Sociosexuality, Commitment, Sexual Infidelity, and Perceptions of Infidelity: Data From the Second Love Web Site

This study explored the association between sociosexuality (behavior, attitudes, and desire) and commitment to understand sexual infidelity in current relationships. We also explored how these variables were associated with attitudes toward infidelity. Participants were romantically involved heterosexuals (N = 252; 51 women, 201 men; Mage = 41.13, SD = 9.82) registered on a dating Web site directed at romantically involved individuals. Results showed that sociosexuality, alongside commitment, was associated with sexual infidelity. Individuals who had (versus had not) previously engaged in infidelity reported more unrestricted sociosexuality, while reporting less commitment. For individuals reporting prior sexual infidelity, unrestricted sociosexual desire and lesser commitment were associated with more permissive perceptions of infidelity. For individuals reporting no prior sexual infidelity, greater commitment was always associated with more strict perceptions of infidelity. No gender differences emerged in the analyses. Also, no differences were found according to individual motivations (i.e., looking to know other people versus seeking casual sex) or relational motivations (i.e., individual registration versus registration as a couple) underlying individuals’ registration on the Web site. These results are an important addition to the literature on infidelity by analyzing a specific sample motivated to engage in infidelity. Implications for future research are discussed.

of mating strategies, however, suggest that individuals also accommodate their predispositions and needs to attain relational goals (strategic pluralism theory; Gangestad & Simpson, 2000). Indeed, empirical evidence shows that individuals tend to restrict their sociosexuality when in a relationship (Penke & Asendorpf, 2008). Following theories of commitment (investment model; Rusbult, Agnew, & Arriaga, 2012), the motivation to remain in the relationship (Arriaga & Agnew, 2001) is one of the crucial factors that helps protect the relationship (Drigotas, Safstrom, & Gentilia, 1999). In this sense, more committed individuals are less likely to engage in infidelity, regardless of their sociosexuality .
In the present study we analyzed the relationships between sociosexuality, commitment, and sexual infidelity, departing from the typical sample of college students or adult mainstream couples. We extend the literature by examining a sample of older individuals who are currently involved in a romantic relationship but at the same time have paid for registration on a Web site that promotes interactions with other romantically involved individuals: Second Love. This sample offers a unique context for studying relationship processes, because these individuals are more predisposed to engage in infidelity behaviors. Researchers have previously argued for the importance of understanding how interactions over the Internet influence romantic relationships (Whitty, 2003), especially because online infidelity is perceived to be as serious as face-to-face infidelity (Merkle & Richardson, 2000). Recent studies extended these findings to online extradyadic involvements over social networking Web site, such as Facebook (e.g., Martins et al., 2016). Nevertheless, research on infidelity behaviors using more specific Web sites such as Second Love is much scarcer in the literature. Hence, it is important to further examine whether sociosexuality and commitment are associated with sexual infidelity in this specific sample and whether these variables are associated with perceptions of infidelity.

Sociosexuality in Human Sexual Behaviors
Sociosexuality is argued to comprise three distinct components: (a) behavior, referring to behavioral tendencies and personal histories of uncommitted sex; (b) attitudes, referring to the evaluative disposition toward uncommitted sex, influenced by socialization; and (c) desire, referring to the interest in uncommitted sex often associated with sexual arousal and sexual fantasies (Penke & Asendorpf, 2008).
However, individual sexuality is influenced by relational and contextual variables, such as social and sex roles in society (Bronfenbrenner & Morris, 2006;Schaller & Murray, 2008;Schmitt, 2005). Hence, a broader understanding of sexuality must integrate biological, psychological, and sociocultural factors, both at individual and relational levels (DeLamater & Hyde, 2004). For instance, although individual attitudes toward sex help understand dyadic sexual behavior and sexual satisfaction (DeLamater & Sill, 2005), attitudes toward extradyadic uncommitted casual sex are not necessarily associated with sociosexual behavior (Penke & Asendorpf, 2008). For instance, an individual with more unrestricted sociosexual attitudes may be more likely to experience greater desire for a variety of potential sexual partners, but at the same time may not be able to act upon such predisposition (e.g., lack of opportunity) or may be constrained by cultural factors (e.g., to preserve a social image). This converges with evidence showing that attitudes are not always a reliable direct predictor of behavior (Ajzen & Fishbein, 2005). In sum, these findings show intragender differences in sociosexuality and how they relate with sexuality. For our present purposes, we now focus on the role of sociosexuality in the course of romantic relationships.

Sociosexuality and Romantic Relationships
For romantically involved individuals, differences in sociosexuality seem to be reflected in relationship quality, such that more unrestricted (versus restricted) individuals tend to develop shorter and less committed relationships (Jones, 1998;Simpson & Gangestad, 1991;Simpson et al., 2004). They are also more likely to engage in infidelity (Seal, Agostinelli, & Hannett, 1994) and to perceive infidelity as more acceptable, albeit only under certain circumstances (e.g., in the context of a bad relationship; Feldman & Cauffman, 1999). Regardless, research also shows that sociosexually unrestricted individuals are motivated to develop stable romantic relationships (Simpson & Gangestad, 1991;Simpson et al., 2004) and are sexually invested in their relationships (Tempelhof & Allen, 2008). Therefore, an unrestricted sociosexual orientation is not a sufficient condition to engage in sexual infidelity.
Research indicates that infidelity is associated with aspects pertaining to the relationship (McAnulty & Brineman, 2007;Shackelford & Buss, 1997). Emotional infidelity tends to arise out of feelings of dissatisfaction RODRIGUES, LOPES, AND PEREIRA and neglect, whereas sexual infidelity arises out of feelings of sexual attraction, need for sexual variety, and wanting more frequent sex (Allen & Rhoades, 2007;Allen et al., 2008;Barta & Kiene, 2005;Feldman & Cauffman, 1999). This clearly illustrates the importance of relational factors and their interplay with dispositional factors to predict infidelity. An individual may have more unrestricted sociosexual attitudes and/or desires but may decide not to engage in casual sexual behavior due to a lack of motivation do to so. This converges with a pluralistic view of mating strategies, according to which individuals accommodate personal motivations and needs to broader contextual conditions and strategically shift their mating strategies to attain specific goals (Gangestad & Simpson, 2000).
In this line of reasoning, commitment, defined as longterm motivation to maintain the relationship (Arriaga & Agnew, 2001), sheds a light on understanding the sociosexuality-infidelity link. On one hand, sexual infidelity is usually associated with negative consequences for the long-term maintenance of relationships, such as resulting in breakup or divorce (Sharpe, Walters, & Goren, 2013;Vangelisti & Gerstenberger, 2004). On the other hand, commitment reliably predicts happiness, sexual adjustment, intimacy, couple well-being (Rusbult, Martz, & Agnew, 1998), fidelity (Drigotas et al., 1999), and relationship persistence (Le & Agnew, 2003). As such, commitment may act upon sociosexuality to promote relationship maintenance. Supporting this argument, research shows that individuals restrict their sociosexuality when initiating a new relationship and become more unrestricted when ending it (Penke & Asendorpf, 2008). Presumably, romantically involved individuals consider both personal and relational consequences of engaging in infidelity and compare them to possible short-term benefits (Drigotas & Barta, 2001). More (versus less) committed individuals are more likely to transform individualistic needs for casual sex into common motivational goals and needs (Drigotas et al., 1999) and activate derogation of alternatives (Lydon, Fitzsimons, & Naidoo, 2003).
Relatively little research has examined the interplay between sociosexuality and commitment. The existing research suggests that commitment does indeed play a crucial role in accommodating individuals' sociosexual orientation when in a highly committed relationship. For example, Mattingly and colleagues (2011) showed that unrestricted individuals were less likely to engage in infidelity when more (versus less) committed. However, participants in the aforementioned study were relatively young (M age = 19.2 years) and in a relationship for a mean length of 17 months. The present study departed from this sample and examined an older sample of individuals, romantically involved in longer relationships, with more diverse characteristics (e.g., cohabitation, children), and registered on a Web site directed at romantically involved individuals.

Sociosexuality and Perceptions of Infidelity
Perceptions of infidelity are important for understanding infidelity in romantic relationships. Individuals have distinct definitions of infidelity, including which behaviors are indicative of it. More often than not, definitions of infidelity are not explicitly discussed within the couple and rely on implicit agreements of what is, and what is not, acceptable (McAnulty & Brineman, 2007). For instance, individuals for whom certain behaviors are not indicative of infidelity (e.g., talking over the Internet) may be more likely to engage in those behaviors because they will not perceive them as being harmful for the relationship. Research supports this reasoning. Individuals with (versus without) a past history of sexual infidelity are more approving of infidelity (Tsapelas, Fisher, & Aron, 2010) and are more accepting of another person's infidelity (Sharpe et al., 2013).
Broadly, behaviors associated with infidelity can range from flirting to sexual intercourse (Roscoe, Cavanaugh, & Kennedy, 1988) and can be categorized into three categories: (a) ambiguous behaviors, such as talking on the Internet or hugging another person; (b) explicit behaviors, such as sexual intercourse or oral sex; and (c) deceptive behaviors, such as lying to, or withholding information from, the relationship partner (Mattingly et al., 2010;Wilson et al., 2011). Ambiguous behaviors are perceived as the least indicative of infidelity; deceptive behaviors as moderately indicative of infidelity; and explicit behaviors as the most indicative of infidelity (Feldman & Cauffman, 1999;Wilson et al., 2011).
Research shows that dispositional and relational factors play a vital role in these perceptions of infidelity. More restricted sociosexuality and greater satisfaction in the relationship are associated with more restrictive perceptions of what behaviors constitute infidelity (Mattingly et al., 2010). Greater commitment is associated with the perception of extradyadic sexual behaviors as more indicative of infidelity and as provoking greater jealousy (Yarab, Allgeier, & Sensibaugh, 1999). Still, only few studies focused on how these perceptions differ according to sexual infidelity in sexually monogamous relationships, and to the best of our knowledge none have examined whether these differences are associated with sociosexuality and commitment.

Overview of the Study
In the present article we aimed at analyzing the role of sociosexuality and commitment to understand sexual infidelity and perceptions of infidelity. This study used a sample of sexually monogamous men and women who were also registered on Second Love, a dating Web site directed at promoting interactions with other individuals in romantic relationships. According to its 2015 brand profile, Second Love was created to provide a dating platform for likeminded people to explore the possibilities outside of their current relationships in a safe and discrete way. Users can register and create an active profile for free, with an option SOCIOSEXUALITY, COMMITMENT, AND SEXUAL INFIDELITY to pay for an upgrade to a premium registration. This acts as a protocol to protect the anonymity and privacy of members. All profiles are individually screened for acceptance into the database to ensure that the participating members are not there to abuse or compromise others.
We decided to use this specific sample because it provides insights regarding infidelity with individuals already motivated to engage in some form of infidelity. Although these individuals have not necessarily engaged in sexual infidelity, they have already engaged in infidelity by paying for their registration. Very little research has examined relational dynamics within such samples. Interactions over the Internet are increasingly popular, and researchers need to understand how these interactions influence the initiations and maintenance of romantic relationships (Whitty, 2003). Online infidelity has more recently become a focus of interest for researchers (e.g., Martins et al., 2016). Individuals perceive online infidelity to be as intimate and real as faceto-face infidelity, and both types of infidelity are likely to have similar consequences for the relationship (Merkle & Richardson, 2000). Therefore, this type of evidence has a high demand, and these samples are highly informative for a broader grasp of infidelity. To promote reliable sexuality and infidelity reports, this was an anonymous online study. Previous research has shown that individuals are less likely to be influenced by social desirability in these situations (Alexander & Fisher, 2003).

Objectives and Hypotheses
The first aim of this study was to examine sociosexuality and commitment according to actual sexual infidelity. Most research conducted on infidelity analyzed responses to hypothetical imagined scenarios and not actual infidelity experiences (Sharpe et al., 2013). In our research we decided to divide the sample according to prior sexual infidelity based on the notions that (a) all participants reported being sexually monogamous in their relationship, (b) extradyadic behaviors should be perceived as infidelity according to a monogamous norm, and (c) all participants were actively engaged in some type of infidelity by being registered on Second Love.
The second aim of the present study was to explore differences in perceptions of what constitutes infidelity. Little research has focused on such perceptions or examined sexual infidelity as an influential factor (Mattingly et al., 2010). Past research suggests that men tend to engage in more infidelity than women (Blow & Hartnett, 2005b), although gender differences are becoming less pronounced (Barta & Kiene, 2005;Brand, Markey, Mills, & Hodges, 2007). Nevertheless, relational factors must be taken into account. Empirical evidence shows that more committed individuals tend to accommodate their sociosexuality (Penke & Asendorpf, 2008) and focus on relational motivations to preserve their relational well-being (Drigotas et al., 1999). Given our specific sample, we expected no differences between men and women in regard to sexual infidelity. Instead, we expected sociosexuality and commitment to be associated with sexual infidelity. Also, participants with a history of sexual infidelity should report more unrestricted sociosexuality (behavior, attitudes, and desire) and lesser commitment than participants without a history of sexual infidelity.
Based on the few studies that focused on infidelity perceptions (Mattingly et al., 2010;Wilson et al., 2011), explicit behaviors should be more indicative of infidelity, followed by deceptive behaviors. Ambiguous behaviors should be the least indicative of infidelity. Individuals with (versus without) sexual infidelity should hold more permissive perceptions of what constitutes infidelity (Sharpe et al., 2013;Tsapelas et al., 2010). For these individuals, their unrestricted sociosexuality should be associated with more permissive perceptions of ambiguous, explicit, and deceptive infidelity behaviors. For individuals without sexual infidelity, on the other hand, greater commitment should be associated with perceptions of greater infidelity Yarab et al., 1999).
Given the specificity of our sample, we also explored differences in sociosexuality, commitment, and perceptions of infidelity according to the motivations of participants for being registered on Second Love: individual (i.e., looking to get to know other people versus looking for casual sex) and relational (i.e., individual registration versus registration as a couple).
All participants were romantically involved in what they identified as a sexually monogamous romantic relationship and were not looking for a new relationship. All participants paid for their registration on the Second Love Web site. Most participants reported sexual infidelity during the current relationship (68.3%) and indicated that they registered on the Web site individually without the partner being aware of it (86.5%). Men were more likely to indicate they were looking for casual dyadic sexual encounters, whereas women were more likely to indicate they were looking to meet other people. No other significant differences across gender were found (Table 1).

Procedure
This study involved human data collection from healthy adult volunteers, in agreement with the ethics guidelines issued by the scientific commission of the hosting institution. There were no physical, financial, social, legal, or other risks connected with the study. The study was noninvasive, no false information or deception was employed, and results were analyzed anonymously. Participants were given a full description of their rights and duties at the beginning of the Web survey.
Participants were recruited through a Web survey hosted on the Second Love Web site and available to all registered users. When users logged in to their personal area on the Web site, a popup window appeared informing them of a Web survey on interpersonal relationships. Should they consider taking part in the survey, a link was included so that participants could be redirected to the study itself. This popup window was visible only to individuals who indicated being heterosexual in their initial registration on Second Love. Ethical considerations, such as confidentiality and anonymity of responses, were explicit in the introductory page. All individuals were informed (a) of the general purpose of the study and the nature of the task involved; (b) that neither their name nor any identifying information were attached to their data; (c) that their participation was voluntary; and (d) that they could withdraw from the study at any time without their responses being considered for analysis. After providing informed consent by clicking on the "I agree" option, participants were directed to the first part of the survey, which included demographic, relationship, and registration information. The second part included our main study variables.
This was a non forced-response Web survey. When applicable, a warning reminded individuals of missing responses, but they were allowed to continue their participation (missing responses: 0.42%). At the end, participants were debriefed about the purpose of the study and were provided with contact information. The average time to complete the survey was 11 minutes. Participants were not paid or given other incentives to participate in the study. Internet protocol (IP) addresses were checked, and no IP corresponded to more than one questionnaire.

Measures
Demographic, Relationship, and Registration Information. The first part of the Web survey started with standard sociodemographic information (gender, age, sexual orientation, education, area of residence, religion). This was followed by relationship factors (relationship status, cohabitation, relationship length, children). Some of these variables were identified in the literature as associated with sexual infidelity (Jackman, 2015;Mark, Janssen, & Milhausen, 2011;Martins et al., 2016;Shaw et al., 2013;Tsapelas et al., 2010) and with perceptions of infidelity (Mattingly et al., 2010). Sexual infidelity was assessed by this question: "Were you involved sexually with other people during your current romantic relationship?" (Yes/ No) (see also Shaw et al., 2013). If they answered Yes, participants were then asked: "Does your partner know of this sexual involvement?" (Yes/No). All participants were additionally asked: "What type of relationship do you have with your partner?" (Sexually monogamous/Individual casual sexual encounters are consented/Open relationship) and "Are you currently looking for a new romantic relationship? (Yes/No). Finally, participants were asked to provide information related to their registration (type of registration, partner awareness of registration, and what they were looking for).
Revised Sociosexual Orientation Inventory. The Revised Sociosexual Orientation Inventory (SOI-R; Penke & Asendorpf, 2008) comprises nine items that assess the willingness to engage in uncommitted sexual encounters. Items are divided in three components: behavior (three items; α = .85; e.g., "With how many different partners have you had sex within the past 12 months?"), attitudes (three items; α = .87; e.g., "Sex without love is okay"), and desire (three items; α = .86; e.g., "How often do you have fantasies about having sex with someone with whom you do not have a committed romantic relationship?"). Responses are given on 7-point scales (scale anchors depend on the item). Higher mean scores on each component signify more unrestricted behaviors, attitudes, and/or desires. A confirmatory factor analysis (CFA) showed a good fit of this measure in our sample: χ 2 (25) = 43.93, p = .011, comparative fit index (CFI) = .97, Tucker-Lewis index (TLI) = .95, standardized root mean square residual (SMSR) = .07, and root mean square error of approximation (RMSEA) = .06. Commitment Scale. The Commitment Scale was taken from the Investment Model Scale (Rusbult et al., 1998) and comprises seven items (α = .89; e.g., "I want our relationship to last for a very long time"). Responses are given on 7-point scales (1 = Do not agree at all, 7 = Agree completely). Higher scores represent greater commitment. A CFA showed a good fit of this measure in our sample: χ 2 (14) = 33.79, p = .002, CFI = .96, TLI = .94, SMSR = .05, and RMSEA = .08 (see also Rodrigues & Lopes, 2013).
Perceptions of Dating Infidelity Scale. The Perceptions of Dating Infidelity Scale (PDIS; Wilson et al., 2011) comprises 12 items that assess the extent to which individuals perceive behaviors as indicative of infidelity. Items are divided in three types of behaviors: ambiguous (six items; α = .90, e.g., "Talking on the phone or Internet"), explicit (four items, α = .93; e.g., "Sexual intercourse"), and deceptive (two items; α = .75, e.g., "Lying to one's partner"). Apart from the deceptive items, all other items are worded to indicate activities with an individual other than the current partner. Responses are given on 7-point scales (1 = Extremely low level of infidelity, 7 = Extremely high level of infidelity). Higher mean scores signify that behaviors are more indicative of infidelity. In the present sample, the CFA indicated a good fit of this measure: χ 2 (49) = 115.94, p < .001, CFI = .96, TLI = .95, SMSR = .05, and RMSEA = .07.

Predictors of Sexual Infidelity
As shown in Table 1, no significant differences were found between women and men in sexual infidelity: 66.7% of women and 68.7% of men reported they had engaged in infidelity, χ 2 (1) = 0.01, Cramer's V = .02.
To examine if sociosexuality and commitment were associated with increased odds of infidelity we conducted a logistic regression analysis in which sexual infidelity (coded: 0 = no, 1 = yes) was the dependent variable. In separate blocks we regressed gender and the variable "looking for" (the only one that showed differences between women and men; see Table 1), followed by each SOI-R component and commitment. As expected, results from Step 2 (Nagelkerke R 2 = .32) show that sexual infidelity was significantly associated with sociosexual behavior (p < .001), sociosexual desire (p = .046), and commitment (p < .001) ( Table 2). Gender was not associated with sexual infidelity either in Step 1 (p = .553) or in Step 2 (p = .382) 2 .

Sexual Infidelity, Sociosexuality, and Commitment
The descriptive statistics and zero-order correlations of all dependent variables are shown in Table 3. To examine if sexual infidelity was associated with differences in sociosexuality and commitment we conducted a 2 sexual infidelity (no, yes) multivariate analysis of covariance (MANCOVA). Multivariate
To further understand whether sociosexuality and commitment are associated with perceptions of infidelity, we conducted a set of hierarchical multiple regression analyses separated by sexual infidelity. In separate blocks we regressed the "looking for" variable (Step 1), followed by sociosexuality (behavior, attitudes, and desire) and commitment measures (Step 2). Each component of the PDIS separately defined the dependent variables. Results are presented in Table 4.
For individuals with sexual infidelity, unrestricted desire and lesser commitment were associated with the perception of ambiguous (p = .011 and p = .005, respectively) and deceptive behaviors (p = .008 and p = .005, respectively) as less indicative of infidelity. For these individuals, perceptions of explicit behaviors as less indicative of infidelity were associated only with unrestricted desire (p = .001). For individuals without sexual infidelity, only greater commitment was associated with perceptions of ambiguous (p = .019), explicit (p = .001), and deceptive behaviors (p = .001) as more indicative of infidelity. No other results reached significance.

Differences According to Motivations for Registration
Given the nature of our sample, we also examined differences in sociosexuality, commitment, and infidelity perceptions according to motivations underlying individuals' registration on the Web site. We first analyzed individual motivations. A multivariate 2 sexual infidelity (no, yes) × 2 looking for (know other people, casual sex) MANCOVA showed a nonsignificant interaction between the sociosexuality factors and commitment, Wilks's Λ = .98, F (4, 235) = 1.26, p = .285, and for perceptions of infidelity, Wilks's Λ = .99, F (3, 246) = 0.81, p = .488.

Discussion
In this study, we examined whether dispositional (sociosexuality) and relational variables (commitment) are associated with sexual infidelity. We additionally examined the role of these variables in perceptions of infidelity. We compared individuals with and without sexual infidelity. As the sample comprised only self-identified sexually monogamous individuals, extradyadic sexual behaviors are perceived as sexual infidelity. The study presented in this article is relevant for three main reasons. First, we used a sample of individuals who, albeit romantically involved, had also paid for registration on a dating Web site for other romantically involved individuals. Accordingly, all participants were behaviorally engaged in online infidelity. Second, unlike the bulk of research in which there is a focus on younger participants, our sample had a mean age of 40 years. This is an interesting and novel aspect for the literature. For instance, the nonexistence of gender differences typically observed in younger cohorts was here extended to this specific sample of individuals. Third, this was an anonymous Web survey. According to the literature, this acts against possible bias in reports of actual behavior, such as overestimation in men and underestimation in women (Blow & Hartnett, 2005a;Fisher, 2009;Whisman & Snyder, 2007). Hence, it is possible that these individuals (versus nonregistered individuals) have more positive attitudes toward sexuality and are less inhibited sexually, which are factors associated with casual sex in both men (Bancroft et al., 2004) and women (Carpenter, Janssen, Graham, Vorst, & Wicherts, 2008). Supporting this argument is the evidence showing that all individuals had highly favorable sociosexual attitudes. Even individuals who had not engaged in sexual infidelity were motivated enough to pay for their registration on the Second Love Web site and engage in (at least) online infidelity. This would probably be less likely if they had negative attitudes toward casual sex.
Results reported in this article show that sexual infidelity was associated with unrestricted sociosexual behaviors and desire, along with lesser commitment. Results also indicate that sexual infidelity is associated with differences regarding which behaviors are perceived to be indicative of infidelity. For individuals with sexual infidelity, unrestricted sociosexual desire and lesser commitment are associated with more permissive perceptions of infidelity. For individuals who have not engaged in sexual infidelity, greater commitment is associated with more restrictive perceptions of infidelity. Furthermore, results show that individuals who registered as a couple (versus individually) are not necessarily more motivated to engage in sexual infidelity due to unrestricted sociosexuality or lack of commitment. Likewise, individuals who are looking for casual sex (versus knowing other people) are Step 1 Step 2 Step 1 Step 2 Step 1 Step 2 Step 1 Step 2 Step 1 Step 2 Step 1 Step 2 .08 .08** .16** .06* .18** .08** Note. All variables were standardized prior to analyses. ΔR 2 = change in R 2 between Step 1 and Step 2. *p ≤ .05; **p ≤ .01; ***p ≤ .001.
RODRIGUES, LOPES, AND PEREIRA not necessarily more motivated to engage in sexual infidelity and less committed to their relationship. In both cases, there were also no differences in perceptions of infidelity. Moreover, in our sample, men were mainly looking for casual dyadic sex with others and women were mainly looking to know other people. This converges with research showing that men are more likely to engage in sexual infidelity, whereas for women infidelity is more often associated with an emotional connection (Martins et al., 2016). Research shows that both types of infidelity are perceived as equally harmful to the relationship (Lishner, Nguyen, Stocks, & Zillmer, 2008). The fact that most women indicated they were looking to meet other people does not necessarily imply they disregarded casual sex. Converging with this, we did not find significant gender differences in regard to incidence of sexual infidelity. This is in line with recent research indicating that gender differences in extradyadic involvement are becoming less pronounced (Brand et al., 2007;Martins et al., 2016;Shaw et al., 2013), including in extradyadic sexual encounters (McAnulty & Brineman, 2007).
Furthermore, and as expected, sexual infidelity was significantly associated with sociosexuality (behaviors and desire) and commitment. Past research shows that infidelity is predicted by relationship quality variables, such as lower commitment (Shaw et al., 2013) or lower satisfaction (Martins et al., 2016). Infidelity is also predicted by dispositional variables related to sexuality, such as lower sexual inhibition, greater sexual excitation (Mark et al., 2011), and greater number of past sex partners (Shaw et al., 2013). These variables directly relate to the definition of sociosexuality (Penke & Asendorpf, 2008;Simpson & Gangestad, 1991).
Likewise, individuals who have engaged in infidelity were more sociosexually unrestricted in their behaviors, attitudes, and desire, while also less committed. The fact that sociosexual behavior had a lower score than the remaining SOI-R components may be explained, in part, by the fact that it refers to sexual behaviors within the past 12 months and the number of sexual partners within this same time frame. Although they were registered on Second Love, our respondnets were also romantically involved. This may contribute to less diversified sexual behavior due to fewer opportunities, or at least less casual sex with different casual partners. This is important evidence suggesting that sociosexually unrestricted individuals do not necessarily engage in casual sex indiscriminately (Tsapelas et al., 2010). Moreover, if this was simply a matter of lack of opportunity, no differences in sociosexuality (and especially on the desire component of sociosexuality) and commitment should have emerged according to sexual infidelity. In fact, these differences emerged, suggesting that prior infidelity is more related to personal or relational factors rather than situational factors (Shaw et al., 2013).
Bridging these results with perceptions of infidelity, individuals with (versus without) sexual infidelity perceive ambiguous and explicit behaviors as less indicative of infidelity. Those without sexual infidelity may hold stricter views on which ambiguous behaviors are indicative of infidelity and may perceive certain actions as unacceptable in certain situations when in a sexually monogamous relationship (Feldman & Cauffman, 1999;Mattingly et al., 2010;Wilson et al., 2011). For instance, talking with another person over the Internet may not be considered infidelity if it takes place on Facebook. But the same situation may be indicative of infidelity if it takes place on a dating Web site. If an individual has engaged in sexual infidelity and considers talking with others on a dating Web site to be an acceptable behavior (regardless of what the partner thinks of it), there is a greater likelihood of repetition without considering it to be an unacceptable behavior.
Our results further show that, regardless of sexual infidelity, participants perceived explicit behaviors (e.g., oral sex, dating) as highly indicative of infidelity (Mattingly et al., 2010;Wilson et al., 2011). This was even more so among individuals without prior sexual infidelity. Those with prior sexual infidelity may develop a more malleable view of infidelity in an attempt to deal with their behavior, legitimize it, and face the potential consequences (Mattingly et al., 2010). The fact that no differences emerged for deceptive behaviors may be grounded in the specificity of our sample. Indeed, most of our participants were withholding information and lying to their partner by having registered on Second Love without the partner being aware of it. Further attesting this, participants perceived deceptive behaviors as moderately indicative of infidelity, possibly acknowledging their behavior as a form of online infidelity.
From the total sample, 13.5% indicated to have registered as a couple. This might seem incongruent with our definition of sexual infidelity. However, in this study, all participants indicated being in a sexually monogamous relationship, and having registered as a couple was not necessarily associated with sexual infidelity. From the 34 individuals that indicated couple registration, 16 indicated no prior accounts of sexual infidelity and 18 indicated accounts of infidelity. In accordance with our definition, these results suggest that, at least for individuals on Second Love, sexual infidelity refers to either individual sexual behaviors without the partner awareness or sexual behaviors without the presence of the partner. In this sense, individuals registered as a couple may not perceive extradyadic sexual behavior as infidelity as long as such behavior is engaged as a couple. When engaged in individually, this behavior may be perceived as infidelity.
There were also differences in the role of sociosexuality and commitment for perceptions of infidelity. More permissive perceptions of explicit infidelity among individuals with sexual infidelity were associated with their unrestricted sociosexual desire. More permissive perceptions of ambiguous and deceptive infidelity for these individuals were associated with their unrestricted sociosexual desire and also with their lower commitment. These latter results were not initially expected but complement our hypotheses. Unlike explicit behaviors, which are consensually perceived as infidelity, perceptions of ambiguous and deceptive behaviors may not be so clear. By being less committed to their relationships, individuals with prior sexual infidelity may SOCIOSEXUALITY, COMMITMENT, AND SEXUAL INFIDELITY view ambiguous (e.g., talking on phone/Internet; go someplace) and deceptive behaviors (e.g., lying; withholding information) as not indicative of infidelity, simply because they engage in them. Otherwise, they would probably not hold such perceptions.
Consistent with this argument are results for individuals without prior sexual infidelity. These individuals report being more committed to their relationships, which is associated with stricter perceptions of what constitutes infidelity. They perceive that explicit behaviors are highly indicative of infidelity, and these perceptions extend to deceptive and ambiguous behaviors. This extends previous research and further shows how dispositional and relational factors interplay with perceptions of what constitutes infidelity (Wilson et al., 2011).

Strengths, Limitations, and Future Directions
This research has three major strengths. First, we conducted an anonymous Web survey, thus reducing the likelihood of bias associated with social desirability (Martins et al., 2016;Whisman & Snyder, 2007). Second, we conducted this study on a dating Web site directed at romantically involved individuals. These were highly motivated individuals that voluntarily took part in a study made available on the Web site platform and were not deceived. This study has ecological validity and examined infidelity in an underrepresented sample in research. Third, our sample departed from the typical college student samples widely used in the literature and was diverse in terms of demographic and relational characteristics, including age, geographic location, relationship length, and relationship characteristics.
This research, however, is not without limitations. First, because of the cross-sectional design, causal associations cannot be inferred in regard to perceptions of infidelity. To overcome this limitation researchers could employ a longitudinal design, for instance examining how incidences of infidelity change perceptions of what constitutes infidelity and how these perceptions are predicted by fluctuations in sociosexuality and commitment. Second, our sample was restricted to heterosexual monogamous relationships and we did not directly assess motivations underlying registration on Second Love. This should be addressed to a greater extent in future research.
Also, future research should seek to extend this evidence to a broader sample of individuals, specifically studying other sexual orientations (e.g., lesbian, gays). Research shows that lesbian and gay individuals differ from heterosexuals in their attitudes and behaviors toward sexuality (Whitton, Weitbrecht, & Kuryluk, 2015). Future research should also consider extending this sample to include other types of romantic relationships (e.g., polyamorous relationships), given the scarcity of available evidence (McCoy, Stinson, Ross, & Hjelmstad, 2015).
Moreover, research should seek to disentangle whether registering to Web sites such as Second Love eventually leads to sexual infidelity. Just as we did not measure the underlying motivations of sexual infidelity, we did not measure for how long individuals without accounts of sexual infidelity were registered on the Web site, nor if they were motivated to pursue sexual encounters.
Third, given the specificity of this study, no information about the behavior or about the perceptions of infidelity from the partner was collected. Indeed, most participants indicated they registered without their partner knowing, and their perceptions of infidelity are possibly quite different from the perceptions held by their partner. Future studies should seek to examine how a complementary or distinct view of what is infidelity influences the likelihood of engaging in infidelity. Whereas similar views may lead to reduced infidelity or to behaviors that are not considered infidelity, distinct views are more likely to have negative consequences for relational well-being (Mattingly et al., 2010;Wilson et al., 2011).
Future venues of research should also examine how infidelity is associated with mate poaching, that is, an intentional pursuit of a person currently in a romantic relationship (McAnulty & Brineman, 2007). Research shows that individuals who were poached by their current partners are more attentive and less committed to their relationships, also engaging in infidelity behaviors (Foster et al., 2014). Hence, future research should analyze the incidence of single individuals on Web sites directed at romantically involved individuals, and study how single individuals make their approaches, the dynamics that are established between individuals, and whether mate poaching is more likely to occur or not given dispositional (e.g., sociosexuality) and relational (e.g., commitment) factors.

Implications and Conclusion
Our results converge with those of self-regulation research (Baumeister, Bratslavsky, Muraven, & Tice, 1998;Ciarocco, Echevarria, & Lewandowski, 2012). Individuals with greater self-control, similarly to those with greater commitment, can counteract and overcome their desires, feelings, needs, and typical patterns of behavior (i.e., unrestricted sociosexual orientation) and prevent infidelity. Indeed, research shows that individuals depleted (versus nondepleted) of self-control are likely to be more attentive to attractive others and to report more attraction (Ritter, Karremans, & van Schie, 2010), to accept a date with a confederate (Ciarocco et al., 2012), and to report greater intent to engage in sexual infidelity (Gailliot & Baumeister, 2007). This could be especially true for individuals for whom commitment is made more salient. For instance, asking individuals to think about their partner makes them more likely to activate pro-relationship mechanisms such as forgiveness (Etcheverry & Le, 2005;Finkel, Rusbult, Kumashiro, & Hannon, 2002). This can potentially constitute a way for clinicians to delineate intervention programs aimed at developing greater self-control and greater focus on commitment. Furthermore, this would help promote greater dedication and work on behalf of the resolution of negative conflicts that can arise after infidelity has occurred. Clinicians can also intervene so that definitions of infidelity are explicit for both members of the couple, thus preventing future conflicts and increasing relationship quality. RODRIGUES, LOPES, AND PEREIRA In conclusion, this is the first study examining sexual infidelity in a sample of romantically involved individuals who have registered on a dating Web site for romantically involved individuals. Based on our data, engaging in actual sexual infidelity in these specific situations may be a matter of commitment. Although it does not prevent individuals from pursuing some type of infidelity, commitment may help individuals to refrain from engaging in actual sexual infidelity and promote more strict views of what constitutes infidelity. These results have important implications for academics to better understand the infidelity phenomenon among individuals who are already engaged in online infidelity, as well as clinicians to delineate intervention programs to strengthen commitment and prevent future infidelity or to solve conflicts after infidelity has occurred.