Dating, marital, and hypothetical extradyadic involvements: how do they compare?

In general, most people express beliefs that relationship partners should be faithful to each other, whether in a dating or marital relationship (Sheppard, Nelson, & Andreoli-Mathie, 1995; Thornton & Young-DeMarco, 2001). However, sexual or romantic relationships outside of a primary marital or dating dyad (i.e., extradyadic involvement or EDI) occur fairly often, particularly in dating relationships (Wiederman, 1997; Wiederman & Hurd, 1999). EDI has been a focus of research interest for testing hypotheses in evolutionary psychology (e.g., Shackelford & Buss, 1996), because of the potential negative impacts on individual, relationship, and sexual health of the involved parties (Allen et al., 2005), and to further basic understanding of relationships. EDI has been studied in different populations, such as undergraduate, marital therapy, and community samples, and with different methods, such as self-reporting on actual personal experiences of EDI, predicting personal experiences of EDI, reacting to hypothetical EDI vignettes, and, in one study, testing actual behavior (Seal, Agostinelli, & Hannett, 1994). However, it is rare to find direct comparisons regarding these different populations or methods of studying EDI. This study explored the similarities and differences in descriptions of EDI from undergraduates reporting on an actual recent (within the last two years) experience of EDI, undergraduates who denied recent EDI and were asked to imagine hypothetically what they might experience if they did engage in it, and community adults reporting on an experience of EDI they had while married or engaged.

We anticipated many differences between undergraduate experiences of EDI in a dating relationship and community adult experiences of EDI during marriage or engagement due to differences in the developmental phase of both the involved person and the primary relationship. The processes of forming close and committed relationships with others, and vacillations in this closeness and commitment, represent a fundamental challenge for emerging adults (e.g., Erikson, 1963). DeGenova and Rice (2005) highlighted several experimental functions of dating relationships for youth, such as exploring gender roles, aspects of self-identity, and sexuality, noted the lack of clear "rules" for dating in contemporary culture, and described the divergent priorities college students have for dating versus marriage. Prior research has shown that, although they generally disapprove of the behavior, college students do not condemn dating EDI as much as marital EDI (Sheppard et al., 1995). In addition, while getting married typically involves a significant commitment to the relationship from both partners (Kline et al., 2004), dating relationships often have lower levels of commitment (Edin, Kefalas, & Reed, 2004), and dating men may even deceive partners about their levels of commitment to the relationship (Tooke & Camire, 1991). Not surprisingly, rates of EDI appear higher in college student dating relationships relative to married relationships (Wiederman, 1997; Wiederman & Hurd, 1999). Due to the relatively lower standards for fidelity, lower commitment, and shorter duration of dating relationships relative to marriage and the subjective importance and relative sanctioning of exploring self and others in romantic and sexual encounters during this stage of life, there may be several differences between dating and marital EDI.

On the other hand, Roscoe, Cavanaugh, and Kennedy (1988) noted similarities in behaviors seen in dating and marital relationships and stated that "in many respects dating is preparation for marriage, and may lead to the establishment of behavioral patterns which will be maintained in marriage" (p. 37). Hence, in regards to EDI specifically, Roscoe et al. anticipated "similarities in infidelity in these two types of dyads" (p. 37). Similarly, Drigotas, Safstrom, and Gentilia (1999) posited that "the root causes of infidelity in dating relationships extend to marriage" (p. 510). Thus, one goal of this study was to compare and contrast dating and marital EDI to identify similarities and differences in these experiences.

In addition to comparing dating and marital EDI, it is also important to ascertain whether individuals asked to imagine their thoughts and feelings in a hypothetical experience of EDI respond differently than persons answering questions based on an actual experience of EDI. This is of interest because many researchers have utilized hypothetical scenarios of EDI, asking participants to imagine that either they or their partners have been unfaithful. Some research has evaluated whether imagined reactions to a partner's EDI vary based on the respondent's actual experiences of partner EDI. For example, Buunk (1995) asked respondents to imagine that their partner informed them that he or she had sexual contact with a third person, and found that projected emotional reactions varied based on how often the respondent's partner had actually been unfaithful. Harris (2002) compared individuals' imagined reactions upon discovering a partner's EDI and their actual recalled reactions to partner EDI, and found no correspondence in these reactions. Thus, prior literature suggests that hypothesized reactions to partner EDI may differ based on actual history of partner EDI and that hypothesized reactions may differ from actual reactions.

Other researchers have asked participants to take on the role of a person engaging in EDI or to speculate about their likely feelings and behaviors if they did engage in EDI. For example, Mongeau and colleagues (Mongeau, Hale, & Alles, 1994; Mongeau & Schulz, 1997) provided an EDI scenario and asked undergraduate participants to take the role of the person engaging in the EDI. Respondents were asked about their level of guilt, blame, and responsibility, and how they might communicate to their primary partner subsequent to the EDI. Wiederman and Allgeier (1996) asked young married participants about their likelihood of engaging in EDI, under what conditions they would allow themselves to engage in EDI, and whether or not they would tell their spouse if they engaged in EDI. No studies have explicitly compared hypothetical accounts of EDI behavior to actual accounts of EDI behavior. However, prior literature suggests that what people anticipate about extradyadic behavior differs from actual EDI. For example, Wiederman and Allgeier found that 87% of respondents indicated that EDI would never be acceptable for them, and 69% reported that, if they did engage in EDI, they would tell the partner. However, lifetime rates of infidelity are generally higher, and rates of partner knowledge are often found to be lower, than these projections (Allen et al., 2005). Moreover, people often have stereotypes or preconceived notions about infidelity which do not mesh with reality. For example, a common assumption is that EDI suggests defects in the primary relationship, but evidence suggests that many individuals engage in EDI in the absence of relationship distress (Glass & Wright, 1985). Individuals asked to imagine their thoughts and reactions to engaging in EDI, particularly individuals without recent EDI experience, may describe different experiences relative to those asked to report on an actual experience of EDI.

In this study, we compared and contrasted several aspects of EDI experience from (a) undergraduates reporting on an actual experience of EDI in a dating relationship in the prior two years, (b) undergraduates without a recent (prior two years) experience of EDI asked to imagine engaging in EDI, and (c) community adults reporting on an actual experience of EDI when married or engaged. First, we hypothesized differences in hypothetical and actual involvement. We hypothesized that participants who denied recent EDI and were asked to imagine a hypothetical experience of EDI would not show the same consistency in responses within and between measures relative to persons reporting on actual EDI. That is, in actual EDI, there are aspects of the experience that converge and diverge in theoretically consistent and empirically supported ways. People without experience of EDI (or, at least, without recent experience) who are asked to imagine EDI may not be able to predict the relationships that are actually found between diverse variables, or they may fluctuate more in their responses due to the lack of an actual recent experience in which to ground their self-report.

Individuals without a history of extramarital involvement are more disapproving of extramarital sex and say they would feel less justified engaging in an affair compared to those who have engaged in extramarital sex (Glass & Wright, 1992; Greeley, 1994). Thus, we also predicted that undergraduates who denied a recent experience of EDI would generally be more disapproving of EDI, and that when asked to imagine themselves engaging in a hypothetical EDI, they would be less likely to justify or minimize the EDI and would be more likely to say the behavior was inconsistent with their values and behavior. Due to the hypothesized negative attitudes toward EDI and the sense that engaging in EDI is a greater violation of their values and typical patterns of behavior, we also predicted that participants reporting on a hypothetical EDI would anticipate greater remorse afterward. This prediction was also based on the fact that, although most people expect that they would never engage in EDI, many individuals who eventually engage in EDI feel justified in doing so based on the various factors that motivated the involvement (Glass & Wright, 1992). Justifications, minimizations, and attachment to the extradyadic partner are all forces which can lead individuals to lower their regret about engaging in the behavior; persons reporting on a hypothetical EDI may not anticipate these mitigating factors.

We also predicted that people reporting a hypothetical EDI will imagine a high level of relationship distress prior to the EDI and posit motivations for the EDI based more on deficits in the relationship. We expected these findings because people reporting on a hypothetical EDI may also operate more on stereotypes of EDI, and thus assume that EDI would only occur in the presence of relationship distress (when, in fact, many individuals engage in EDI in the absence of relationship distress). In addition, because people who deny EDI likely hold more negative attitudes toward it, they may speculate that they would require a high level of relationship distress to precipitate such behavior and that their behavior would be more motivated by relationship problems.

We also expected to find differences between marital and dating EDI, primarily because dating relationships may be less stable (i.e., shorter duration) and committed, on average, than marital relationships, and because the standards for monogamy may be lower in a dating union. Given these differences, people reporting on dating EDI should (a) consider EDI a less serious violation of relationship standards and thus report lower levels of remorse subsequent to the EDI compared to persons reporting marital EDI, and (b) require a lower threshold of relationship distress to motivate an EDI and thus report lower levels of relationship distress prior to the onset of EDI compared to persons reporting marital EDI.

It is difficult to predict whether the type of EDI (e.g., casual involvement or more intense emotional involvement) will differ between dating and marital EDI. College student romantic relationships are often characterized as either casual sexual arrangements or very intense involvements (Glenn & Marquardt, 2001). If these patterns hold in EDI, we might expect that dating EDI may be often be described as a casual fling or as an intense, obsessive affair. However, in marital EDI, some individuals have casual "flings" or one-night stands, whereas others feel driven to violate marital vows because of intense attraction or needs; thus, the same patterns likely exist.

We included gender in the analyses comparing groups on different aspects of the EDI, given that some characteristics of EDI differ for males and females. For example, we have already reported main effects for gender on motivations for EDI (Allen & Baucom, 2004), with women reporting more emotional intimacy motivations (e.g., feeling lonely or neglected in the primary relationship and wanting to feel cared about by the affair partner). This is consistent with Glass and Wright's (1992) review and findings regarding women's relatively higher focus on love and emotional intimacy motivations for EDI (compared to men's relatively higher focus on sexual motivations). Additionally, EDI is more strongly linked to primary relationship dissatisfaction for women as compared to men (Glass & Wright, 1985; Prins, Buunk, & Van Yperen, 1993). Thus, we anticipated gender differences in ratings of the primary relationship prior to the onset of the EDI, with men rating the relationship more positively than women. We also expected to replicate the typical finding that men are more approving of EDI (e.g., Oliver & Hyde, 1993). We were also interested in the interaction term between type of EDI (dating, hypothetical, marital) and gender to evaluate if males and females in the various groups differ.

METHOD

See Allen and Baucom (2004) for more information regarding participants, procedures, and measures.

Participants

We evaluated three separate samples: an undergraduate sample reporting on an actual experience of EDI ("dating EDI"), an undergraduate sample reporting on a hypothetical experience of EDI ("hypothetical EDI"), and a community sample reporting on an actual experience of EDI while married or engaged ("marital EDI"; approximately 90% of these experiences occurred during marriage).

The total undergraduate sample consisted of 504 students from a large southeastern university. All participants were required to have been involved in a primary dating relationship of at least one month in duration within the last two years. Of this overall sample, 345 (69%) reported some degree of EDI (defined as romantic or sexual behavior with someone other than the primary partner while in a primary relationship) within the last two years. The behaviors encompassed in this definition of EDI were quite broad, ranging from falling in love to sexual intercourse. A 69% rate of EDI is consistent with prevalence rates for an undergraduate population found in other research that also included a broad range of extradyadic activity (Wiederman & Hurd, 1999). These participants who had engaged in EDI were asked to report in detail on their most recent EDI; this group was the "dating EDI" group.

The 159 (31%) participants who had not engaged in EDI in the prior two years were asked to answer questions based on hypothetical involvement; this group was the "hypothetical EDI" group. Specifically, the directions were as follows: "If you have NOT had sexual or romantic contact with anyone outside of your primary relationship, please IMAGINE that you end up having a sexual or romantic encounter with someone other than your primary dating partner. Please answer the following questions as if you are actually in the situation of having a steady dating relationship and have sexual or romantic contact with someone else. That is, think about your personality, emotions, and values, situations you might be in either deliberately or accidentally, things that attract you to people sexually or romantically, and other factors that might relate to this. Based on these factors, imagine what kind of relationship this might be, how you might feel about it, and what sorts of things might motivate you to have this sexual or romantic encounter when answering the following questions." Thus, the directions were designed to engage the respondent in imagining the hypothetical EDI experience in a personal way. The directions preceding this prompt also made it clear that we were inquiring about concurrent sexual or romantic involvement; that is, involvement with another person that occurred while in the primary relationship.

The two undergraduate subsamples (i.e., those reporting on dating EDI and those reporting on hypothetical EDI) were nearly identical in their demographic information, so we present information on the overall undergraduate sample. The sample was 60% female and 40% male, averaged 19 years old (SD = 1.33), and were mostly (81%) White (non-Hispanic); 12% were African American, 1% Hispanic, 4% Asian American, 1% Native American, and 1% other.

The third sample consisted of volunteer married or formerly married community participants. From a larger volunteer community sample, 115 respondents indicated that they had engaged in EDI during a marriage or engagement. This sample consisted of 52 females and 63 males. Ages in this sample ranged from 24 to 74 (M = 42.49, SD = 8.60). Most (85%) of the sample was White/non-Hispanic, with 13% African-American, 1% Asian American, and 1% other. The sample was highly educated: 72% reported a bachelor's degree or higher level of education. We asked these respondents to report in detail on their most recent EDI. Two individuals did not provide complete information on some dependent variables; thus, we only included 113 participants in some analyses.

Procedure

Undergraduates participated in a study of "Experiences in Relationships" in exchange for introductory psychology course credit. Participants came to the laboratory to complete the self-report measures confidentially. Since we asked all participants to complete the entire battery of measures (based on either real or hypothetical experience), all participants took approximately the same amount of time to complete measures; thus, their EDI status could not be identified by the amount of time required to complete measures. After completion of the questionnaires, all participants were debriefed with a written explanation of the basic research questions, the development of the materials, and the research procedure; verbal explanations and elaborations were provided to interested participants.

For the community sample, packets were distributed in areas of heavy pedestrian traffic, such as an airport and a shopping area. Packets were not individually identified, were enclosed in a business reply envelope addressed to the primary investigator, and included the set of questionnaires as well as an informed consent form for the participant to keep. Using this procedure, participants retained anonymity. Adults who either were alone or in same sex-groups were approached and asked if they were or had been married and if they were interested in completing a survey about experiences in marriage. If the individual expressed interest, the investigator explained that it was about a highly personal experience that he or she may or may not have had in marriage and that he or she should feel free to discard the survey if it was considered too personal. All questions that potential participants had about the study were answered (excluding any information about specific hypotheses), and individuals were frequently given more information about the topic of the study, affiliation of the investigator, and protections of anonymity. Volunteers then took the packets to complete and mail back without charge and at their convenience.

Measures

Patterns of EDI. We developed a measure to assess a range of thoughts, feelings, and behaviors associated with the EDI experience entitled the Extradyadic Experiences Questionnaire (EEQ; Allen & Baucom, 2004). This measure was used to assess the consistency of responses for the hypothetical EDI group and to assess differences among all three groups on aspects of the EDI. The EEQ consists of general questions about experiences of EDI followed by subscales assessing various aspects of the most recent EDI (see below). All items on these scales were face valid and asked the respondent to self-appraise thoughts and feelings, such as being "obsessed" with the affair partner or "wanting space" from the primary partner. Subscale questions consisted of declarative statements with response options on a 7-point Likert scale, usually from 1 (not at all true) to 7 (very true). Scale scores are the average of all items on the scale.

In this study, we report on 3 EEQ scales which assessed the type of relationship the participant had with the extradyadic partner: Casual EDI (a fun, "no-strings-attached" relationship without emotional investment), Close EDI (a comfortably loving, close, and open relationship), and Obsessive EDI (a needy, obsessive relationship with high desires for attention and reassurance). Three subscales measured the appraisals of the primary relationship prior to the onset of the EDI: Close primary (satisfaction with amount and quality of relationship intimacy), Autonomy complaints (description of the primary relationship as smothering and primary partner as not providing enough space), and Intimacy complaints (description of the primary partner as unavailable, uncaring, or rejecting). Four scales assessed motivations for the EDI: Autonomy reasons (wanting freedom and space from primary relationship), Intimacy reasons (needing closeness and caring not provided by the primary relationship), Self-esteem reasons (wanting an increase in self-esteem), and Love reasons (strong feelings for the extradyadic partner). Five scales assess remorse and distress related to the EDI: Remorse (general remorse regarding the EDI), Abandon (distress due to fears of losing either or both partners and being alone), Hurt partner (distress due to potential of hurting the primary partner), Disapprove (distress regarding potential negative evaluations from others), and Unlike me (distress due to the EDI being inconsistent with general values and patterns of behavior). One final scale, called Justify, assessed beliefs that justified or minimized the EDI as "no big deal."

Disapproval of EDI. We predicted that the hypothetical group would generally be more disapproving of EDI. We assessed disapproval with the Justifications for Extramarital Relationships Scale (JER; Glass & Wright, 1992). This mea sure was used in the undergraduate sample only; thus, comparisons were made between the dating EDI and the hypothetical EDI group. The JER asks respondents how justified they would feel engaging in extradyadic involvement for 17 different reasons, such as for fun, for sexual excitement, or for falling in love with another person. Glass and Wright used this questionnaire for a marital sample; questions were rephrased here for extradyadic, rather than extramarital, involvement. The JER preceded the EEQ, as recommended by Johnson (1970). Items are answered on a 4-point scale. Scores are the average of the 17 items; higher scores indicate feeling more justified in engaging in EDI for the various reasons (thus, more approving of EDI). In the whole undergraduate sample, the internal consistency of this scale was .85.

External measure of personality. We predicted that the hypothetical group would show lower consistency of responses between different measures. We chose to include a measure of personality that has shown reliable associations with aspects of EDI, as measured by the EEQ, in both the dating and the marital EDI groups in order to test whether these same associations were found in the hypothetical group. We used the Experiences in Close Relationships Inventory (ECRI; Brennan, Clark, & Shaver, 1998), a self-report, adult romantic attachment measure containing two 18-item scales assessing anxiety about abandonment and avoidance of closeness in adult romantic relationships. Following Brennan et al.'s (1998) recommended procedures, we placed participants into four attachment categories of secure, fearful, preoccupied, and dismissive, based on their scores on the two scales. Brennan et al. reported excellent internal consistency on the scales and good convergent validity for the attachment categories.

RESULTS

Consistency of Responses

We hypothesized that participants reporting on a hypothetical experience of EDI would not show the same patterns of convergence and divergence within and between measures as those reporting on an actual experience. First, we evaluated the internal consistency of the EEQ scales (see Table 1). Contrary to hypotheses, respondents basing their answers on a hypothetical involvement completed the scales of the EEQ with similar internal consistency, and thus similar general patterns of convergence and divergence between items within scales, compared to the two groups reporting on actual EDI. Items were interwoven in their order of presentation within global content areas (e.g., type of relationship with extradyadic partner), and thus respondents were not simply answering continuous sets of a few similar questions.

Next, we evaluated the relationships between selected scales of the EEQ for the three groups (see Table 2). Several scales should theoretically have negative interrelationships. ff a respondent described the extradyadic relationship as casual (casual EDI), then he/she should be less likely to endorse close and loving feelings for the extradyadic partner (close EDI), and the respondent should also be less likely to report obsessive feelings for the extradyadic partner (obsessive EDI). If a respondent described the level of closeness in the primary relationship as good (close primary), then he/she should be less likely to complain that the primary relationship did not provide either sufficient intimacy or autonomy (intimacy complaints and autonomy complaints). Even with some variability in the strength of the associations, the predicted relationships were generally supported in all three groups. The only exception was that data from undergraduates reporting on a hypothetical EDI did not show a significant negative relationship between casual EDI and obsessive EDI. Fisher's r to z transformation indicates that the correlation observed for the hypothetical EDI respondents is significantly lower relative to that obtained for both the dating EDI (t = 3.13, p < .01) and marital EDI groups (t = 3.23, p < .01).

Other scales should theoretically have positive interrelationships. Respondents who complained that their primary relationship did not provide sufficient intimacy (intimacy complaints) should be more likely to cite intimacy needs as a reason for their EDI (intimacy reasons). Similarly, respondents who complained that their primary relationship did not provide sufficient autonomy (autonomy complaints) should cite higher autonomy needs as a reason for their EDI (autonomy reasons). These predicted positive relationships emerged strongly for all three groups. This was found despite that, within the measurement battery, the scales assessing complaints about the primary relationship were given well before the scales assessing the reasons for the EDI. Thus, contrary to predictions, persons answering EDI questions hypothetically demonstrated the same types of divergent and convergent associations between items and scales as those found with persons answering questions based upon an actual experience of EDI.

In addition to the theoretically consistent patterns of convergence and divergence among the EEQ scales, we also tested the relationship between certain EEQ scales and the respondents' adult romantic attachment style. Specifically, we examined (a) the degree of complaints regarding the level of intimacy in the primary relationship prior to the EDI and (b) intimacy or autonomy reasons for engaging in the EDI as a function of attachment style. Relationships between these aspects of EDI and attachment style have been empirically confirmed in prior research in both the dating and the marital samples (Allen, 2001; Allen & Baucom, 2004). Thus, the question here was whether data from individuals reporting on hypothetical EDI would replicate these empirically-validated links. The instructions for participants reporting on a hypothetical involvement explicitly asked them to think about their personality, emotions, and values, and, based on this, to imagine what EDI would be like for them. Thus, the relationship between the respondents' personality style as measured by an attachment questionnaire and aspects of the EDI is an appropriate question for all three groups.

We used ANOVA to evaluate different levels of intimacy and autonomy variables for the four attachment styles (see Table 3). Pairwise comparisons were used to compare specific attachment styles to one another. Across all three samples (dating, hypothetical, and marital EDI), individuals higher in anxiety (the fearful and preoccupied groups) reported feeling particularly neglected in the primary relationship prior to the EDI (intimacy complaints) and said a need for greater intimacy (intimacy reasons) was a stronger motivation for the EDI relative to groups lower in anxiety (the secure and dismissive groups).

In prior research with samples reporting on actual EDI (Allen, 2001; Allen & Baucom, 2004), individuals categorized as dismissive reported higher levels of autonomy complaints and autonomy reasons relative to all other attachment groups. Thus, dismissive persons who have actually engaged in EDI are more likely to say that they felt smothered or trapped in the primary relationship prior to the EDI and that their reasons for EDI were more focused on getting space and freedom relative to all other attachment groups. When this association was tested in the hypothetical sample, we found some differences. A main effect of attachment was found for autonomy complaints, but not for autonomy reasons. In pairwise comparisons for both autonomy complaints and autonomy reasons, the dismissive group still earned the highest mean score, but this was significant only relative to the preoccupied group, and was not significantly higher than the secure or fearful groups. Thus, results did not cleanly replicate for autonomy complaints and reasons in the hypothetical EDI group. However, two separate two-way factorial ANOVAs with EDI status (dating EDI, hypothetical EDI, and marital EDI), attachment style, and the interaction of EDI status and attachment style as the independent variables, and autonomy complaints and autonomy reasons as the dependent variables did not manifest interaction effects (autonomy complaints F(6,604) = 1.41, p = .21; autonomy reasons F(6,604) = .68, p = .66). Thus, the levels of autonomy reasons and autonomy complaints as a function of attachment style did not significantly differ for the three samples. Moreover, these two-way ANOVAs both showed a main effect for attachment, in which the dismissive group was higher than the other three attachment groups, and no other differences were seen among attachment groups. Thus, when the hypothetical sample was subsumed in the larger sample of actual dating/marital EDI, the results replicated that finding with the actual samples only.

For all the ways we examined this question (internal consistency of EEQ scales, consistency between EEQ scales, relationship between EEQ scales and attachment), individuals who based their responses on a hypothetical experience of EDI generally showed the same patterns of convergence and divergence within and between measures as those persons reporting on an actual experience.

Differences Between Groups

The second set of hypotheses focused on predicted differences in descriptions of aspects of the EDI experience between those reporting on dating EDI, hypothetical EDI, and marital EDI using ANOVA (see Table 4).

Attitudes toward infidelity. The first prediction was that participants who denied engaging in EDI would more often endorse attitudes consistent with relationship fidelity. Consistent with hypotheses, persons reporting on hypothetical EDI scored significantly lower (thus, feeling less approving regarding EDI in general) on the Justifications for Extramarital Relationships (JER) measure relative to persons reporting on actual dating EDI, lower on minimization and justification of the specific imagined EDI experience (Justify), and higher on saying the imagined EDI was inconsistent with their values and general behavior (Unlike me) relative to persons reporting on actual dating or marital EDI.

Remorse. Also consistent with predictions, participants reporting on a hypothetical EDI reported higher global levels of remorse (remorse) than persons reporting on actual dating or marital EDI. Moreover, this pattern generally held for specific subtypes of distress (further suggesting consistency in hypothetical responses). People reporting on hypothetical EDI reported higher levels of distress regarding potential abandonment (abandon) relative to persons reporting on actual dating or marital EDI, and they reported higher levels of distress regarding hurting the primary partner (hurt partner) and being judged negatively by others (disapprove) relative to persons reporting on actual dating EDI.

We also predicted that the dating EDI sample would show lower levels of remorse than the marital EDI sample; however, undergraduates reporting on dating EDI did not have overall lower levels of remorse (remorse) or distress focused on potential abandonment (abandon) compared to the sample reporting on marital EDI. On the other hand, they did report relatively lower levels of concern regarding hurting the primary partner (hurt partner), being judged negatively by others (disapprove), or feeling like the EDI was inconsistent with their values and typical behaviors (unlike me) relative to persons reporting on marital EDI. Interestingly, an interaction effect emerged for hurt partner (see Figure 1). Whereas women were slightly higher on this measure than men in the undergraduate groups, married women were lower than married men on this variable (t[110]= 1.98, p = .05, two-tailed).

[FIGURE 1 OMITTED]

On average, the primary relationship of the undergraduates reporting on dating EDI was of shorter duration than the marriages of the individuals in the marital EDI sample. One might think that the amount of time in the primary relationship would affect global remorse regarding the EDI, yet overall remorse did not differ in the two samples, and further analyses in the dating EDI sample found length of time in the primary relationship at the time of the EDI to be uncorrelated with overall remorse (r = .04). However, as noted above, dating EDI respondents were less likely to be concerned about hurting the partner, being judged negatively, or feeling that the EDI behavior was unlike them. Length of time in the primary dating relationship at the time of the EDI was significantly correlated with these three sources of distress (hurt partner r = .13, p < .05; disapprove r =. 16, p < .01; unlike me r =. 12, p < .05). Thus, the areas in which dating EDI did result in lower distress relative to marital EDI may be partially explained by the relative brevity of dating relationships. Although it is not feasible to control for length of time in the relationship, given the inherently vast differences on this variable, we did subsequently reanalyze these three distress variables (disapprove, unlike me, hurt partner) using only participants who were in dating relationships lasting 12 months or longer at the time of the EDI. Specifically, we conducted two-way ANOVAs comparing males to females and those reporting marital EDI to dating EDI for those undergraduates who reported that the primary relationship was 12 months or longer at the time of the EDI, as well as the interaction term. Once we limited the dating sample in this manner, there was no main effect for EDI type (dating EDI vs. marital EDI) on unlike me, although the gender effect held (F[1,194] = 9.18, p < .01), such that women scored higher on this scale. We did not find an interaction effect. However, main effects for EDI type held for disapprove (F[1,195] = 5.34, p < .05) and hurt partner (F[1,194] = 4.24, p < .05), such that individuals reporting on dating EDI had lower scores on these scales. We did not find gender or interaction effects. Thus, even in dating relationships of longer duration, people reporting on dating EDI were relatively unconcerned (compared to persons reporting on marital EDI) about hurting the partner or being disapproved of by others due to the EDI.

Relationship satisfaction. Contrary to our predictions, participants reporting on hypothetical EDI actually endorsed higher levels of satisfying relationship intimacy (close primary) relative to those reporting on actual dating or marital EDI, and there were no differences in the reported level of satisfaction with relationship intimacy prior to the EDI between individuals with dating EDI and marital EDI.

However, when asked about specific reasons for engaging in the EDI, marital EDI respondents were more likely to attribute their EDI to feeling neglected and lonely in the primary relationship (intimacy reasons) relative to respondents describing a dating EDI. We found no differences between undergraduates reporting on an actual dating EDI versus a hypothetical EDI in terms of intimacy reasons, nor were any differences found between groups on autonomy reasons. Marital EDI respondents also scored most highly on reports of need for a self-esteem boost (self-esteem reasons) and love for the extradyadic partner (love reasons) as reasons for the EDI.

Nature of extradyadic relationship. No differences emerged on the reported emotional quality of the extradyadic relationship between groups reporting on dating, hypothetical, or marital EDI.

DISCUSSION

Asking individuals to imagine a hypothetical experience of EDI and to report in detail about that experience is a challenging task. Rather than being able to answer all questions based on an actual recent experience of reference, hypothetical responders need to imagine what their thoughts, feelings, and behaviors might be if they engaged in EDI. Of course, because we only explicitly assessed EDI within the last two years in the undergraduate sample, some individuals in the group reporting on hypothetical EDI may actually have been drawing upon a more remote history of adolescent EDI. However, since we asked individuals to think about current personality or situational factors when imagining the EDI, even individuals with a remote history of EDI may have been able to differentiate their current values and feelings from past values and feelings. Our procedure of asking individuals to imagine engaging in EDI maps onto other research procedures in which participants are asked to predict their thoughts, feelings, and reactions to EDI, regardless of their own personal history of the behavior (or of their history of having a partner engage in EDI). Most importantly, the findings indicated that this group is distinct from the groups that reported engaging in EDI in terms of attitudes toward EDI and imagined precipitants of and responses to EDI. Thus, although this is clearly an imperfect assessment of EDI history, the distinction appears meaningful.

Even though participants in the hypothetical EDI group had either never had an experience of EDI or at least had not had one in the prior two years, we found impressive internal consistency within scales assessing aspects of the EDI for this group. Also, scales assessing different aspects of the involvement generally converged and diverged in expected ways, even when these scales were separated by many other questions. Finally, the personality styles of the hypothetical group were generally found to differentiate aspects of the EDI in the same manner found within groups reporting on actual EDI. This suggests that individuals reporting on hypothetical EDI were generally able to imagine a consistent scenario of how, given their personalities, they might experience EDI and what experiences would and would not go together.

The minor exceptions are intriguing. For example, while dismissive individuals in all groups were highest on autonomy motivations for the EDI, the means for dismissive persons reporting on hypothetical involvement were not as extreme as those reporting on actual involvement, which contributed to an incomplete replication of findings regarding significant differences between dismissive individuals and all other attachment styles on autonomy motivations. Dismissive individuals who have not engaged in EDI may have somewhat lower autonomy motivations and thus feel less "driven" to engage in EDI. Alternatively, dismissive individuals who have not engaged in EDI do may not anticipate how strong autonomy motivations would be for them in precipitating an EDI. However, because there were no significant interaction effects between attachment and type of EDI (dating, hypothetical, or marital), these differences are minor.

Although the consistency within and among measures for the hypothetical EDI group was impressive, there were some important differences in aspects of EDI as imagined by participants reporting on a hypothetical involvement compared to those reporting on an actual involvement. Individuals reporting on a hypothetical involvement generally projected higher levels of remorse and various types of distress subsequent to an EDI compared to the levels of remorse and distress reported by individuals who had actually engaged in EDI. This could be influenced by at least two factors. First, people who have not engaged in EDI were found to be less likely to approve of or justify EDI and more likely to feel that the behavior was inconsistent with their values and typical behavioral patterns. Thus, these individuals could be predicting accurately that they would feel higher levels of remorse if they did engage in EDI. On the other hand, this finding could suggest that individuals generally imagine they will feel more regret after engaging in EDI than persons who engage in EDI actually experience. As noted earlier, many individuals who engage in EDI feel justified in doing so based on the various factors that motivated the involvement, or they may minimize the importance of the involvement to reduce dissonance. Justifications, minimizations, and attachment to the extradyadic partner are all forces which can lead individuals to lower their regret about engaging in the behavior. People reporting on a hypothetical EDI may not anticipate these forces. One of the primary methods in the literature of eliciting hypothetical responses is to provide participants with a standardized EDI scenario and ask about their likely thoughts, feelings, and behaviors. Our results suggest that individuals without (at least recent) EDI experience will predict higher levels of remorse and distress regarding their own hypothetical EDI behavior relative to persons who have actually engaged in the behavior.

A surprising result was that individuals reporting on a hypothetical involvement predicted higher levels of satisfaction with the intimacy in their primary relationship prior to engaging in EDI relative to that endorsed by persons reporting on actual dating or marital EDI. We expected that individuals who had not engaged in EDI and who generally felt less justified in engaging in this type of behavior would imagine that they would need to experience a high level of relationship distress before they would engage in EDI. Perhaps individuals who have not engaged in EDI had generally experienced greater satisfaction in their primary relationships and transferred this state of satisfaction to the hypothetical EDI scenario. This finding may also converge with those of Roscoe et al. (1988), who asked undergraduates to generate reasons why a person in a serious dating relationship would engage in EDI and found that individuals who had engaged in EDI in the past were more likely to cite relationship dissatisfaction reasons, compared to people who denied a history of EDI. People who have engaged in EDI may focus more on contextual reasons for the EDI (consistent with the fundamental attribution error; Ross, 1977). Predicting generally higher levels of satisfaction with the intimacy provided by the primary partner prior to engaging in EDI appears consistent with the anticipated higher levels of remorse and concern about hurting or losing the primary partner reported by the hypothetical EDI group.

Relative to the group reporting on dating EDI, undergraduates imagining a hypothetical EDI did not differ on the degree of various reasons for the EDI or on their ratings of the type of relationship with the extradyadic partner. Thus, data based on hypothetical involvement appears comparable to data based on actual involvement on the level of autonomy, intimacy, and self-esteem reasons cited for the involvement, as well as on the intensity of feeling for the extradyadic partner.

Overall, our findings suggest that people who have not engaged in EDI are able to answer questions about EDI consistently and in a way that fits with their personality. Thus, surveying this group may be helpful when developing and piloting measures. Hypothesis testing with this group is tentatively supported because many of the established relationships between attachment style and EDI were replicated in the hypothetical EDI group. However, this replication was incomplete, and thus, hypothesis testing with this group should not be considered definitive. Moreover, when asked about attitudes toward EDI, anticipated remorse and distress regarding EDI, and imagined satisfaction with the intimacy in the primary relationship, there were significant differences in the hypothetical EDI group compared to those with actual extradyadic experience. Thus, it cannot be assumed that results from participants asked to report on hypothetical involvement are equal to responses based on actual involvement.

We also found several differences between groups reporting dating and marital EDI. Participants reporting on marital EDI reported significantly higher levels of intimacy and self-esteem reasons for the EDI, as well as motivations for the EDI based on love for the extradyadic partner, relative to participants reporting on a dating EDI. Given the relative seriousness of the infringement of the marital contract represented by an extramarital affair, individuals appear to require higher levels of these types of motivations to engage in an extramarital involvement relative to a dating EDI. The seriousness of the violation of the marital contract represented by an extramarital affair may also underlie the relatively greater concern regarding the negative evaluations of others and the concern about hurting the partner seen in the marital EDI group relative to the dating EDI group. However, there were no differences in overall remorse or fears regarding potentially losing a partner (either the primary or extradyadic partner) for dating and marital EDI groups. The higher levels of motivation for the EDI seen in the marital group may lessen the overall regret and fears of loss in the context of EDI.

Other aspects of the EDI were similar between dating EDI and marital EDI groups. Despite the importance of experimenting with different relationships as a developmental task in young adulthood, the dating EDI group did not report greater autonomy motivations or more casual extradyadic relationships. Participants reporting on dating EDI were also not more likely to report obsessive extradyadic relationships, even with the frequent intensity of adolescent relationships (Glenn & Marquardt, 2001). Overall, there were no differences in the ratings of the emotional connection with the extradyadic partner between dating and marital EDI groups.

We included gender in our analyses to evaluate possible interaction effects between type of EDI and gender on various aspects of the EDI. Several main effects of gender were found which were consistent with prior literature (and some of which replicated prior findings analyzing the dating EDI and marital EDI data separately; Allen, 2001; Allen & Baucom, 2004). Women reported more intimacy and self-esteem reasons for the EDI, endorsed higher levels of feeling that the behavior was unlike them, and felt more concerned about the negative judgments of others subsequent to the EDI, whereas men generally reported more acceptance of EDI, endorsed more satisfaction with the intimacy in the primary relationship prior to the EDI, and described the extradyadic relationship as more casual. Remarkably, only one interaction effect emerged for all aspects of the EDI: women in the undergraduate groups appeared slightly more concerned about hurting their partner, while women in the marital EDI group were less concerned than men about hurting the partner. In the marital context, women's relatively lower satisfaction with the intimacy provided in their marriage may lead to a sense of justification for the involvement and mitigate their concerns for their husbands' feelings.

The differences found between groups in this study have implications for intervention. For example, intimacy and self-esteem reasons were stronger forces in marital as compared to dating EDI. Also, women (relative to men) and the marital EDI group (relative to the dating EDI group) were more likely to say the EDI was inconsistent with their values and general patterns of behavior. Premarital education programs which seek to prevent marital distress and dissolution (commonly associated with EDI; Amato & Previti, 2003) may educate engaged couples regarding theses issues in order to help couples understand that intentions are not completely protective and to underscore the importance of keeping intimacy and validation strong in the marriage.

Future research in this area could explore several other important questions. For example, Harris (2002) explicitly compared participants' own hypothesized reactions to partner EDI to their actual reactions to past partner EDI. Future explorations could similarly compare what individuals imagine about engaging in EDI to actual past experiences of engaging in EDI. We also did not take on the more challenging task of asking undergraduate participants to imagine responses to an event farther out of their experience (e.g., a spouse engaging in infidelity when there are children and several years of marital investment). For example, Hupka and Eshett (1988) asked introductory psychology students to imagine themselves as older and married, with men specifically asked to imagine themselves as a junior executive and women asked to imagine themselves as the wife of a junior executive. The participants were then asked to imagine seeing their wife or husband passionately kissing someone of the opposite sex at an office party attended by many coworkers and to rate the likelihood of having certain affective, cognitive, and physiological responses to this scene. It has not been established how well undergraduate participants can identify with such specific and far-off scenarios. Hupka and Eshett found that women expected lower degrees of distress if the other person was their best friend (as compared to a stranger or some other person), whereas clinical experience suggests that this dual betrayal might, in fact, be the context of the greatest distress.

Overall, our study suggests that even when diverse groups are able to answer questions about EDI with good internal consistency, the actual reported experiences of EDI often differ between groups. Reported experiences of EDI appear to differ based on the context of the EDI, so it may often be inappropriate to generalize findings from one group to another. This is consistent with other findings which have found differences in EDI based on whether the sample was a community marital sample or a marital therapy sample reporting EDI (Glass, 2003). We expected some of the group differences in this study, but others were surprising. Therefore, it is difficult to predict how diverse groups may differ on various aspects of EDI. The context for examining EDI appears critical in trying to unravel this most important, complex, and at times bewildering aspect of interpersonal behavior within an intimate relationship.

Note. This research was supported in part by an American Psychological Association Dissertation Research Award.

Manuscript accepted February 5, 2006

REFERENCES

Allen, E. S. (2001). Attachment styles and their relation to patterns of extradyadic and extramarital involvement. Unpublished doctoral dissertation, University of North Carolina at Chapel Hill.

Allen, E. S., Atkins, D., Baucom, D. H., Snyder, D., Gordon, K. C., & Glass, S. P. (2005). Intrapersonal, interpersonal, and contextual factors in engaging in and responding to extramarital involvement. Clinical Psychology: Science and Practice, 12, 101-130.

Allen, E. S., & Baucom, D. H. (2004). Adult attachment and patterns of extradyadic involvement. Family Process, 43, 467-488.

Amato, P. R., & Previti, D. (2003). People's reasons for divorcing: Gender, social class, the life course, and adjustment. Journal of Family Issues, 24, 602-626.

Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46-76). New York: The Guilford Press.

Buunk, B. P. (1995). Sex, self-esteem, dependency and extradyadic sexual experience as related to jealousy responses. Journal of Social and Personal Relationships, 12(1), 147-153.

DeGenova, M. K., & Rice, F. P. (2005). Intimate relationships, marriages, and families (6th ed.). Boston, MA: McGraw Hill Higher Education.

Drigotas, S. M., Safstrom, C. A., & Gentilia, T. (1999). An investment model prediction of dating infidelity. Journal of Personality and Social Psychology, 77, 509-524.

Edin, K., Kefalas, M. J., & Reed, J. M. (2004). A peek inside the black box: What marriage means for poor unmarried parents. Journal of Marriage and Family, 66, 1,007-1,014,

Erikson, E. (1963). Childhood and society. Second edition, revised and enlarged. New York: W. W. Norton & Co.

Harris, C. R. (2002). Sexual and romantic jealousy in heterosexual and homosexual adults. Psychological Science, 13, 7-12.

Hupka, R. B., & Eshett, C. (1988). Cognitive organization of emotion: Differences between labels and descriptors of emotion in jealousy situations. Perceptual and Motor Skills, 66, 935-949.

Glass, S. E (2003). Not "just friends": Protect your relationship from infidelity and heal the trauma of betrayal. New York: Free Press.

Glass, S. P., & Wright, T. L. (1985). Sex differences in type of extramarital involvement and marital dissatisfaction. Sex Roles, 12, 1,101-1,120.

Glass, S. P., & Wright, T. L. (1992). Justifications for extramarital relationships: The association between attitudes, behaviors, and gender. The Journal of Sex Research, 29, 361-387.

Glenn, N., & Marquadt, E. (2001). Hooking up, hanging out, and hoping for Mr. Right: College women on dating and mating today. New York: Institute for American Values.

Greeley, A. (1994). Marital infidelity. Society, 31, 9-13.

Johnson, R. E. (1970). Some correlates of extramarital coitus. Journal of Marriage and the Family, 32, 449-456

Kline, G. H., Stanley, S. M., Markman, H. J., Olmos-Gallo, P. A., St. Peters, M., Whitton, S. M., et al. (2004). Timing is everything: Pre-engagement cohabitation and increased risk for poor marital outcomes. Journal of Family Psychology, 18, 311-318.

Mongeau, P. A., Hale, J. L., & Alles, M. (1994). An experimental investigation of accounts and attributions following sexual infidelity. Communication Monographs, 61, 326-344.

Mongeau, P. A., & Schulz, B. E. (1997). What he doesn't know won't hurt him (or me): Verbal responses and attributions following sexual infidelity. Communication Reports, 10, 143-152.

Oliver, M. B., & Hyde, J. S. (1993). Gender differences in sexuality: A meta-analysis. Psychological Bulletin, 114, 29-51.

Prins, K. S., Buunk, B. P., & Van Yperen, N. W. (1993). Equity, normative disapproval, and extramarital relationships. Journal of Social and Personal Relationships, 10, 39-53.

Roscoe, B., Cavanaugh, L. E., & Kennedy, C. A. (1988). Dating infidelity: Behaviors, reasons, and consequences. Adolescence, 23, 35-43.

Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10). New York: Academic Press.

Seal, D. W., Agostinelli, G., & Hannett, C. A. (1994). Extradyadic romantic involvement: Moderating effects of sociosexuality and gender. Sex Roles, 31, 1-22.

Shackelford, T. K., & Buss, D. M. (1996). Betrayal in mateships, friendships, and coalitions. Personality and Social Psychology Bulletin, 22, 1,151-1,164.

Sheppard, V. J., Nelson, E. S., & Andreoli-Mathie, V. (1995). Dating relationships and infidelity: Attitudes and behaviors. Journal of Sex and Marital Therapy, 21, 202-213.

Thornton, A., & Young-DeMarco, L. (2001). Four decades of trends in attitudes toward family issues in the United States: The 1960s through the 1990s. Journal of Marriage and the Family, 63, 1,009-1,037.

Tooke, W., & Camire, L. (1991). Patterns of deception in intersexual and intrasexual mating strategies. Ethology and Sociobiology, 12, 345-364.

Wiederman, M. W. (1997). Extramarital sex: Prevalence and correlates in a national survey. The Journal of Sex Research, 34, 167-174.

Wiederman, M. W., & Allgeier, E. R. (1996). Expectations and attributions regarding extramarital sex among young married individuals. Journal of Psychology and Human Sexuality, 8, 21-35.

Wiederman, M. W., & Hurd, C. (1999). Extradyadic involvement during dating. Journal of Social and Personal Relationships, 16, 265-274.

Elizabeth S. Allen

University of Colorado at Denver and Health Sciences Center

Donald H. Baucom

University of North Carolina at Chapel Hill

Address correspondence to Elizabeth Allen, Department of Psychology, University of Colorado at Denver and Health Sciences Center, CB 173, P.O. Box 173364, Denver, CO, 80217; e-mail: [e-mail].

Table 1. Reliability of Selected EEQ Scales Across Samples

Scale                   Number     Dating   Hypothetical   Marital
                        of items    EDI         EDI          EDI

Casual EDI *               5       .83 *        .80         .81 *
Close EDI *                5       .91 *        .88         .92 *
Obsessive EDI *            5       .85 *        .83         .83 *
Close primary              3       .73          .74         .72
Autonomy complaints        5       .92          .91         .88
Intimacy complaints        5       .90          .90         .87
Autonomy reasons *         5       .91 *        .90         .85 *
Intimacy reasons *         5       .90 *        .91         .81 *
Self-esteem reasons *      3       .88 *        .82         .87 *
Love reasons               4       .90          .83         .91
Remorse                    6       .85          .89         .85
Abandon                    4       .85          .84         .79
Hurt partner               4       .78          .81         .84
Disapprove                 2       .77          .78         .82
Unlike me                  3       .81          .83         .83
Justify                    4       .80          .73         .85

* Previously reported in Allen & Baucom, 2004.

Table 2. Bivariate Correlations Between Select EEQ Scales

                              Dating    Hypothetical   Marital
                               EDI          EDI          EDI
Divergent Scales
  Casual EDI/Close EDI       -.70 ***     -.54 ***     -.70 ***
  Casual EDI/Obsessive EDI   -.39 ***       -.11       -.47 ***
  Close primary/autonomy
    complaints               -.46 ***     -.36 ***     -.37 ***
  Close primary/intimacy
    complaints               -.15 **      -.41 ***     -.31 **
Convergent Scales
  Intimacy
    complaints/intimacy
    reasons                  .78 ***      .77 ***      .76 ***
  Autonomy
    complaints/autonomy
    reasons                  .84 ***      .81 ***      .72 ***

** p <.01 *** p <.001 (all two-tailed)

Table 3. Relationship Between Adult Romantic
Attachment Style and Aspects of EDI

                                          F
                        Dating      Hypothetical       Marital
                          EDI            EDI             EDI

Intimacy complaints     9.24 ***       17.17 ***       37.98 ***
Autonomy complaints     3.63 *          3.20 *          7.74 ***
Intimacy reasons        5.93 (a) **     6.37 ***       13.98 (a) ***
Autonomy reasons        4.97 (a) **     2.60            4.85 (a) **

(a) Previously reported in Allen & Baucom, 2004.

* p < .05 ** p < .01 *** p < .001 (all two-tailed)

Table 4. ANOVA of Gender and Type of EDI on
the JER and EEQ Subscale Scores

                    Female         Male
                   n = 352        n = 265

JER                 2.02           2.21
                    (.53)          (.53)
EEQ Scales
  Justify           1.78           2.64
                    (.97)         (1.49)
  Unlike me         5.25           4.68
                   (1.58)         (1.74)
  Remorse           4.63           4.37
                   (1.61)         (1.50)
  Abandon           3.33           3.23
                   (1.78)         (1.64)
  Hurt partner      5.18           5.12
                   (1.56)         (1.50)
  Disapprove        4.66           4.32
                   (1.80)         (1.80)
  Close primary     4.27           4.75
                   (1.50)         (1.37)
  Intimacy          3.29           2.82
    reasons        (1.79)         (1.48)
  Autonomy          3.10           3.14
    reasons        (1.86)         (1.53)
  Self-esteem       4.01           3.55
    reasons        (1.81)         (1.69)
  Love reasons      3.69           3.37
                   (1.96)         (1.78)
  Casual EDI        3.50           4.20
                   (1.55)         (1.61)
  Close EDI         4.17           3.94
                   (1.74)         (1.72)
  Obsessive EDI     2.36           2.27
                   (1.28)         (1.22)

                      Dating      Hypothetical   Marital
                       EDI            EDI          EDI
                     n = 345        n = 159      n = 113

JER                 2.23             1.80          --
                    (.51)            (.46)
EEQ Scales
  Justify           2.28 (a)         1.74 (b)      2.35 (a)
                   (1.29)            (.93)        (1.58)
  Unlike me         4.71 (a)         5.59 (b)      5.10 (c)
                   (1.66)           (1.47)        (1.77)
  Remorse           4.26 (a)         5.22 (b)      4.33 (a)
                   (1.51)           (1.44)        (1.63)
  Abandon           3.20 (a)         3.74 (b)      2.93 (a)
                   (1.72)           (1.73)        (1.60)
  Hurt partner      4.83 (b)         5.58 (a)      5.53 (a)
                   (1.53)           (1.42)        (1.47)
  Disapprove        4.10 (b)         5.10 (a)      4.98 (a)
                   (1.77)           (1.56)        (1.92)
  Close primary     4.42 (a)         4.78 (b)      4.22 (a)
                   (1.46)           (1.39)        (1.54)
  Intimacy          2.94 (a)         3.12 (a,b)    3.51 (b)
    reasons        (1.69)           (1.67)        (1.59)
  Autonomy          3.21             2.99          3.00
    reasons        (1.81)           (1.64)        (1.58)
  Self-esteem       3.72 (a)         3.62 (a)      4.36 (a)
    reasons        (1.74)           (1.59)        (1.99)
  Love reasons      3.38 (a)         3.60 (a,b)    4.00 (b)
                   (1.90)           (1.64)        (2.14)
  Casual EDI        3.86             3.71          3.75
                   (1.65)           (1.50)        (1.67)
  Close EDI         3.98             4.08          4.33
                   (1.47)           (1.52)        (1.96)
  Obsessive EDI     2.32             2.45          2.15
                   (1.29)           (1.18)        (1.24)

                               F

                                      EDI
                     Gender           Type
                   (df = 1,611)   (df = 2,611)

JER                12.78 ***       80.25 ***

EEQ Scales
  Justify          57.56 ***       11.95 ***

  Unlike me        17.95 ***       16.24 ***

  Remorse           2.12           20.74 ***

  Abandon            .11            9.05 ***

  Hurt partner       .02           17.40 ***

  Disapprove        6.17 *         23.00 ***

  Close primary    19.92 ***        6.64 **

  Intimacy         13.30 ***        6.68 **
    reasons
  Autonomy           .03             .92
    reasons
  Self-esteem      13.72 ***        8.89 ***
    reasons
  Love reasons      3.66            5.82 **

  Casual EDI       28.68 ***         .74

  Close EDI         2.12            2.26

  Obsessive EDI      .23             .15

* p < .05 ** p < .01 *** p <.001
(all two-tailed)

Note. For EDI type, means within each row
whose subscripts differ are different at
p < .05. For analyses on JER, df = 1,484.