex with a new partner. Also consistent with popular lore, those who were ‘‘dumped’’ by their partners were more distressed and angry and more likely to have sex to cope and to get back at or get over their ex-partner. Finally, individuals who reported having sex to cope with negative feelings or to get over their ex-partner at the beginning of the study were more likely to have sex with a stranger and to continue having sex with new partners over time. Results were discussed in terms of widely held but largely untested beliefs about rebound and revenge sex. Keywords Romantic relationships  Breakup  Psychological recovery  Sex motives  Rebound sex  Revenge sex Introduction According to popular belief, people ‘‘on the rebound’’ are emotionally vulnerable due to increased feelings of distress, L. L. Barber  M. L. Cooper (&) Department of Psychological Sciences, University of Missouri, 124 Psychology Building, Columbia, MO 65211, USA e-mail: [email protected] L. L. Barber Ohio Department of Public Safety, Columbus, OH, USA

anger, and loss and diminished self-esteem that typically accompany loss of a relationship partner. Consequently, they may have sex to boost their self-esteem, to ease their pain and loneliness, or to‘‘get over’’the breakup, a phenomenon known as‘‘rebound sex.’’Alternatively, feelings of heightened distress and anger might instead (or also) cause one to have sex to ‘‘get back at’’the ex-partner, a phenomenon sometimes called ‘‘revenge sex.’’And finally, both rebound and revenge sex are thought to be more common among those who were ‘‘dumped’’ or were in more committed or longer-lasting relationships, presumably because relationship loss under such circumstances is more distressing. Despite the widespread nature of these beliefs, almost no scientific data exist on sexual experiences in the aftermath of a romanticrelationshipbreakup.Consider,forexample,thatentering the terms ‘‘rebound sex’’ or ‘‘revenge sex’’ in the PsycINFO search engine yielded 12 hits, of which only four were actually relevant whereas entering the same terms into the Google search engine yielded nearly 18,000,000 total hits. Although one cannot draw simple, straightforward conclusions based on these results about the relative quantity of scientifically valid information versus folk wisdom or speculation on these topics, the sheer magnitude of difference in the number of hits suggests that while beliefs about these phenomena abound, little actual evidence exists. To address this evidentiary gap, the present study examined feelings of distress, self-esteem, sexual motives (i.e., the self-reported reasons why people have sex), and sexual experiences following the breakup of a romantic relationship, using a weekly online diary among a sample of 170 college students who had recently broken up from a relationship. Common Beliefs About Being on the Rebound To gain insight into the nature of popular beliefs on rebound and revenge sex, the top websites listed under each of the

123

Arch Sex Behav

previously described Google searches were examined to determine how both phenomena are commonly defined and viewed. According to the Urban Dictionary, for example, a ‘‘rebound’’is‘‘someone you date or go out with…to keep your mind off your ex, for whom you still have feelings’’whereas rebound sex is defined as ‘‘the engagement of a sexual act while on the rebound.’’ Interestingly, the Urban Dictionary also advises men that ‘‘rebound day’’ (the day after Valentine’s Day) is an opportune time to ‘‘pick up recently single women in bars. These women are usually looking for someone for a quick night of forgettable sex to help them release the anger and pain of breaking up.’’ A similar definition was found on Yahoo! Answers:‘‘Rebound sex is when you’ve just gotten out of a relationship—typically a serious one, and you have sex with another person to either stick it to the one who dumped you or try to quiet your emotional hurt… or both!’’ (http://answers.yahoo.com/question/index?qid=200903061 62139AAFTs2y). Revenge sex is commonly defined as a ‘‘random, meaningless hook-up just to make the ex jealous’’ (see http://www. lemondrop.com/2010/03/31/revenge-sex/). According to a post on Huffpost Women (di Donato, September 30, 2012), revenge sex can be driven by other goals as well, including having sex to remind a former flame what he or she is missing; to express anger; to pay back someone who cheated on you by cheating on them; or to use someone else because you feel used. Thus, although some variability exists across websites, most definitions assume that instances of rebound or revenge sex are motivated by a desire or need to cope with the negative feelings thought to accompany the loss of a romantic partner. Most also identify distinct motivational antecedents for rebound and revenge sex. In the case of rebound sex, the motive is primarily to ease the pain of loss and to forget about or get over the ex-partner whereas anger and the desire to hurt or pay back the ex-partner are the primary motivators for revenge sex. Thus, both rebound and revenge sex are defined primarily by the motives they subserve and are, therefore, distinct from hook-ups, one-night stands, or serial short-term relationships to the extent that such sexual partnerships could be motivated by a range of different needs or desires including, for example, pleasure seeking or conquest. Common Psychological Consequences of Relationship Breakups As rebound and revenge sex are thought to occur in response to negative feelings experienced in the aftermath of a breakup, we first consider what is known about the most common psychological responses following a romantic relationship breakup. Both cross-sectional (Frazier & Cook, 1993; Sprecher, 1994) and prospective (Sbarra & Emery, 2005; Slotter, Gardner, & Finkel, 2010) research shows that heightened negative affect, including sadness, distress, and anger, are common responses to

123

relationship loss among both men and women (Simpson, 1990; Sprecher, 1994; Tashiro & Frazier, 2003) although intense levels of these affects may be relatively short-lived. In one of the only studies to chart emotional responses following a breakup over time, Sbarra and Emery (2005) found that individuals who had broken up from a dating relationship reported greater sadness than participants in intact dating relationships at 0–2 weeks post-breakup, but not at 4–6 weeks. In contrast, heightened anger persisted until their last follow-up at 6 weeks post-breakup. Less is known, however, about the effects of relationship loss on feelings about the self. Although several studies have examined how personal characteristics of the individual or of the breakup itself are associated with feelings about the self (e.g., Davis, Shaver, & Vernon, 2003; Frazier & Cook, 1993), we found only one study that directly assessed the impact of a recent breakup on an aspect of the self. Slotter et al. (2010) followed a group of students, all of whom were in a relationship at the start of the study, and found that those who experienced a breakup in the prior week showed a significant drop in self-concept clarity relative to their pre-breakup levels. However, we found no study that directly assessed the impact of relationship loss on self-esteem per se. Finally, past research shows that people who were‘‘dumped’’ or were in more committed, higher quality, or longer-lasting relationships typically experienced greater distress and lower self-esteem (Davis et al., 2003; Fine & Sacher, 1997; Frazier & Cook, 1993; Helgeson, 1994; Perilloux & Buss, 2008; Sprecher, 1994). Such findings suggest that we should see higher levels of sex to manage negative mood states, to repair or boost one’s self-esteem, and to get over or get back at one’s ex-partner when these circumstances apply (for contrary findings, see Bancroft et al., 2003). Past Research on Sexual Behavior in the Aftermath of a Breakup As previously discussed, although there appears to be a good deal of consensus on the existence and nature of rebound sex, only a small number of studies provide data that are even indirectly relevant to judging the accuracy of these beliefs. Several studies (e.g., Anderson et al., 2004; Leigh, Temple, & Trocki, 1993; Smith, 1991) have shown that the overwhelming majority (up to 90 % in one national study; Leigh et al., 1993) of individuals who described themselves as currently divorced or separated had had at least one sex partner in the past year and a substantial minority (33 % in the Leigh et al. study) had had two or more partners. However, none of these studies ascertained whether the sexual experiences even occurred after the breakup, much less whether the participants considered the experiences instances of rebound or revenge sex. One study, however, did examine sexual experiences that occurred in the aftermath of a recent breakup (Wade &

Arch Sex Behav

DeLamater, 2002). Using data from the 1992 National Health and Social Life Survey, 536 individuals whose most recent breakup involved either a cohabiting or marital relationship retrospectively reported the number of sex partners since the breakup or, for those who broke up more than 1 year ago, the number in the past year. Relative to a comparison group of never-married individuals, those who had broken up within the past year had significantly more new sex partners whereas those who broke up more than a year ago did not differ from the comparison group. From this, it was inferred that acquiring a new sexual partner(s) might aid adjustment among recently separated individuals by fulfilling intimacy or revenge needs or by enhancing self-esteem. Although Wade and DeLamater’s (2002) data were consistent with popular notions of rebound sex, they were limited in important ways. First, participants were not asked about their motives for sex nor did they disclose how long after the breakup the sexual experiences occurred. Second, the three groups may have differed on a variety of dimensions that could have accounted for the observed differences in sexual experience. For example, differences between the recent breakup and nevermarried comparison groups might reflect social deficits among the never-married group (e.g., DiTommaso, Brannen-McNulty, Ross, & Burgess, 2003) rather than excess sexual activity in the recent breakup group. Similarly, the lack of differences between the remote breakup group (which included partnered individuals) and the never-married comparison group could also reflect social deficits among the never-married individuals given that single individuals would ordinarily have more sex partners than their partnered counterparts (Santelli, Brener, Lowry, Bhatt, & Zabin, 1998). Finally, this study relied exclusively on a crosssectional design and retrospective recall of past sexual experiences. Past Research on the Use of Sex for Instrumental Reasons Consistent with rebound lore, research on motives for sex documents that people do self-report using sex for instrumental reasons, including the use of sex to cope with distress and to boost or repair self-esteem (Cooper, Shapiro, & Powers, 1998; Gebhardt, Kuyper, & Greunsven, 2003; Grossbard, Lee, Neighbors,Hendershot,&Larimer,2007;Meston&Buss,2007).Atthe same time, studies consistently show that coping and self-esteem enhancement motives are relatively uncommon motives for sex, at least as compared to motives like increased intimacy and physical pleasure (e.g., Cooper et al., 1998; Patrick, Maggs, Cooper, & Lee, 2011). Both rebound and revenge motives were identified in one study of college students who were asked to list ‘‘all of the reasons they could think of’’ that either they or someone they know had had sexual intercourse in the past (Meston & Buss, 2007). Revenge-related reasons such as ‘‘I wanted to get even with someone,’’‘‘I wanted to make someone else jealous,’’and‘‘I

wanted to even the score with a cheating partner’’were all listed, but were among the 50 least common reasons given (out of 237 total reasons). Rebound-related reasons, such as ‘‘I was on the rebound from another relationship’’ and ‘‘I was trying to ‘get over’ an earlier person/relationship,’’ were also identified, but did not appear in either the top or bottom 50 lists, suggesting that their occurrence fell somewhere in between the two extremes. Finally, ABC News (2004) conducted a random sample phone survey of 1,501 U.S. adults (response rate = 92 %) on a wide range of sexual experiences. Approximately 20 % of those who responded to the survey reported having had rebound sex (defined as‘‘sex with someone just to help you get over a failed relationship’’) and 10 % reported having had revenge sex (defined as‘‘sex with someone just to get back at someone else’’), with nearly equal percentages of men and women endorsing both behaviors. Rates were even higher (33 and 16 %, respectively) among young adults, aged 18–29 years, indicating that a significant minority of American adults, especially young adults, reported having had at least one experience of rebound or revenge sex. Together, these data suggest that people do have sex to cope with negative emotions and boost their self-esteem, as well as more specifically for rebound and revenge motives. Although all of these motives appear to be relatively uncommon motivations for sex, none of these studies focused on individuals who had experienced a recent breakup for whom such motives might be both more relevant and more common. Overview of the Current Study The present study examined trajectories of recovery from loss of a romantic relationship over 8 months, with a particular focus on testing widely held beliefs about rebound and revenge sex. A sample of 170 college students who experienced a recent breakup completed longitudinal, online diaries over the course of a semester. Hypotheses regarding patterns of change and predictors (including relationship and breakup characteristics) of these patterns were tested for three broad categories of outcomes: emotional distress and feelings of self-worth, motives for sex, and sexual partners. Hypothesis 1: Patterns of Change in Negative Feelings and Self-Esteem We expected to replicate and extend Sbarra and Emery’s (2005) findings on the time course of emotional recovery from a breakup. Specifically, distress over the breakup and anger toward the ex-partner should decline over time following the breakup. However, because our study extended well beyond the 6 weeks post-breakup studied by Sbarra and Emery, these declines should eventually level off. Conversely, feelings of self-worth should show the opposite pattern—that is, increase and then level off.

123

Arch Sex Behav

Hypothesis 2: Patterns of Change in Motives for Sex To the extent that sex is used as an antidote to dysphoria, loneliness, or low self-esteem, having sex to cope with distress, to boost one’s sense of self-worth, or to get over or get back at one’s ex-partner should decline over time and eventually level off, much like the expected trajectories for feelings of distress and low self-worth. Hypothesis 3: Patterns of Change in Sexual Partners To the extent that sex is used to cope with the psychological aftermath of a breakup, patterns of change in actual sexual experience should mirror those for motives. However, these patterns may be difficult to discern because actual sexual behavior depends on a variety of factors, including the presence of a willing partner. Thus, we examined trajectories of sexual behavior to determine whether such patterns emerged. Hypothesis 4: Relationship and Breakup Characteristics as Predictors of Patterns of Recovery Based on prior research, adverse emotional responses should be intensified and recovery slowed among individuals who were ‘‘dumped’’ or were in a highly committed or longerlasting relationship. Similar patterns should emerge between relationship/breakup characteristics and trajectories of sex motives and actual sexual experiences though, as previously discussed, the behavioral patterns may be weak due to the dyadic nature of sexual behavior. Hypothesis 5: Psychological and Motivational Predictors of Sexual Behaviors Following a Breakup Finally, to the extent that actual sexual encounters are used instrumentally as a way to get over or get back at one’s expartner, the highest initial levels of sexual involvement and the slowest declines in these behaviors should be seen among those who experience the greatest distress and the lowest levels of selfesteem, as well as among those who strongly endorse the use of sex to deal with or manage these negative feeling states.

Method Participants Participants were recruited from the introductory psychology research pool at a large midwestern university. To participate, students must have (1) been at least 18 years old; (2) self-identified as heterosexual; (3) had sex at least once in their lifetime; (4) broken up from a romantic relationship in the past 12 months; and (5) been single (i.e., not in a relationship) at the outset of the

123

study. Twenty-six participants got back with their ex-partner during the first 3 weeks of the study and were excluded from the sample. Five additional participants got back with their partner later in the study; weekly reports from these individuals were retained up to and through the week prior to getting back with the ex-partner. This resulted in a final sample of 170 participants who contributed a total of 1,673 weekly distress reports and 1,503 weekly self-esteem and sex reports. The final sample was 66 % female and 88 % Caucasian; 89 % were freshman. On average, participants were 18.3 years old; their most recent relationship lasted 13.5 months (range, 1– 42 months), and the breakup occurred, on average, 12.9 weeks prior to the initial survey (range, \1–32 weeks ago). Gender comparisons revealed that males did not differ significantly from females on relationship duration, time since breakup, or who initiated the breakup. However, women were more committed than men to the lost relationship (3.3 vs. 2.8 on a 1–5 scale), t(168) = 2.16, p = .03. All analyses were based on the full sample of individuals and weekly reports, except for analyses of sexual motives and sexual behavior which were based on smaller subsets, as described more fully below. Procedure Questions regarding the experience of a recent breakup were included in the mass screening survey completed by all introductory psychology students during the first week of the fall semester. Students meeting our eligibility criteria were sent an initial e-mail informing them of their eligibility to participate in a study on‘‘the process of recovering from a romantic relationship breakup’’in exchange for course research credit and a chance to earn cash ($3 foreach weekly survey completed after thefirst 10) and win a $200 cash prize. Follow-up phone calls to determine willingness to participate were then made to those who did not spontaneously respond to the initial e-mail. Calls were made according to the recency of the breakup such that those who had most recently broken up were contacted first. Enrollment into the study was discontinued during week 4 of the semester, allowing all participants the opportunity to contribute at least 10 weekly reports. Participants who agreed to participate and met eligibility criteria (described above) completed three phases of the study—an initial session, a 10–12 week online diary, and an exit interview. As described, participants were compensated with a combination of research credits, money, and lottery prize tickets, with total compensation increasing as the percentage of possible surveys completed increased. On average, participants completed 82 % of weekly reports during the diary phase. The percent of reports completed by each participant was uncorrelated with gender, race, age, who initiated the breakup, relationship commitment, relationship duration, initial feelings toward the ex-partner, and initial self-esteem. However, participants who broke up longer

Arch Sex Behav

ago completed fewer reports, r(168) = -.44, p\.001, possibly indicating that the study held less interest for them. In the initial session, the study was described, informed consent was obtained, and a brief baseline survey was completed. During the diary phase, participants were sent an email on Wednesday afternoons containing a link to the weekly survey, which was to be completed by Thursday evening, after which the survey was closed. Weekly surveys queried participants about their recent mood, self-esteem, and sexual experiences. Those who had had sex in the past week completed questions about their sexual experiences with up to two different partners. To avoid ‘‘training’’ participants to deny sexual experience (cf. Kessler, 1995), participants who had not had sex in the past week completed an alternate set of questions (of roughly the same length) about sexual desire, masturbation, etc. Consequently, the weekly survey took about 20 min regardless of the participant’s sexual experiences in the past week. At the end of the diary phase, participants completed a short questionnaire regarding their participation, were debriefed, and compensated.

Table 1 Descriptive statistics for Level 2 covariate and predictor variables and Level 1 outcome variables Variable

Ma

SD

Reliability

ICC

Initial survey (Level 2 predictors) Gender Time since breakup (weeks)

.34

.47

na

na

12.88

8.25

na

na

Partner initiation Relationship duration (months)

2.56

1.11

.75

na

13.37

11.23

na

na

3.13

1.40

.88

na

Commitment Weekly surveys (Level 1 outcomes) Feelings toward the ex-partner Distress

2.57

1.33

.93

.60

Anger

1.38

0.66

.86

.36

Appear self-esteem

3.58

1.06

.89

.77

Social self-esteem

3.30

1.04

.86

.64

Self-esteem

Sex motives

Measures

Coping motives

1.92

0.99

.73

.51

Self-affirm motives

2.17

1.39

.89

.66

Rebound motives

1.56

1.12

.79

.59

Revenge motives

1.25

0.59

.79

.44

p. new partner

.09

.10

na

.004

p. stranger sex

.02

.06

na

.02

Sexual partners

Table 1 shows descriptive information and internal consistency estimates of reliability for all variables and intraclass correlations (ICCs) for weekly variables. All items were scored so that higher scores equal more of the measured construct. Relationship and Breakup Characteristics Four characteristics of the relationship and the breakup were assessed before the online phase of the study. Time since breakup (in weeks) and relationship duration (in months) were both assessed by single items. Responsibility for initiating the breakup was assessed by two items indicating the extent to which the expartner initiated and wanted the breakup (e.g.,‘‘While you were still in the relationship, how much did you want to breakup compared to how much your ex-partner wanted to end the relationship?’’). Items were rated on a 1 (It was completely my idea/I wanted the breakup much more than my ex-partner did) to 5 (It was completely my ex-partner’s idea/My ex-partner wanted the breakup much more than I did) scale. Relationship commitment was assessed by three items indexing the extent to which the individual was more involved, more committed, and viewed the relationship as more important than the ex-partner (e.g.,‘‘Which one of you, if either, was more involved in the relationship?’’). All items were rated on a 1 (My ex-partner much more than me) to 5 (I much more than my ex-partner) scale. The mid-point of both the commitment and initiation scales denoted equality between partners.1 Examination of the correlations among predictors 1

We tested for nonlinear effects of both scales given that equality between partners (the mid-point of our scale) might not lay equidistant

p probability Gender is coded 0 = women, 1 = men. Number of weekly reports (n) for feelings about ex-partner range from 1,652 to 1,663; ns for self-esteem range from 1,496 to 1,498; ns for sex motives range from 311 to 313; and ns for sexual partners range from 1,105 to 1,121 a

Mean values for dependent variables reflect the average within-person mean collapsed across all weeks of participation

revealed that time since breakup was unrelated to the other three variables, rs(168)\-.10, ps[.19, thus indicating that perceptions of the relationship and the breakup were not biased (e.g., seen as increasingly less serious) as a function of elapsed time since the breakup. In addition, and as expected, commitment and (partner)initiationweremoderatelypositivelycorrelated,r(168) = .47, p\.001, whereas relationship duration was weakly negatively related to (partner) initiation, r(168) = -.13, p = .10, and commitment, r(168) = -.18, p = .02. Footnote 1 continued between the two poles, as a linear model assumes. Consistent with this possibility, several non-linear effects were found for both commitment and partner initiation. Plotting the non-linear interactions revealed that findings for those who chose the mid-point of the scale were more similar to those observed for individuals who were more committed to the relationship and did not want the breakup than to findings observed for those whose partner was more committed and did not want the breakup. Although such findings do not alter the basic conclusions drawn from the present study, they do substantially increase its complexity. For these reasons, we elected not to include the results of the nonlinear tests of commitment and initiation effects.

123

Arch Sex Behav

Negative Feelings in the Aftermath of the Breakup

Self-Esteem

Two different feelings toward the ex-partner and lost relationship were assessed in both the initial and weekly surveys: (1) distress about the breakup and love for the ex-partner (6 items); and (2) anger and desire for revenge (2 items). Sample items included: ‘‘How emotionally attached did you feel to your expartner?’’ (distress), ‘‘How distressed were you about the breakup?’’(distress), and‘‘How much did you wish bad things on your ex-partner?’’(anger). Initially, participants indicated the extent to which they had experienced each feeling in the past few weeks (or for those who had broken up more recently, since the breakup) whereas, in the weekly surveys, participants rated the experience of each feeling in the past week. Responses were provided on a 7-point scale, where 1 = Not at all and 7 = Extremely/A great deal. An exploratory factor analysis of data from the initial survey supported the two factor structure with all items loading on their respective factors C 0.50; cross loadings on secondary factor were B 0.19. As shown in Table 1 (bottom panel), both measures were moderately to highly reliable, but exhibited varying amounts of stable between-person variance (as indicated by the ICCs), with feelings of distress showing substantially more stability and less variability from week to week than anger. Mean levels of endorsement collapsed across the diary period fell below the midpoint of the scale on both measures, but were especially low for anger (1.38 on a 1–7 scale). For this reason, anger was dichotomized for analytic purposes, such that 0 = no anger, 1 = any feelings of anger in the past week.2

Social and appearance-related self-esteem were measured in the weekly reports by Heatherton and Polivy’s (1991) state selfesteem measure. Subscales were computed according to published guidelines, except for two items that were excluded from the social scale because they failed to load on the intended factor (loadings B 0.06). Appearance self-esteem (four items) assessed the extent to which the participant felt attractive and was satisfied with his/her appearance and the way his/her body looked (e.g., ‘‘I am pleased with my appearance right now’’). Social self-esteem (3 items) assessed the extent to which the participant felt confident about how others viewed or felt about him/her (e.g., ‘‘I feel concerned about the impression I am making’’ [reverse scored]). Participants rated how much they agreed with each statement on a 1 (strongly disagree) to 6 (strongly agree) scale, based on their feelings in the past day. As shown in Table 1, subscale alphas were acceptably high. ICCs were also high, indicating that self-esteem was more stable than variable from week to week.

2

Three outcomes were dichotomized in the present study due to low base rates: anger, rebound sex, and revenge sex. To determine whether our results were affected by dichotomizing these three outcomes, analyses reported in Tables 2 and 3 for these outcomes were re-estimated using a log transformed version of each variable. Results were identical in terms of sign (i.e., either both positive or both negative) and significance (i.e., either both significant or both non-significant) across the two analyses for all three normative trajectories reported in Table 2. Results were likewise identical in terms of sign and significance for the predictors of variability in the rebound sex trajectory reported in Table 3. Slight differences between the dichotomous and transformed continuous outcomes were found, however, for anger and revenge sex. Specifically, the significant effects of partner initiation on anger and revenge sex reported in Table 3 were weaker and non-significant (though still the same sign) using the transformed continuous measures, whereas the corresponding non-significant effects reported in Table 3 for commitment were stronger and significant. Thus, there is instability across alternative operationalizations of the outcome measures in which effect emerges (partner initiation or commitment) in the multivariate context. However, given that both partner initiation and commitment are significant positive predictors of anger and revenge motives when estimated independently of one another, and that partner initiation and commitment are positively correlated (r = .47), we believe that both patterns of results can be seen as supporting the same general conclusion: Consistent with rebound lore, those who were more psychologically invested in the lost relationship (whether indexed by a measure of psychological commitment or the fact that the partner initiated the

123

Motives for Sexual Behavior Participants described their reasons for having sex with up to two different partners in the last week. Having sex to cope with distress (3 items) and to reassure oneself of his/her desirability or boost self-esteem (2 items) were assessed by a composite of the highest loading items on each subscale from Cooper et al.’s (1998) sex motives measure. Representative items included having sex…‘‘to cheer up or to feel better’’(coping motives) and ‘‘to reassure yourself of your desirability’’ (self-affirmation motives). Four additional items (2 per motive construct) were created to assess revenge (e.g.,‘‘to ‘get back’ at your expartner’’) and rebound (e.g., ‘‘to help you ‘get over’ your expartner and the breakup’’) motives. Revenge and rebound motives were administered only when the sex partner was not the expartner. For each item, participants indicated on a 7-point scale (1 = Not at all, 7 = Extremely/A great deal), the degree to which they had sex with each partner for that reason in the past week. Factor analysis of the sex motive items (averaged across weekly reports) revealed that all items loaded on their intended factors, including the four newly created items which loaded cleanly on separate rebound and revenge sex factors (i.e., primary loadings C 0.81, secondary loadings B 0.23). As shown in Table 1, all motive scales evidenced adequate reliability, particularly given the shortened format. Moreover, all motives exhibited moderate levels of stability within a person Footnote 2 continued breakup) experienced more anger and stronger revenge sex motives. In the end, we opted to report the results of the dichotomous analyses because of their greater interpretability. Details of the analyses of the transformed continuous outcomes are available from the second author on request.

Arch Sex Behav

across weeks. Average rates of endorsement (collapsed across weeks) were low for all motives. Although as previously described, this pattern of endorsement was consistent with normative data from other studies, scores on the most skewed motives (rebound and revenge motives) were dichotomized (none vs. any) for analytic purposes (see footnote 2).

center time since breakup at 4 weeks); (2) a linear growth component, assessing linear change from week 4; and (3) a quadratic component, assessing the eventual slowing or leveling off of change over time.

Results Sexual Partners Preliminary Analyses A total of 113 of the 170 participants had sex one or more times over the course of the study and together contributed 390 weekly sex partner reports: 20 % of these events involved the ex-partner, 26 % a new or first-time partner, and 54 % a previous (but non-ex) partner. Roughly 20 % of the first-time partners were strangers, defined as someone met that same day or evening. Because rebound lore focuses primarily on sex with new or casual partners, we created two dichotomous measures (0 = no, 1 = yes) for use in our primary trajectory analyses, indicating whether, in a given week, the individual had sex with a new partner (i.e., someone with whom the participant had never had oral sex or sexual intercourse) or with a stranger, defined as someone the participant had ‘‘just met.’’ Overview of Analyses Hierarchical linear modeling (HLM version 6.08) (Raudenbush, Bryk, Cheong, & Congdon, 2004) was used to characterize patterns of change in weekly reports of distress, self-esteem, sex motives, and sexual behavior across the 10–12 week period. To account for dependencies among the weekly reports, they were modeled at Level 1 (L1) and nested under individuals (modeled at Level 2; L2). Patterns of change in feelings and behaviors were analyzed as a function of time since the breakup, centered at 4 weeks, which provided a more stable solution than centering at 3 or fewer weeks. This approach, which correctly accounts for heterogeneity in time since breakup at the initial assessment (Mehta & West, 2000), is analogous to growth curve analyses of data from a cohort-sequential (or accelerated) design. In such designs, adjacent segments of repeated data from different cohorts (in the present case, subsets of individuals who broke up with their partners differing lengths of time ago) were used to estimate a common developmental trend or growth curve over a longer period of time than any single individual or groups of individuals actually participated. For example, in the present study, we had subsets of individuals who contributed data from 1 to 12 weeks post-breakup and others who contributed data from 23 weeks to 34 weeks post-breakup. Nevertheless, although no one contributed data across the full 8 months, this method of analysis allowed us to estimate trajectories of growth over the entire 8-month period. Finally, each growth curve consisted of three terms: (1) an intercept or average level of the outcome estimated at 4 weeks post-breakup (reflecting our decision to

Identifying Significant Covariates for the Base Growth Curve Models A set of base models was estimated to identify the relevant control variables for inclusion in our primary analyses. Whether the survey was completed in the morning or afternoon and Thanksgiving break were treated as time varying covariates and modeled at L1. The former (coded a.m. vs. p.m.) was included to control for diurnal mood effects (Watts, Cox, & Robson, 1983) and the latter to control for possible consequences of seeing or talking with the ex-partner or other changes in routine which might have occurred over Thanksgiving break. Gender (0 = female, 1 = male), freshman (vs. more advanced) standing, and week of the semester in which the first weekly report was filed (i.e., Week 4, 5, or 6) were also modeled at L2. Freshman standing and start week were controlled to help rule out adjustment-to-school effects that typically diminish over the course of the first semester of college (Gall, Evans, & Bellerose, 2000). Whether each participant had sex during the week was also controlled in models predicting mood and self-esteem, thus allowing for estimation of normative patterns of change over time, net of any influence that having sex might have on mood. Initial analyses treated each component of the growth curve (i.e., the intercept, linear term, and quadratic term) as the dependent measure and included the full set of L1 and L2 covariates as predictors. Following recommendations by Bryk and Raudenbush (1992), trimmed models, which have been shown to yield more stable parameter estimates, were developed for each dependent measure. Summary of Covariate Results Although we offered no a priori hypotheses regarding the effects of the covariates on patterns of recovery, we were particularly interested in whether men and women responded similarly to loss of a romantic relationship. Examination of gender effects in these analyses revealed only one significant effect of 30 tested (10 outcomes X 3 growth parameters): Men relative to women reported higher levels of appearance selfesteem (intercept = 0.81, p\.001) at four weeks postbreakup. Based on these analyses, it appears that men and women responded in a similar manner to the loss of a romantic relationship.

123

Arch Sex Behav Table 2 Results of trimmed base growth curve models describing change over time since the breakup in outcomes Outcome

Growth parameters Intercept

Linear

Quadratic

4.11***

-0.17***

0.004***

0.31*** a

-0.04**



Feelings toward the ex-partner Distress p. anger Self-esteem Appear self-esteem

3.57***





Social self-esteem

3.03***

0.02**



Sex motives -0.02 

Coping motives

2.37***

Self-affirm motives

2.18***

p. rebound motives

0.35 a

-0.06**



p. revenge motives

0.23** a

-0.06*



-0.11*

0.003 



– –

Sexual partners p. new partner

0.20*** a

p. stranger sex

a

0.03***





Base model includes the intercept (centered at 4 weeks post-breakup), linearbreakup timeago,and quadraticbreakup timeago (i.e.,breakup time ago-squared). Results are unstandardized regression coefficients from trimmed base models in which non-significant (i.e., p[.10) higher-order terms were dropped from the model (denoted as‘‘–’’in the table) a

For dichotomous variables, intercept values are predicted probabilities (denoted as‘‘p’’) of endorsement calculated from the unstandardized logistic regression coefficient

 

p\.10; * p\.05; ** p\.01; *** p\.001

Hypotheses 1–3: Normative Patterns of Recovery Following a Breakup Our initial analyses describe normative patterns of change in three broad classes of outcomes (emotional upset and selfesteem, motives for sex, sexual partners) over an 8-month period following the breakup of a romantic relationship. A total of 10 multilevel models were estimated, one for each of the outcomes examined in the present study. Results for the final trimmed models (dropping non-significant terms) are shown in Table 2. The intercept or predicted values of the outcome at Week 4 are shown in the first column and linear and quadratic changes in the outcome are shown in the second and third columns, respectively. Hypothesis 1: Patterns of Change in Negative Feelings and Self-Esteem As shown in Table 2 (second column) and as predicted, both distress and the likelihood of experiencing anger declined significantly across time. Also as predicted, a significant quadratic effect was found for distress, indicating that the decrease in distress decelerated or ‘‘leveled off’’ over time, reaching its

123

Fig. 1 Top panel base model for breakup time ago predicting distress. Bottom panel base model for breakup time ago predicting probability of having a new sex partner

lowest level around 25–28 weeks post-breakup. This trajectory is illustrated in the top panel of Fig. 1. Effects for self-esteem are shown in the second panel of Table 2. As expected, levels of social self-esteem increased over time, though contrary to expectation, appearance selfesteem did not change. Also contrary to expectation, there were no quadratic effects for either self-esteem measure, indicating that the increase in social self-esteem did not level off during the study period. Hypothesis 2: Patterns of Change in Motives for Sex As shown in the third panel of Table 2, from one-fourth to one-third of participants reported having sex to get back at (revenge) or get over (rebound) the ex-partner, to at least some extent, at 4 weeks post-breakup. Moreover, rebound, revenge, and coping motives showed significant declines over time, as expected, although they did not level off (as indicated by the absence of a significant quadratic component). Also contrary to expectation, self-affirmation motives did not decline over time.

Arch Sex Behav Table 3 Results of trimmed growth curve models describing change in outcomes over time since the breakup as a function of relationship and breakup characteristics Outcome

Growth parameters Intercept

Linear

Quadratic

Partner initiation Distress

0.63***

p. anger

0.40**

-0.02*

Social self-esteem

0.20

-0.03*

0.001*

Coping motives

0.96**

-0.18***

0.01***

Self-affirm motives

0.94**

-0.18***

0.01**

p. rebound motives

1.08**

-0.06**



p. revenge motives

2.46**

-0.36**

0.01**



– –

Relationship duration Distress

0.03***

Appear self-esteem

0.02 

Social self-esteem

0.01*

p. revenge motives

0.11*

Commitment Coping motives

– -0.001* – -0.02**

-0.05

0.06*

Self-affirm motives

-0.20

0.09**

p. new partner

-0.41 

p. stranger sex

-0.16

– – – 0.001*** -0.002* -0.003**

-0.08*

-0.003*

0.10 

-0.004 

p predicted probability Non-significant lower-order terms were included in the model as needed to provide valid tests of significant higher-order terms. However, nonsignificant higher-order terms were dropped from the final trimmed models, and are denoted as ‘‘–’’ in the table. Tabled values are unstandardized regression coefficients  

p\.10; * p\.05; ** p\.01; *** p\.001

Hypothesis 3: Patterns of Change in Sexual Partners As shown in the bottom panel of Table 2, nearly 20 % of participants reported having sex with a new partner at 4 weeks postbreakup. Moreover, as shown in the bottom panel of Fig. 1, this probability declined and then leveled off, as expected, starting at about 13–16 weeks post-breakup. Finally, only about 3 % of participants reported having sex with a stranger at week 4, and, contrary to expectation, this probability did not decline over time, perhaps due to its low initial base rate. Hypothesis 4: Relationship and Breakup Characteristics as Predictors of Patterns of Recovery To determine whether the course of recovery from the breakup differed systematically as a function of relationship or breakup characteristics, a series of 10 multilevel models were estimated, one for each outcome. For each outcome, the intercept, linear, and quadratic components of the growth curve were regressed

on the set of three predictors (partner initiation, relationship duration, and commitment). Relationship/breakup predictors (as fixed attributes of the individual’s past experience) were modeled at L2. All models controlled for significant L1 and L2 covariates, as described above. Effects on the intercept tested whether variation in each of the predictors was uniquely associated with average levels of the outcome at week 4 (e.g., whether individuals whose relationships lasted longer were more upset on average at 4 weeks postbreakup). Significant effects on the linear and quadratic components indicated the presence of cross-level interactions, which signify that the relevant aspect of change over time varied reliably as a function of the L2 predictor. For example, a significant effect for partner initiation on the linear component of revenge sex motives would indicate that the rate of linear change in revenge motives increased (for a positive coefficient) or decreased (for a negative coefficient) with increasing levels of the partner initiation variable. Following recommendations of Bryk and Raudenbush (1992), full models including all higher-order terms were initially tested. Non-significant higher-order terms were then dropped to develop trimmed models in which only significant effects or lower-order effects required to provide valid tests of significant higher-order effects were retained. Finally, when patterns of change were found to differ significantly as a function of either relationship or breakup characteristics, separate growth curves were estimated for subgroups that were low or high on the predictor to determine whether the observed patterns of change over time conformed to expectation. For example, in models where partner initiation predicted either the linear or quadratic component of the growth curve, separate growth curves were estimated for those who wanted the breakup more than the ex-partner (the 15th percentile on partner initiation) as well as for those whose ex-partner wanted the breakup more (the 85th percentile). Results for the trimmed models are shown in Table 3. Hypothesis 4a: Effects of Partner Initiation on Recovery Patterns As shown in Table 3 (top panel, intercept column), individuals whose partners initiated the breakup reported stronger feelings of distress and greater anger at 4 weeks post-breakup, as expected, though they did not experience lower levels of self-esteem. In addition, and also as expected, they reported significantly higher levels of coping, self-affirmation, rebound, and revenge motives for sex. As shown in the second (linear) and third (quadratic) columns of Table 3, there was a consistent pattern of cross-level interactions involving partner initiation. Plotting these interactions showed in all but one case (predicting social self-esteem) that those who were‘‘dumped’’started out higher but recovered more quickly such that their levels converged with those who were not

123

Arch Sex Behav

Fig. 2 Top panel interaction of partner initiation X breakup time ago predicting feelings of distress. Bottom panel interaction of partner initiation X breakup time ago predicting probability of having rebound sex

‘‘dumped’’by about 22 weeks post-breakup. Two of these interactions are shown in Fig. 2. Hypothesis 4b: Effects of Relationship Duration on Recovery Patterns Overall,onlyafeweffectswereobservedforrelationshipduration and they were neither robust across outcomes nor uniformly consistent with expectation (see second panel of Table 3). On the one hand, individuals from longer (vs. shorter) relationships reported greater distress, as expected, at 4 weeks post-breakup, but they also reported higher social self-esteem and marginally higher appearance self-esteem, perhaps reflecting the fact that those who were able to maintain longer relationships were generally more competent. In addition, two trajectories (revenge sex, appearance selfesteem) differed significantly as a function of varying relationship lengths. Examining these interactions again revealed one pattern that was generally consistent with rebound lore and one that was not. As shown in the top panel of Fig. 3, individuals from longerlasting relationships reported significantly elevated levels of

123

Fig. 3 Top panel interaction of relationship duration X breakup time ago predicting probability of having revenge sex. Bottom panel interaction of relationship commitment X breakup time ago predicting coping motives for sex

revenge sex in the immediate aftermath of the breakup, which then declined steeply and leveled off around 13 weeks. In contrast, those from shorter relationships showed a different pattern of change over time but, more importantly, reported substantially lower levels of revenge sex in the immediate aftermath of the breakup. Contrary to expectation based on rebound lore, however, trends over time in appearance selfesteem showed a cross-over pattern in which self-esteem declined slightly among those from longer lasting relationships, but increased significantly among those from shorter relationships. Hypothesis 4c: Effects of Relationship Commitment on Recovery Patterns As shown in Table 3 (third panel), only one intercept effect was found for commitment and it did not conform to the expected pattern: Those who were less (not more) committed to their relationship were more likely to have a new sex partner at 4 weeks post-breakup. However, commitment also interacted with time since breakup to explain change in coping and selfaffirmation motives. Plotting separate trajectories for individuals

Arch Sex Behav

low and high in commitment revealed similar patterns for these two sex motives. As shown in Fig. 3 (bottom panel), those who were less committed to their prior relationship reported steeper declines in coping motives over time, whereas those who were more committed showed an increase followed by a gradual decline starting around 16 weeks post-breakup. Finally, commitment also interacted with time since breakup to predict change in the probabilities of having sex with a new partner or a stranger. Plotting these interactions (not shown) revealed that less committed participants were more (not less) likely to have sex with a new partner at 4 weeks post-breakup, as previously described, after which time the probability declined and then leveled off. Among those who were more committed, the probability of sex with a new partner increased up to about week 15 and then declined. The pattern for stranger sex was similar but weaker. Hypothesis 5: Psychological and Motivational Predictors of Sexual Behaviors Following a Breakup The final series of analyses examined whether feelings of distress and anger related to the ex-partner and the breakup, low self-esteem, or the use of sex to manage these states predicted the likelihood of having a new sex partner or sex with a stranger. Following our earlier analytic strategy, we regressed the intercept, linear, and quadratic components of the growth curve for new partners and stranger sex on the average level (aggregated across the diary period) of negative feelings, self-esteem, or motives, modeled at L2. As before, significant covariates were also included. The intercept effect in the L2 model tested between-person differences in the outcome (estimated at week 4) as a function of between-person differences in the average level of negative feelings, self-esteem, or motives whereas the L2 effects on the linear and quadratic components tested whether sex partner trajectories varied across individuals who differed in average levels of each predictor. Results revealed no intercept effects on the probability of having sex with a new partner. Thus, people who differed in average levels of distress and anger, self-esteem, and motives were no more or less likely to have had sex with a new partner at 4 weeks post-breakup. Likewise, there were no intercept effects on the probability of having sex with a stranger as a function of average differences in negative feelings or self-esteem. However, there was a consistent pattern of motive effects: Those who were high on rebound (intercept = 1.17, p = .01), revenge (intercept = 1.53, p = .01), coping (intercept = 0.36, p = .04), and affirmation (intercept = 0.35, p = .01) motives were significantly more likely to have had sex with a stranger at 4 weeks post-breakup than were their low motive counterparts. In addition, three significant or marginally significant crosslevel interactions were also observed for self-affirmation (b = 0.03, p = .04) and rebound motives (b = 0.16, p = .04) predicting patterns of change over time in new partners, and

Fig. 4 Top panel interaction of average within-person rebound motives X breakup time ago predicting probability of having sex with a new partner. Bottom panel interaction of average within-person appearance self-esteem X breakup time ago predicting probability of having sex with a stranger

for appearance self-esteem (b = -0.03, p = .06) predicting patterns of change in the probability of stranger sex. Plotting these interactions revealed patterns for rebound and self-affirmation motives that were nearly identical and consistent with expectations based on rebound lore. Figure 4 (top panel) illustrates this pattern for rebound motives. As shown in the top panel, although people who were low versus high in rebound motives were equally likely to have had sex with a new partner at 4 weeks postbreakup, those who were low on rebound motives showed a steep decline over time in the probability of having sex with a new partner, whereas their high motive counterparts continued to have sex with new partners at the same or even higher rates. Finally, as shown in the bottom panel of Fig. 4, plotting the appearance selfesteem X time-since-breakup interaction revealed a pattern in which individuals with high appearance self-esteem were somewhat more likely than their less confident peers to have sex with a stranger at 4 weeks post-breakup. However, rates of stranger sex declined steeply for those with high appearance selfesteem such that they were eventually less likely than those with low appearance self-esteem to have stranger sex.

123

Arch Sex Behav

Discussion The present study used a longitudinal, online diary method to explore widely held beliefs about recovery from the loss of a romantic relationship and in particular whether people used sex as a way to get over or get back at their ex-partners, phenomena commonly known as rebound and revenge sex. Overall, our results provided support for these beliefs, indicating that people have sex to cope with feelings of distress, anger, and diminished self-esteem in the aftermath of a breakup, and that those who have sex for these reasons were more likely (not initially, but over time) to continue having sex with different‘‘new’’partners, suggesting that they may be slower to recover from the breakup. Individual differences in patterns of recovery were also explored, comparing those who were ‘‘dumped’’ versus initiated the breakup or were in more (vs. less) committed and longer-lasting (vs. shorter) relationships. Findings generally supported widely held beliefs about the effects of being left by one’s partner, indicating that those who were‘‘dumped’’were more distressed, angrier, and more likely to use sex as a way to deal with the loss. Results were more complex, however, for relationship commitment and length, indicating that how people from more committed or longer relationships behave in the aftermath of a breakup only partially conforms to widely held beliefs on the topic. Normative Patterns of Recovery Following a Romantic Relationship Breakup The present study replicated and extended past research on recovery from a romantic relationship breakup, showing that the average person experienced heightened distress in the immediate aftermath of the breakup, which then declined and leveled off at about 6 months post-breakup. People also appeared, on average, to suffer a loss in social self-esteem, which slowly recovered over time. The average person also reported higher levels of coping, rebound, and revenge motives for sex immediately after the breakup, which then declined over time. Likewise, the probability of having sex with a new partner declined over time, leveling off between 13 and 16 weeks post-breakup. In contrast, no temporal trends were found for appearance selfesteem, self-affirmation motives, or the likelihood of having sex with a stranger, suggesting that these outcomes were, on average, unaffected by the breakup. Moreover, as previously described, modal patterns of recovery did not differ for men and women. These data (shown in Table 2) were consistent with widely held beliefs about‘‘being on the rebound’’following romantic relationship loss. In particular, evidence showing that people are more likely to have sex to cope with feelings of distress and to get over or get back at their ex-partners immediately following a breakup, and that these motives diminish over

123

time, provides clear support for the idea that at least some people use sex as a short-term coping strategy in the immediate aftermath of a relationship loss. Interindividual Differences in Patterns of Recovery Examining patterns of interindividual differences in recovery (see Table 3) showed that, of the three predictors examined (viz., which partner initiated the breakup, relationship duration, commitment; though see footnote 2), results for partner initiation were strongest and most consistent with rebound lore. Those who were ‘‘dumped’’ were initially more distressed and angry and were more likely to have sex to cope with a host of negative feelings experienced in the aftermath of the breakup. Although cross-level interactions showed that people who were dumped‘‘caught up’’to those who were not by 20–25 weeks post-breakup on most dimensions, they remained angrier across time. These data suggest that individuals who are‘‘dumped’’have more trouble letting go of the past relationship even after 8 months. As shown in Table 3, results for relationship duration were substantially less consistent across outcomes, suggesting (perhaps not surprisingly) that the simple passage of time provides a crude measure of the psychological significance of a relationship. In contrast, results for commitment were more complex and nuanced, indicating that more committed individuals were less likely (especially initially) to have sex, but more likely to use it as a way to cope with the emotional aftermath of the breakup when they do. In contrast, those who were less committed were more likely to have sex in the immediate aftermath of the breakup, but less likely to use it as a way to cope, presumably because they were less invested in the lost relationship and more ready to move on. These findings, together with the strong pattern of results for partner initiation, suggest that the fact of not choosing or wanting the relationship to end may be the single best predictor of the suite of responses that characterize being ‘‘on the rebound.’’ Finally, individuals who reported higher average levels of rebound, revenge, coping, and self-affirmation motives were more likely to have had sex with a stranger at four weeks postbreakup and (as indicated by the lack of significant linear or quadratic effects) remained more likely to do so over time. In contrast, individuals who were high (vs. low) on rebound and self-affirmation motives showed differential patterns of change over time in the probability of having sex with a new partner (Fig. 4, top panel). Whereas both those who were high and low in these motives reported the same elevated rates of sex with new partners at the outset, only those low in these motives declined over time. Those who were high continued to have sex with new partners, suggesting that people who use sex to cope with the psychological aftermath of a breakup have greater difficulty transitioning to a stable new partnership.

Arch Sex Behav

Clarifying the Nature of ‘‘Rebound’’ Phenomena Based on an informal analysis of lay beliefsabout the meaning of rebound and revenge sex, we argued that both phenomena are defined primarily by the motives they subserve—in the case of rebound sex, the motivation to get over feelings of pain and loss associated with the breakup, and in the case of revenge sex, the motivation to strike back at or get even with one’s ex-partner. On the whole, our findings lend support to this view, showing the clearest pattern of results for negative feelings and the use of sex to cope with or manage these feelings. Indeed, self-reported motives indexing sex to cope with a range of negative feelings following the breakup (including distress over loss of the ex, anger at the ex, general upset or sadness, and insecurities or selfdoubts) were the only predictors to: (1) exhibit the expected patterns of change over time, (2) show reliable differences as a function of partner initiation and relationship commitment, and (3) differentially predict the likelihood of having sex with a stranger as well as patterns of change in new partners over time. Together, these data support the contention that having sex to cope with the emotional aftermath of a breakup lies at the heart of the rebound phenomenon and suggests that the best and most direct indicators of this phenomenon may well be the explanations people provide for their sexual experiences. In contrast, results were less consistent for self-esteem. Although social self-esteem showed a significant increase over time, suggesting that confidence in one’s social worth was diminished as a result of the breakup and then steadily recovered, appearance self-esteem remained constant throughout the study period. In addition and also contrary to expectation, both self-esteem variables were higher (not lower) among those who had been in longer relationships, perhaps reflecting the greater ability of high self-esteem individuals to maintain a relationship over time. These findings in conjunction with the fact that social and appearance self-esteem were among the most stable of all outcomes examined during the diary phase (see Table 1) suggest that self-esteem, at least as measured in the current study, is a relatively stable property of the individual and, as such, may be relatively unaffected by relationship loss. Moreover, as suggested by the pattern shown in Fig. 4 (bottom panel), those who were dispositionally high in appearance self-esteem may be better equipped than their low esteem counterparts to attract sexual partners in the immediate aftermath of the breakup and, as indicated by the decline in sex with strangers over time, ultimately to enter into a new, stable relationship. In short, although there is some evidence that social self-esteem suffers immediately after a breakup, appearance self-esteem appears to be relatively impervious to the adverse effects of relationship loss and, if anything, may ultimately facilitate re-entry into a new relationship. The data for actual sexual experiences were also complex and inconsistent, thus raising questions about how best to think of sexual experiences in the aftermath of a breakup. Should having

sex with a new or unknown partner, particularly those instances motivated by a need to cope with the emotional sequelae of the breakup, be viewed as risky and ill-advised, or should they be seen as useful steps toward leaving an old relationship behind and moving on to a new one? Based on our findings, we would argue that, for some people, having sex with a new partner is a healthy and necessary part of moving on. However, the fact that those who reported the use of sex to cope with or get over the relationship loss continued to have sex with new partners even after 8 months suggests that, for at least these individuals, this behavior signified a lack of recovery and an inability to move on. Further complicating the interpretation of these data, sexual behaviors were only partially controlled by the individual, making them highly indirect and imperfect reflections of internal psychological states. By pointing this out, we do not mean to imply that actual sexual experiences should not be measured in studies on rebound phenomena. On the contrary, their inclusion is crucial if we hope to develop a better understanding of the role sexual behaviors play in recovery from relationship loss. Rather, we point out this limitation to highlight the complex nature of sexual experiences and the fact that such behaviors may not conform perfectly to predictions based on rebound lore. Strengths and Limitations The current research provided a number of important contributions to the literature on recovery from a romantic relationship breakup.First,byreplicatingresultsofpaststudiesonthenatureof psychological responses toa breakupusing alongitudinal, withinperson, diary method, and by extending these findings over time, the present study provides the most compelling evidence to date on the nature and time course of emotional recovery following a romanticrelationshipbreakup,atleast among youngadult college students. The present study also provided important new insights into the motivational and behavioral consequences of a breakup. Indeed, we are aware of no study using prospective methods to examine the sexual motives and behaviors of people‘‘on the rebound.’’Consequently, the current study provided novel information on a largely unexamined aspect of the recovery process, as well as the strongest test to date of widely held beliefs about rebound and revenge sex. Despite these strengths, several limitations must be acknowledged. First, trajectories estimated in the present study represent an amalgam of within-person changes over time and betweenperson differences. As previously pointed out, individuals who broke up one week ago contributed to the growth curve between weeks 1 and 12 whereas those who broke up 23 weeks ago contributed to the growth curve for weeks 23–34. Unfortunately, the fact that the two groups might differ in ways (other than elapsed time since the breakup) that also (or instead) influenced the level and pattern of observed changes cannot be ruled out. However, confidence in our trajectory analyses is enhanced by the fact that

123

Arch Sex Behav

these findings were for the most part intuitively sensible, consistent with past research on the psychological sequelae of relationship loss, and predicted on the basis of rebound lore. Further, although all possible confounding influences cannot be ruled out, supplemental analyses (not reported) revealed no significant differences in demographics or personality variables (i.e., the Big 5), or in relationship or breakup characteristics, as a function of elapsed time since the breakup. Second, our sample consisted primarily of students in their first semester of college, thus raising questions about the degree to which observed changes were due to the transition to college itself rather than recovery from a breakup. Even though markers of the transition process (e.g., freshman status) were controlled in all analyses, thisissuecannot be resolved definitively basedon the present data. That our findings conform to empirical evidence on relationship loss and that recovery trajectories varied in theoretically meaningful ways as a function of breakup characteristics serve to allay this concern. Nevertheless, future research using a control group of entering freshmen who did not experience a recent breakup, or using a sample of individuals who were not freshmen entering college, would provide a stronger basis for inferences about recovery trajectories distinct from adjustment-to-college effects. Third, although our sample was the largest ever used in a diary study of romantic relationship breakups, only a small number of participants had broken up in the three weeks prior to the study. This unfortunately introduced uncertainty into the estimation of effects during the first few weeks, which is reflected in larger SEs and lower power to detect differences (Raudenbush & Bryk, 2002). Given that individuals are presumably the most vulnerable during this period, future studies should aim to collect data from individuals as soon as possible after the breakup, though we recognize that obtaining large samples of this sort is difficult. Fourth, due to the characteristics of our sample, questions concerning the generalizability of our results to other populations, to loss of other types of relationships (e.g., marriage), and to breakups at different life stages also arise. However, as with all questions of generalizability, systematic replications will be required to address these questions. Finally, the issue of whether having sex with a new partner or a stranger in the aftermath of a breakup is functional and adaptive or ill-advised and risky was not directly examined in the present study. Although we believe that both logic and empirical evidence (Perilloux, Duntley, & Buss, 2011) strongly suggests that having sex with a stranger (i.e., someone you just met) imposes excess risk of negative outcomes (e.g., being victimized), and may therefore be both ill-advised and risky, it is also possible that such encounters could become part of the healing process. For example, Owen, Fincham, and Moore (2011) found that young adults who engaged in hook ups (sex with no mutual expectation of further interaction) felt less depressed and lonely at follow-up than those who did not, at least if they initially

123

reported high levels of depression and loneliness. Although not identical to the sexual behavior outcomes used in the present study, such findings suggest that the effects of having sex, including casual sex, are complex. Thus, future research could usefully examine the individuals for whom, or the circumstances under which, having sex on the rebound serves to undermine versus facilitate recovery from relationship loss. These limitations notwithstanding, the present study provides support for widely held beliefs about rebound phenomena, indicating that people use sex to cope with distressed feelings following a romantic relationship breakup, although actual instances of ‘‘rebound’’ sex may be relatively uncommon. Future research replicating these findings in more diverse samples with data collected closer in time to the breakup will enhance confidence in these conclusions and provide an opportunity to flesh out a more complete and nuanced picture of the potentially important role that sexual motivations and behavior play in recovering from romantic relationship loss.

References ABC News. (2004, October 21). The American sex survey: A peek beneath the sheets. Retrieved from http://abcnews.go.com/images/ Politics/959a1AmericanSexSurvey.pdf. Anderson, E. R., Greene, S. M., Walker, L., Malerba, C., Forgatch, M. S., & DeGarmo, D. S. (2004). Ready to take a chance again: Transitions into dating among divorced parents. Journal of Divorce & Remarriage, 40, 61–75. Bancroft, J., Janssen, E., Strong, D., Carnes, L., Vukadinovic, Z., & Long, J. (2003). The relation between mood and sexuality in heterosexual men. Archives of Sexual Behavior, 32, 217–230. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Cooper, M. L., Shapiro, C. M., & Powers, A. M. (1998). Motivations for sex and risky sexual behavior among adolescents and young adults: A functional perspective. Journal of Personality and Social Psychology, 75, 1528–1558. Davis, D., Shaver, P. R., & Vernon, M. L. (2003). Physical, emotional, and behavioral reactions to breaking up: The roles of gender, age, emotional involvement, and attachment style. Personality and Social Psychology Bulletin, 29, 871–884. DiTommaso, E., Brannen-McNulty, C., Ross, L., & Burgess, M. (2003). Attachment styles, social skills and loneliness in young adults. Personality and Individual Differences, 35, 303–312. Fine, M. A., & Sacher, J. A. (1997). Predictors of distress following relationship termination among dating couples. Journal of Social and Clinical Psychology, 16, 381–388. Frazier, P. A., & Cook, S. W. (1993). Correlates of distress following heterosexual relationship dissolution. Journal of Social and Personal Relationships, 10, 55–67. Gall, T. L., Evans, D. R., & Bellerose, S. (2000). Transition to first-year university: Patterns of change in adjustment across life domains and time. Journal of Social and Clinical Psychology, 19, 544–567. Gebhardt, W. A., Kuyper, L., & Greunsven, G. (2003). Need for intimacy in relationships and motives for sex as determinants of adolescent condom use. Journal of Adolescent Health, 33, 154–164. Grossbard, J. R., Lee, C. M., Neighbors, C., Hendershot, C. S., & Larimer, M. E. (2007). Alcohol and risky sex in athletes and nonathletes: What roles do sex motives play? Journal of Studies on Alcohol and Drugs, 68, 566–574.

Arch Sex Behav Heatherton, T. F., & Polivy, J. (1991). Development and validation of a scale for measuring state self-esteem. Journal of Personality and Social Psychology, 60, 895–910. Helgeson, V. S. (1994). Long-distance romantic relationships: Sex differences in adjustment and breakup. Personality and Social Psychology Bulletin, 20, 254–265. Kessler, R. C. (1995). The national comorbidity survey: Preliminary results and future directions. International Journal of Methods in Psychiatric Research, 5, 139–151. Leigh, B. C., Temple, M. T., & Trocki, K. F. (1993). The sexual behavior of US adults: Results from a national survey. American Journal of Public Health, 83, 1400–1408. Mehta, P. D., & West, S. G. (2000). Putting the individual back into individual growth curves. Psychological Methods, 5, 23–43. Meston, C. M., & Buss, D. M. (2007). Why humans have sex. Archives of Sexual Behavior, 36, 477–507. Owen, J., Fincham, F. D., & Moore, J. (2011). Short-term prospective study of hooking up among college students. Archives of Sexual Behavior, 40, 331–341. Patrick, M. E., Maggs, J., Cooper, M. L., & Lee, C. M. (2011). Measurement of motivations for and against sexual behavior. Assessment, 18, 502– 516. Perilloux, C., & Buss, D. (2008). Breaking up romantic relationships: Costs experienced and coping strategies deployed. Evolutionary Psychology, 6, 164–181. Perilloux, C., Duntley, J. D., & Buss, D. M. (2011). Susceptibility to sexual victimization and women’s mating strategies. Personality and Individual Differences, 51, 783–786. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage.

Raudenbush, S., Bryk, A., Cheong, Y. F., & Congdon, R. (2004). HLM 6 hierarchical linear modeling [Computer software]. Chicago: Scientific Software International. Santelli, J. S., Brener, N. D., Lowry, R., Bhatt, A., & Zabin, L. S. (1998). Multiple sexual partners among U.S. adolescents and young adults. Family Planning Perspectives, 30, 271–275. Sbarra, D. A., & Emery, R. E. (2005). The emotional sequelae of nonmarital relationship dissolution: Analysis of change and intraindividual variability over time. Personal Relationships, 12, 213–232. Simpson, J. A. (1990). Influence of attachment styles on romantic relationships. Journal of Personality and Social Psychology, 59, 971–980. Slotter, E. B., Gardner, W. L., & Finkel, E. J. (2010). Who am I without you? The influence of romantic breakup on the self-concept. Personality and Social Psychology Bulletin, 36, 147–160. Smith, T. W. (1991). Adult sexual behavior in 1989: Number of partners, frequency of intercourse and risk of AIDS. Family Planning Perspectives, 23, 102–107. Sprecher, S. (1994). Two sides to the breakup of dating relationships. Personal Relationships, 1, 199–222. Tashiro, T., & Frazier, P. (2003).‘‘I’ll never be in a relationship like that again’’: Personal growth following romantic relationship breakups. Personal Relationships, 10, 113–128. Wade, L. D., & DeLamater, J. D. (2002). Relationship dissolution as a life stage transition: Effects on sexual attitudes and behaviors. Journal of Marriage and Family, 64, 898–914. Watts, C., Cox, T., & Robson, J. (1983). Morningness-eveningness and diurnal variations in self-reported mood. Journal of Psychology: Interdisciplinary and Applied, 113, 251–256.

123

Rebound sex: Sexual motives and behaviors following a relationship breakup.

The present study used a longitudinal, online diary method to examine trajectories of psychological recovery and sexual experience following a romanti...
404KB Sizes 0 Downloads 0 Views