Romantic infidelity is accompanied by various costs. Infidelity is a leading cause of separation and divorce (Allen and Atkins 2012). Those who suspect or discover their partner’s infidelity may suffer from major depression, anxiety, and relationship dissatisfaction (Betzig 1989; Cano and O’Leary 2000). Individuals who commit infidelity may contract sexually transmitted diseases from an extra-pair partner, thereby placing their regular partner at risk of contracting those diseases. A man whose regular partner commits sexual infidelity is at risk of investing resources into offspring to whom he is genetically unrelated.

Given the costs of partner infidelity, individuals perform mate retention behaviors to reduce the likelihood of partner infidelity. A widely used mate retention instrument is the Mate Retention Inventory (MRI; Buss 1988; Buss et al. 2008), which measures performance frequencies of specific mate retention behaviors. Behaviors assessed by the MRI include those predicted and found to be sexually differentiated, according to evolutionary psychological theory. For example, men more than women display material resources and deploy violence to retain a mate, whereas women more than men enhance their physical appearance and use sexual inducements to retain a mate (Buss 1988; Buss and Shackelford 1997). Virtually, all (102 of 104) behaviors included in the MRI are individual-level behaviors—behaviors that individuals perform alone (i.e., without assistance from others).

Guided by social exchange theory, Stafford and Canary (1991) developed the relational maintenance strategy measure—an inventory of relational maintenance behaviors that focus on maintaining relationship stability through coordinated efforts between partners. Behaviors in this inventory focus on cooperation (e.g., “share in the joint responsibilities that face us”) and relationship equity (e.g., “help equally with tasks that need to be done”). One of the five tactics in this inventory—network—includes behaviors that use social networks to maintain relationship stability (e.g., “like to spend time with our same friends,” “focus on common friends and affiliations”). However, Stafford and Canary did not explore specific behaviors that friends perform to assist in promoting relationship stability. The current research investigates specific behaviors that comprise coalitional mate retention (CMR), defined here as mate retention with assistance from allies (i.e., friends).

Documented Cases of Coalitional Mate Retention

Nonhuman animals that recruit allies for CMR can overcome limitations of individual mate retention. For example, in lions, a coalition of males is more successful than a single male at defending sexual access to a female against rivals (Grinnell et al. 1995; Schaller 1972). Macaque males may form coalitions to overpower higher-ranking males to secure sexual access to fertile females (Bissonnette et al. 2011). Males of other species also sometimes form alliances with other males to defend sexual access to females as has been documented in cheetahs (Caro 1994), chimpanzees (Watts 1998; Nishida and Hosaka 1996), and dunnocks (Davies 1992).

Historically, humans have practiced CMR. For example, kings recruited guards to protect their harem (Lad 2011). The inclusion of two items in the MRI—“had my friends check up on my partner” and “got my friends to beat up someone who was interested in my partner”—suggests that humans sometimes perform CMR (Buss 1988; Buss et al. 2008). The inventory developed by Stafford and Canary (1991) provides an indirect evidence of CMR: Maintaining social networks shared by romantic partners is important for maintaining stable relationships. Individuals whose partner has sex with their friend—relative to with a stranger or with an enemy—feel most betrayed by their partner (Shackelford and Buss 1996), providing indirect evidence that individuals understand that one should not pursue a friend’s romantic partner. Indeed, trustworthiness is important for maintaining friendships (Cottrell et al. 2007; Vigil 2007). Unlike some nonhuman species in which males cooperatively breed with females (see Watts 1998), humans do not typically tolerate allies having sex with their partner.

Benefits of Coalitional Mate Retention

CMR can overcome the costs of excessive individual mate retention. Individuals who perceive their partner to be overbearing, excessively jealous, and mistrustful (i.e., they perceive that their partner performs excessive mate retention) also report greater relationship dissatisfaction and sometimes terminate the relationship as a result (Guerrero et al. 1995). Individuals can reduce their partner’s perception of excessive mate retention by employing CMR. Individuals who use CMR can decrease the likelihood of their partner’s infidelity without appearing overbearing to their partner. For example, individuals may recruit allies to accompany their partner to a social gathering to ensure that their partner does not flirt with others, while simultaneously appearing trusting to their partner because they are absent.

CMR can overcome time (e.g., restrictive work schedules) and geographic (e.g., long-distance relationships) restrictions of individual mate retention. For example, individuals may be less likely to commit infidelity during a trip away from their partner if, during that trip, they are accompanied by their partner’s friends. Similar to vigilance-related behaviors in the MRI, individuals can employ CMR to perform proxy vigilance through their allies.

Individuals can boost their mating desirability to their partner with assistance from allies, thereby decreasing the likelihood of their partner’s infidelity. Individuals mated to a less (versus more) desirable partner are more likely to terminate the relationship or commit infidelity (Shackelford and Buss 1997). Individuals can ask their friend to advertise their mating qualities for them (e.g., your friend says positive things about you), thereby avoiding the costs of appearing narcissistic if they bragged about themselves and validating those advertisements by providing corroborative evidence from friends.

Costs of Coalitional Mate Retention

CMR has potential costs. Those who recruit a same-sex ally for CMR risk their ally poaching their partner—luring their partner away or having sex with their partner. Kings were aware of potential mate poaching by guards and, therefore, recruited eunuchs—instead of sexually functional men—to police their harem (Lad 2011). Two items included in the MRI—“told my same-sex friends how much my partner and I were in love” and “refused to introduce my partner to my same-sex friends”—indicate that individuals are aware that same-sex friends are potential mate poachers.

Those who recruit an opposite-sex ally for CMR risk being the target of their ally’s mate poaching (Mogilski and Wade 2013). The majority of men, and some women, are sexually attracted to their opposite-sex friends (Kaplan and Keys 1997) and report uncertainty about sexual boundaries with those friends (Swain 1992). Individuals sometimes befriend opposite-sex friends with the explicit intention of mate poaching (Schmitt and Buss 2001). Some men report that “there wasn’t enough sex in our relationship” as a reason for ending an opposite-sex friendship (Bleske-Rechek and Buss 2001).

Overview of the Current Research

Despite substantial evidence that friends play important roles in helping to maintain romantic relationships (Canary and Stafford 1992; Stafford and Canary 1991), research has yet to investigate how friends assist in CMR. Here, we develop and provide initial psychometric validation of the Coalitional Mate Retention Inventory (CMRI). In study 1, participants provided open-ended responses in which they nominated CMR behaviors. In study 2, a new sample of participants provided frequency reports on the CMR behaviors identified in study 1.

Study 1

Method

Participants and Procedure

We recruited 100 participants from Facebook. Participants were at least 18 years of age. Participation was anonymous and unrewarded.

Following previous act nomination procedures (Buss and Craik 1983), participants reported up to ten acts that they (or someone they know) asked their friend or relative to do to prevent their partner from becoming interested in someone else. Participants recorded their responses on an online survey host.

Results

We collected 190 act nominations. We eliminated vague acts (e.g., “asked my friend to help with my relationship”) and irrelevant acts (e.g., “I am never jealous”) to construct an initial 44-item CMR Inventory for further investigation in study 2 (complete list of acts is available upon request).

Study 2

Method

Participants

We recruited from Amazon’s Mechanical Turk (MTurk) 387 participants (176 women) in a committed, heterosexual relationship lasting at least 1 year. The mean participant age was 32.1 years (SD = 9.1), and the mean relationship length was 66.0 months (SD = 88.5). Participants reported on interactions with two friends (one man and one woman, see below), each of whom they considered as a good friend, and each of whom they have known for at least 1 year. The mean length of the friendship was 88.7 months (SD = 90.2) with the male friend and 76.6 months (SD = 89.6) with the female friend. We implemented MTurk filters recommended by Peer et al. (2013). MTurk workers who could access had successfully completed at least (1) 500 MTurk jobs and (2) 95 % of their jobs.

Materials

In line with Simmons et al. (2011), we fully disclose all variables. Participants reported their age, sex, and romantic relationship length. Participants reported the number of miles they lived from their partner, and how often they have sex with their partner during a typical week. Participants completed the Mate Retention Inventory Short Form (MRI-SF), in which they reported on a four-point scale (0 = never, 1 = rarely, 2 = sometimes, and 3 = often) how often they performed 38 mate retention behaviors during the past year (Buss et al. 2008). Participants also completed the HEXACO-60, which secures scores on personality traits (Ashton and Lee 2009).

Participants were instructed to think of one heterosexual man and one heterosexual woman, each of whom they considered to be good friend, and each of whom they have known for at least 1 year. For the remainder of the survey, participants answered each question twice: once for their male friend and again for their female friend.

Participants reported their friend’s age, the length of their friendship (in months), whether their friend was in a committed or romantic relationship, and whether their friend was friends with their partner. Participants completed the McGill Friendship Questionnaire (Mendelson and Aboud 1999), which assesses friendship quality. Participants completed the 44-item initial CMRI, in which they reported on a four-point scale (0 = never, 1 = rarely, 2 = sometimes, and 3 = often) (1) how often they requested their friend to perform each behavior during the past year and (2) how often they believed their friend performed that behavior during the past year. These two frequency reports are not necessarily related. For example, individuals can frequently request a CMR behavior, but their friend never performs the behavior (i.e., their friend disregards the CMR request), or individuals might never request a CMR behavior, but their friend frequently performs the behavior (i.e., their friend performed an unrequested CMR behavior). Two attention check questions were incorporated into the CMRI: one question asked participants to respond “never” and the other question asked participants to respond “often.”

Procedures

Potential participants viewed an advertisement for this study on MTurk’s job listings. Those interested in participating were provided a link to a consent form. Those who electronically signed the consent form could access the survey, and those who did not sign were exited from the study. Participants who correctly answered the attention check questions were compensated $4.00. Responses from participants who failed the attention check questions (31 of 387 or 8.0 %) were excluded from analyses.

Results

Participants reported for each CMR behavior how often they requested that behavior from (1) their male friend (α = 0.97 across all items) and (2) their female friend (α = 0.97) and the frequency with which that behavior was performed by (3) their male friend (α = 0.96) and (4) their female friend (α = 0.96). Because participants and their friends were heterosexual, some CMR behaviors only applied to same-sex friends (e.g., “seduced my partner to help me test my partner’s faithfulness to me” might only apply to same-sex friends). Thus, we used each participant’s sex to generate CMR behavior request frequencies for (5) their same-sex friend (α = 0.97) and (6) their opposite-sex friend (α = 0.97) and CMR behavior performance frequencies for (7) their same-sex friend (α = 0.95) and (8) their opposite-sex friend (α = 0.96). In total, there were eight versions of the CMRI.

We conducted principal component analyses (PCAs, followed by direct oblimin rotation) on each of the eight versions of the CMRI. For each version, we extracted and retained components with an eigenvalue greater than 1.0 (Kaiser-Guttman criterion; see Jackson 1993). The PCA for each of the eight versions produced five or six components. After comparing PCA results from all eight versions, the male friend’s CMR performance version provided a useful starting point for constructing the CMRI because this version produced the most interpretable components (see Table 1 for component loadings and labels for male friend’s CMR performance; analyses of all eight PCAs are available upon request). For this version, each item had a component loading of at least 0.30 on at least one component. Because the development of the CMRI is in its early stages, we adopted a liberal loading requirement (0.30) and, therefore, none of the 44 items were eliminated.

Table 1 Component loadings on CMR performed by male friend (principal components analysis, direct oblimin rotation)
Full size table

First, we assigned each item to the component onto which it loaded most highly. Then, we moved a few items from one component to another because of their conceptual resemblance with items in the new components. We moved “included my partner in group activities” and “studied with my partner” to component 4 (monopolizing time). We moved “followed my partner around” to component 3 (vigilance). We moved “reassured my partner that I liked my partner,” “listened to my partner’s relationship concerns,” and “told my partner that cheating was wrong” to component 5 (therapy). There is a compelling theoretical reason to group violence-related CMR behaviors into a unique component: Men more than women employ violence as mate retention (Buss 1988; Buss et al. 2008; Buss and Shackelford 1997). Thus, we moved three items to a new violence component: “fought someone who was interested in my partner,” “hit someone who was flirting with my partner,” and “intimidated someone who was interested in my partner” (see Appendix for item placements in the final CMRI).

Components of Coalitional Mate Retention

The manipulation component includes behaviors in which an ally deceives the partner into admitting (e.g., “got my partner drunk to see what my partner said”) or demonstrating (e.g., “tried to ‘hook up’ my partner with someone else to see what my partner did”) the partner’s propensity to commit infidelity. The praise component includes behaviors in which an ally says positive things to the partner and to others (e.g., “said nice things about me when my partner and other people were around”), thereby increasing the romantic partnership’s desirability. The vigilance component includes behaviors in which an ally watches the partner’s behavior (e.g., “observed how my partner acted around people interested in my partner”). The monopolizing time component includes behaviors in which an ally spends time with the partner (e.g., “accompanied my partner to a party”). The therapy component includes behaviors in which an ally strengthens the romantic partnership by repairing relationship problems and listening to relationship concerns (e.g., “told my partner how much I liked my partner”). The gifts component includes behaviors in which an ally secures information about appropriate gifts for the partner (e.g., “told me what gifts my partner wanted”). The violence component includes behaviors in which violence is employed against potential mate poachers (e.g., “hit someone who was flirting with my partner”).

Relationships between MRI Tactics and CMRI Tactics

It is not clear whether individuals who employ more CMR will also employ more individual mate retention (i.e., concurrent strategies) or less individual mate retention (i.e., compensatory strategies). If CMR functions primarily to overcome constraints associated with individual mate retention, then individual mate retention and CMR may be concurrent strategies because individuals may be attempting to exhaust possible strategies to reduce the likelihood of partner infidelity or relationship defection. If CMR functions primarily to disguise jealousy and mistrust, then individual mate retention and CMR may be compensatory strategies because individual mate retention, not CMR, overtly demonstrates jealousy.

We investigated the relationship between these two classes of mate retention strategies. First, we computed tactic scores by computing the mean scores across items within each tactic from the MRI-SF and from the CMRI. Then, we correlated each tactic score from the MRI-SF with each tactic score from the CMRI. The CMRI and MRI-SF use identical four-point scales (never, rarely, sometimes, and often), and both measure performance frequencies over the past year (see Table 2). Of the 72 significant correlations, 71 were positive, suggesting that individuals who deploy more individual mate retention also deploy more CMR.

Table 2 Zero-order correlations between CMRI tactics (columns) and MRI tactics (rows)
Full size table

Sex Differences in Coalitional Mate Retention

Table 3 presents the results of tests for sex differences in CMR performance frequencies. Men (relative to women) more frequently recruit their female friends to determine appropriate gifts for their partner. Female friends (relative to male friends) more frequently provide relationship therapy. Unexpectedly, both men and women more frequently recruit their female friend (relative to their male friend) to deploy violent CMR (e.g., hitting, fighting, and intimidating potential poachers). To verify that our sample was not unusual, relative to samples recruited for other mate retention studies, we investigated whether the sexes differ in the violence tactic from the MRI-SF (complete analyses of sex differences for all MRI-SF tactics are available upon request). Consistent with previous research (Buss 1988; Buss and Shackelford 1997), men (M = 1.25, SD = 0.52) relative to women (M = 1.13, SD = 0.42) more frequently perform individual mate retention involving violence—t(348) = 2.18, p < 0.05. Thus, despite the unexpected sex difference in violent CMR, our sample is similar to samples from previous mate retention studies.

Table 3 Mean performance frequencies (standard deviations) of each CMR tactic, separated by participant’s sex and by friend’s sex
Full size table

Discussion

The current research aimed to develop and provide initial psychometric validation for an inventory that assesses coalitional mate retention (CMR) using a balance of data- and theory-driven methods. The CMRI includes 44 behaviors that individuals deploy with assistance from allies to reduce the likelihood of partner infidelity or relationship defection. The components extracted from the CMRI range from negative, cost-inflicting behaviors (e.g., manipulation) to positive, benefit-provisioning behaviors (e.g., therapy).

Because this is a new inventory, replication of the current research is especially warranted. We recommend that future research recruit participants using diverse recruitment methods. Buhrmester et al. (2011) found that participants comprising MTurk samples differed demographically (e.g., age, ethnicity) from participants comprising standard internet samples or college samples. Future research could replicate the current research using different recruitment methods, including in-person surveys.

We suspect potential cultural differences in CMR. For example, adolescents who self-identify as more collective (versus those who identify as more individualistic) are closer to their same-sex friends and are more sensitive to those friends’ needs (Verkuyten 1996), suggesting that unrequested same-sex CMR may be deployed more frequently in collectivist cultures than in individualistic cultures. Future research could recruit multicultural samples to explore cultural differences in CMR.

We did not investigate CMR with assistance from relatives. Relatives—compared to friends—can sometimes generate stronger alliances. For example, when forming coalitions, individuals are more likely to receive help from relatives than from non-relatives (Ackerman and Kenrick 2008). Thus, individuals who assist relatives receive benefits resulting from genetic relatedness. Additionally, CMR with relatives—compared to CMR with friends—carries different costs. For example, it is less likely that an individual would sabotage their opposite-sex relative’s relationship—relative to their opposite-sex friend’s relationship—because of romantic feelings for the target. Furthermore, relatives—compared to friends—may perform more unrequested CMR. CMR may be perceived more explicitly as a favor with friends than with relatives (i.e., reciprocal altruism; Trivers 1971), which may motivate individuals to perform more unrequested CMR for relatives than for friends. Future research might investigate differences in CMR involving friends versus relatives.

The CMRI may provide an important tool for research on long-distance, romantic relationships. In the USA, as many as one million people annually are in a long-distance relationship, up to 50 % of first-year college students are in a long-distance relationship, and 10 % of job relocations cause romantic relationships to become long-distance relationships (reviewed in Aylor 2003). Additionally, work schedules and career choices are causing more couples to live apart (Armour 1998; reviewed in Aylor 2003). The MRI might produce uninterpretable results for such logistically constrained relationships. For example, two of the mate retention tactics assessed by the MRI are monopolizing a partner’s time and remaining vigilant about a partner’s whereabouts. However, individuals who spend time apart from their partner because of logistics (e.g., long-distance relationships, work schedules, business trips) would score lower on these two tactics, but not necessarily because they lack motivation to perform mate retention. In fact, the time individuals spend apart from their partner is the time they cannot account for their partner’s behaviors and is, therefore, the time during which there may be greatest motivation to perform mate retention (Dainton and Aylor 2001; McKibbin et al. 2011; Pham and Shackelford 2013; Shackelford et al. 2007, 2002).

This research explores the intersection of two ubiquitous features of human life: mate retention behavior and coalitional behavior. Interest in mate retention behavior has penetrated several research domains, including personality psychology (McKibbin et al. 2014), sexual behavior (Goetz et al. 2005), and ovulatory cycling (Gangestad et al. 2002). The current research adds to this literature by highlighting the importance of mate retention with assistance from allies.