Have you ever questioned your sexuality?

Relationship satisfaction and infidelity: one connection, two directions


How are unfaithful behavior in a relationship and satisfaction with the partnership related? Initially, the influence of the relationship quality on loyalty behavior appears plausible: the more dissatisfied a partner is in a relationship, the more likely he or she is to seek external relationships. However, this does not rule out that the quality of the partnership or its assessment can also change after the infidelity episode from the perspective of the perpetrator, e.g. B. due to conflicts and marital crises triggered by infidelity or to reduce feelings of cognitive dissonance. Using fixed-effects models based on the data from the relationship and family panel pairfam Over the observation period from 2008 to 2016, we are investigating both possible directions of action in a longitudinal section with a one-year and two-year interval between the measurement times. It turns out that there are interactions between the two factors, with the effects of infidelity on relationship satisfaction being greater in women than in men. While the relationship between relationship satisfaction and the risk of infidelity is rather low, it has been shown that the risk of cheating increases significantly when the long-term orientation in the relationship decreases.


How are unfaithful behavior and relationship satisfaction connected? It seems plausible that relationship quality has an influence on unfaithfulness: the less satisfied a partner is in a relationship, the more he or she will look for external relations. However, this does not exclude that relationship quality changes after the infidelity episode, for example, due to conflicts and marriage crises triggered by the unfaithfulness or to reduce feelings of cognitive dissonance. Using fixed-effects models based on data from the German Family Panel pairfam during an observation period from 2008 to 2016, we investigated both causal pathways with a longitudinal design and measurement points 1 or 2 years apart. We show that relationship satisfaction and infidelity are associated and that the influence of infidelity on relationship satisfaction is greater for women than for men. Although the association between relationship satisfaction and risk of infidelity is relatively low, it is shown that the risk of being unfaithful increases sharply if long-term commitment to the relationship decreases.


Even if science and society have discovered polyamorous arrangements as objects for themselvesFootnote 1, Exclusivity still applies in most intimate relationships. A more recent study from the USA by the American research group of Haupert et al. (2017) speaks of a prevalence of around 20% of all respondents who have ever had experience with a partnership in which exclusivity was not a basic requirement. This means that, conversely, 80% of the respondents lived exclusively in monogamous partnerships.

There are hardly any reliable figures for Germany in this regard, but over 90% of those surveyed would like to see a study by Schmidt et al. (2006, p. 133) sexual fidelity from her current partner. This is in line with the information provided by young people in current youth studies from Germany. For example, in the Shell Youth Study, loyalty was described as "trendy" by the majority of young people for years (Albert et al. 2015) and in the SINUS Youth Study 2016, young people describe loyalty as an important element of a partnership (Calmbach et al. 2016) .

These findings are in clear contrast to the prevalence of unfaithful behavior known to date, which is also likely to be underestimated due to the presumably high number of unreported cases. About 40% of all German-speaking women and men have cheated on at some point in their lives, with men tending to report more infidelity than women (for an overview see Kröger 2010). This is in line with the figures from the USA, as Hall and Fincham (2009) say that around 35% of Americans have cheated at some point in their lives. The prevalence among married couples seems to be even higher (Fincham and May 2017).

It is neither clear what the respondents mean by “cheating” or “unfaithful behavior”, nor has a uniform set of conceptual instruments emerged in science.Footnote 2 Most of the studies focus on extradyadic sexual activities (e.g. Atkins et al. 2001; Maddox Shaw et al. 2013; Mark et al. 2011; Previti and Amato 2004). It is critically noted with this type of conceptual handling that inter alia. Emotional relationships without sexual intercourse are not cheating, even though those affected perceive it as a betrayal in the partnership (see Fincham and May 2017 for an overview). This criticism led to a terminological separation of the concepts more sexual and emotional infidelity. While the former can be defined relatively clearly, definitions of the second concept vary widely (Gibson et al. 2016; Guitar et al. 2017; Thompson and O’Sullivan 2016a). For example, respondents feel cheated by their partner when feelings are involved (e.g. also when they only go to a restaurant with another person), but do not apply the same standards to their own behavior. Thompson and O’Sullivan (2016b) speak of one in relation to emotional infidelity Actor-observer biasbased on Nisbett et al. (1973) goes back: The behavior of others is assessed under stricter standards than one's own. A comprehensive definition of infidelity comes from Drigotas et al. (1999), according to which infidelity consists in any interaction with a third person who on the one hand violates existing partnership norms and on the other hand is associated with feelings of jealousy and rivalry, and thus does justice to both sexual and emotional infidelity.

Regardless of the definition, unfaithful behavior usually causes stress in a relationship, which often leads to the breakup of the relationship. It is not for nothing that cheating is one of the main causes of divorce (Allen and Atkins 2012; Amato and Previti 2003; Hall and Fincham 2009) and that is culturally evenly distributed (Betzig 1989). It is therefore not surprising that research into causes is being advanced in psychology and increasingly also in sociology, despite great difficulties in obtaining data and a high number of unreported cases (see literature review by Fincham and May 2017).

A characteristic that is often cited as a cause of infidelity is relationship satisfaction (Maddox Shaw et al. 2013; Previti and Amato 2004). The more dissatisfied a partner is in a relationship, the more likely he or she is to seek external contacts and relationships. However, this consideration contradicts the statements made by unfaithful people that the infidelities did not result from dissatisfaction with and doubts about the relationship, but rather the “allure of the new” and sexual attraction were decisive (Schmidt et al. 2006, p. 135 ). So the question arises as to whether infidelity is actually a result of poor relationship satisfaction.

In addition, the causal connection between relationship satisfaction and infidelity cannot be checked using cross-sectional data (with simultaneous measurement of infidelity and relationship satisfaction), since it cannot be ruled out that the quality of a relationship may decline after an episode of infidelity. Obviously this is for the relationship satisfaction of the cheated person, who - if he or she is aware of the breach of trust - feels deceived and is forced to revise his or her opinion about his or her partner. The fraudster's relationship satisfaction can also decrease, for example due to conflicts and crises in the relationship triggered by the episode (which has become known), the contrast between the current and the potential new partner, or due to feelings of cognitive dissonance. Accordingly, empirical research on the mutual influence of both factors is inadequate, especially since so far - especially for Germany - only cross-sectional studies are available which, due to the nature of the data structure alone, cannot clarify whether or to what extent a poor relationship quality is the cause or consequence of infidelity.

In this article, based on the data from the pairfam relationship and family panel, it is analyzed longitudinally over the observation period from 2008 to 2016 whether relationship satisfaction contributes to explaining the risk of infidelity and / or whether unfaithful behavior reduces the quality of relationships.

We are following the recommendations of Fincham and May (2017) as well as Munsch (2012) and deliver one of the first longitudinal studies on the subject of "infidelity" and, to the best of our knowledge, the first scientific longitudinal study German Data. In addition, we counter the criticism of small-numbered samples by using pairfam, the data from a large nationwide random sample from the period 2008 to 2016, which also includes both non-cohabiting and cohabiting couples. In this way, more recent social developments are taken into account, as many of the existing studies are based on data from the 1980s and often disregard non-cohabiting couples.

The interviewee's relationship satisfaction at time t is linked to the behavior of this person in the 1 to 2 years after this point in time (reported in the following wave), which ensures that the (previous) low relationship satisfaction cannot be the result of the infidelity episode. With regard to the opposite direction, the longitudinal data is used to check whether the relationship satisfaction changes from time t to t + 1 after the occurrence of infidelity between the two measurement times. The estimators of the fixed effects models (Sect. 6) show interactions between the two factors. Relationship satisfaction has a negative effect on the risk of infidelity regardless of gender and is determined by the so-called long-term orientation (Commitment) mediated. Infidelity, in turn, significantly reduces relationship satisfaction, and more so in women than in men.

Why does relationship satisfaction affect loyalty?

So far, there are hardly any sociological approaches to explicitly explain unfaithful behavior. However, since cheating represents an individual decision to act, theories with an action-theoretical core on the micro level offer an explanatory approach. If one also considers that cheating is one of the main causes of separations and divorces (Allen and Atkins 2012; Amato and Previti 2003; Hall and Fincham 2009), it makes sense to use theoretical ideas that have so far been used to explain relationship stability. Even if unfaithful behavior does not necessarily lead to a separation or divorce, it often does and at least from the perspective of the unfaithful person signals the willingness to accept a breakup of the relationship.

In order to investigate separation behavior, exchange theoretical models have proven to be fruitful, which are based (in a broad sense) on the assumptions of a rationally calculating actor (for an overview of rational choice approaches see Braun and Gautschi 2011): Optimizing actors make use-maximizing decisions Basis of a subjective cost-benefit calculation under given restrictions and resources. They evaluate situations and alternative courses of action according to anticipated rewards or expected punishmentsFootnote 3 (Lindenberg 1981) and are interdependent with other people (Nye 1982). Since people only have limited resources, the social exchange of goods and services enables individuals to improve their own equipment. Social exchange is thus guided by utility-theoretical motives for the satisfaction of needs, with the - not necessarily intended - (positive) by-product of initiating and improving social interactions (Hill and Kopp 2015).Footnote 4

From a theoretical exchange perspective, partnerships can be understood as a market in which the respective partners exchange goods, such as affection, social recognition, sex, care, love, childcare or housework. Exchange theory generally distinguishes between two categories of goods: economic and social or reciprocal goods, with the latter being particularly relevant for the analysis of couple relationships. In the exchange of social goods - unlike in economic exchange - goods are not (simultaneously) exchanged for goods or money, but (affective) elements are transferred at different times. Economic and social goods also differ with regard to the occurrence of saturation effects (see Blau 1964, p. 90 f. For a description of the principle of decreasing marginal utility in social exchangeFootnote 5): Social goods such as love, care or recognition cannot be saved for the future, which is why the need for these does not diminish over time (Arránz Becker 2008). Since at the time of your own investment, i. H. of one's own action, the reciprocal consideration of the partner can be expected but is by no means certain, the lack of saturation effect with social goods increases the probability of a reciprocal response. If there is no consideration, this has an impact on relationship satisfaction and the likelihood of separation.

Thibaut and Kelley (1959) process this idea in the Interdependence Theory. Accordingly, the resolution of a relationship (or a one-sided exit option such as infidelity) depends on the extent of the discrepancy between the Outcome a relationship, i.e. the status quo of the relationship or the success of the relationship, and the expectations of the partners, the so-called level of comparison (Comparison level [CL]). The status quo of the relationship (Outcome) is composed of the weighted sum of subjectively assessed attributes of the partner, for example intelligence, attractiveness, humor or even sexual compatibility (Rusbult 1980; Thibaut and Kelley 1959). These properties can easily be reinterpreted and expanded as social exchange goods, because the exchange of positively assessed interactions within a couple relationship is only possible if said attributes, or at least some of them, are also positively evaluated. For example, social recognition can be given as a compliment on the attractiveness of the partner.

Individual demands that result from past partnership experiences, observations of other couples or from social norms form the comparative level (CL) against which the status quo of the relationship must be measured. Thibaut and Kelley (1959) define the difference between the status quo of the relationship (“outcome”) and the level of comparison as relationship satisfaction, which in turn influences the stability of couple relationships. "Relationships the outcomes of which fall above CL would be relatively 'satisfying' and attractive to the member; those entailing outcomes that fall below CL would be relatively 'unsatisfying' and unattractive ”(Thibaut and Kelley 1959, p. 21). The dissolution of a partnership and, if necessary, one-sided exit strategies in the form of unfaithfulness occurs when the reward from reciprocal actions fails to materialize or is assessed negatively compared to previous or other relationships.

The investment model according to Rusbult (1980, 1983) seems particularly suitable for further transferring the basic idea of ​​exchange theoretical considerations to unfaithful behavior, since in its original version it uses the comparison level and the alternative comparison level from the Interdependence Theory von Thibaut and Kelley (1959) as well as the concept of relationship-specific investments, as they also appear in Becker's (1960) family economic model. Another advantage of the investment model is the fact that the concept has already been applied to infidelity for the first time by Drigotas et al. (1999).

In the investment model according to Rusbult (1980, 1983) the so-called Commitment both partners play the central role, i.e. the long-term orientation of the partners in the couple relationship (Häring 2017) or the bond with the partner and the will to linger in a relationship. This long-term orientationFootnote 6 directly influences the risk of separation and at the same time acts as a mediator for three further factors: relationship satisfaction, quality of the alternatives and the (financial and emotional) investments in the relationship. It has already been defined by definition that relationship satisfaction and long-term orientation do not describe the same concept. "Satisfaction [...] refer [s] to the degree of positive affect associated with a relationship. The individual’s commitment to an association, however, is related to the probability that he / she will leave the relationship "(Rusbult 1980, p. 174).

Rusbult (1980, 1983) understands relationship satisfaction to be the discrepancy between the comparative level (CL) and the status quo of the current relationship, following the model explained above by Thibaut and Kelley (1959). Relationship satisfaction has a positive effect on long-term orientation.The more satisfied a person is in their relationship, that is, the more the difference between the status quo of the relationship and the level of comparison deviates upwards, the more likely a person will think long-term and therefore want to maintain the partnership. Accordingly, it may well be that an individual is trapped in an unsatisfying relationship: “High investments and / or poor alternatives may sometimes serve to 'trap' the individual in an unhappy, unsatisfying relationship - commitment may be high while satisfaction [… is ] low "(Rusbult 1980, p. 175).

In addition to relationship satisfaction, investments play an important role as an influencing factor on the likelihood of separation, mediated through long-term orientation (Commitment). They are characterized by a sharp drop in value in the event of a separation or divorce and act as the cement of a relationship. Investments encompass everything that a person would lose in the end of the relationship, such as mutual friends or shared property (Drigotas et al. 1999, p. 178). High investments increase the long-term orientation of both partners and thus have a positive effect on the likelihood of staying in a relationship.

Attractive alternatives, as a third influencing factor, reduce the long-term orientation or bond with the partner, since after the relationship has been dissolved, a new relationship, which is perceived as better, could be entered into, which would replace the current partnership. Actors compare their ideas of an alternative relationship with their own level of comparison in order to estimate an anticipated relationship satisfaction with the alternative. Subsequently, the relationship satisfaction in the current partnership and the anticipated satisfaction with the alternative are compared to decide whether one should switch to the alternative partnership.

If you no longer see a future for yourself and your partner (low long-term orientation), the step out of the relationship, whether in the form of a separation or in the form of infidelity as a one-sided exit strategy, is only a matter of time. The less investments are made in the relationship, the better the alternative performs compared to the current partnership and the more dissatisfied one of the partners is in a relationship, the lower your own long-term orientation (or that Commitment) and this increases the likelihood of a breakup or unfaithful behavior.

However, the long-term orientation is not only influenced by the three factors, and thus influences the likelihood of unfaithful behavior, but also has a reciprocal effect on the three factors. A high long-term orientation stabilizes a relationship in three ways (cf. Drigotas et al. 1999): Firstly, it has the effect that possible alternative partners are belittled in the perception of the actor, i.e. are judged worse than they would be objectively. Second, a high long-term orientation leads to lower present preferences: long-term oriented partners consider the consequences of their current actions that lie far in the future (such as infidelity) and rate their costs (separation, jealousy, feelings of guilt) and benefits no less than advantages that are closer in time (e. B. Gaining pleasure and self-affirmation through an affair). For these reasons, long-term partners are less likely to succumb to the temptation of fling. In addition, the cost-benefit considerations of long-term oriented actors include the future well-being of the partner in their considerations, which thus finds its way into the relationship satisfaction of the actors (e.g. "I am satisfied when my partner is doing well") and accordingly the decision barometer in Towards loyal behavior. Thirdly, investments in a relationship are only profitable with a high level of long-term orientation, since their nature only yields returns in the future.

The great advantage of the investment model according to Rusbult (1980, 1983) is the interaction of the four elements and the resulting possible explanations for unfaithful behavior. This makes it clear why actually satisfied people cheat (e.g. because too little investment has been made in the relationship or because of an alternative that is classified as particularly attractive) and why dissatisfied people remain loyal (e.g. because too many are already there investments made are at stake or no suitable alternatives are available) (cf. Drigotas et al. 1999).

In summary, the following hypothesis can be drawn up:


The lower a person's relationship satisfaction, the greater the likelihood that that person will cheat in the following time.

This relationship is conveyed through the long-term orientation of the person, from which hypothesis H1b results:


The lower a person's relationship satisfaction, the lower their long-term orientation (Commitment) and therefore the greater the likelihood of cheating in the following time.

Why does infidelity affect relationship satisfaction?

In order to clarify the question of the causal effect of infidelity on relationship satisfaction, two perspectives must first be distinguished: perpetrator vs. victim. If a person has been cheated on, it is obvious that the relationship satisfaction decreases due to the breach of loyalty by the partner (which has become known) or the relationship is ended entirely. This case only plays an indirect role in our analyzes, because in the empirical analysis we focus on the effects of infidelity on the perpetrator.

For the explanation of the extent to which infidelity can lead to decreased relationship satisfaction in the unfaithful person, it depends first of all on whether the partner is aware of the breach of loyalty or not. If the cheating confessed after the offense or if the cheated person has otherwise learned of the partner's infidelity (or at least has a strong suspicion) and the couple nevertheless decided not to separate (yet), this is due to the perpetrator's behavior The relationship is burdened by negative interactions and feelings such as guilt, reproach and increased jealousy, which can lead to decreased satisfaction with the relationship (Abrahamson et al. 2012; Charny and Parnass 1995; Gordon et al. 2007; Olson et al. 2002 ).

The episode can trigger a relationship crisis in the person who has been betrayed in the sense of a fundamental reflection on the relationship (Esser 2002). Reframing radically changes the previous view of the relationship ("frame", to use the frame selection theory) and the previous lack of unquestioning relationship disappears. As a result, further investments in the relationship, which is now being put to the test as a whole, are suspended: Why invest time on vacation together when the relationship is on the verge of failure? This in turn leads to further impairment of the relationship.

The pair dynamics after infidelity were described in detail in qualitative studies. Olson et al. (2002), the process of coming to terms with it takes place in three phases, beginning with an emotional “roller coaster ride”, which is often associated with severe, sometimes physical, conflicts, followed by a moratorium phase characterized by the withdrawal of the partners, and finally the rebuilding of the relationship in which the partners get closer to each other again, trust is restored and new structures are created. At the end of the final phase, the relationship may even be perceived as better than it was before the incident. The reports of the respondents to a qualitative study by Abrahamson et al. (2012) clearly.

But regardless of whether the cheated partner has knowledge of the case of fraud, the infidelity episode can have consequences for the unfaithful person's satisfaction with the relationship.

On the one hand, infidelity changes the behavior of the unfaithful person, for example, as in Abrahamson et al. (2012) described by being less attentive to her partner (and mentally with the lover), but also because feelings of guilt represent an emotional burden that is reflected in the behavior of the unfaithful person. This in turn can lead to decreased relationship satisfaction, jealousy and suspicion of the (ignorant) cheated person as well as to conflicts which ultimately lead to a lower relationship satisfaction of the unfaithful person.

On the other hand, it is conceivable that the unfaithful person only becomes aware of the inadequacies of the existing partnership through their affair or infidelity. The external relationship is thus perceived as an attractive alternative, but cannot be implemented as a new partnership, for example due to social norms, high exit barriers (e.g. common children or common property in the existing partnership) or because the "other" is not interested in a serious relationship. By experiencing the attractive alternative in contrast to the current partnership, the long-term orientation can now decrease, which in turn has a negative effect on relationship satisfaction (see Fig. 1). Here, too, the above-described concept of the crisis according to Esser (2002) can be used to explain: The unfaithful person comes to the point of reframing through the external relationship, i.e. H. triggered by the experience of the alternative, the existing relationship loses its unquestionability and is reassessed.

Finally, lower relationship satisfaction can also appear as a strategy for resolving states of cognitive dissonance. Assuming that over 85% of the respondents in pairfam classify infidelity as a “serious relationship problem”Footnote 7, that loyal behavior in a partnership is preferred by the vast majority of the population and can therefore be viewed as normatively set (see also Burkart 2018, Chapter 11), an act of cheating could trigger states of cognitive dissonance. Cognitive dissonance arises when various cognitive elements (opinions, norms, attitudes, behavior) cannot be brought into harmony with one another. B. Attitudes do not fit behavior or opinions do not fit certain thoughts (Festinger 1962). One speaks of a dissonant relationship between two cognitions if the agent perceives the emotional state resulting from the incompatibility of the elements as unpleasant and / or frustrating (Opp 1964).Footnote 8 An action from one of the two cognitive elements can then take place, which significantly increases the cognitive dissonance again. In the case of infidelity, the desire to cheat (be it for reasons of lust, affirmation, being in love, etc.) would be contrary to the norm of perceiving cheating as immoral (because one offends one's partner, abuses trust, etc.), and would thereby trigger cognitive dissonance. If the actual cheating takes place in the form of an affair or an affair, the cognitive dissonance is further intensified.

Affected by states of cognitive dissonance, people strive to reduce them and thereby reduce the state of tension (Festinger 1962). This can be achieved through different strategies. Actors can change their wishes and actions or their convictions, they can suppress the relationship between the cognitions or they can try to justify their own action. Applied to infidelity, this means that the perpetrator can try not to “want” to cheat or end the affair (of course, an affair cannot be reversed). Alternatively, he or she could change his or her own attitude towards unfaithful behavior by denying that cheating is harmful to the partner, which at times can even lead to the claim that the affair "is good for the relationship". Finally, the unfaithful behavior can be reinterpreted as an inevitable consequence of undesirable characteristics of the relationship with the partner, which should result in a changed perception of one's own relationship satisfaction. In other words: The perpetrator justifies his or her own behavior, which is perceived as immoral, in retrospect by interpreting the relationship satisfaction as low.

Therefore, in summary, the following hypothesis can be formulated:


If a person cheats, their relationship satisfaction decreases as a result.

With the data available from the relationship and family panel, the mechanisms mentioned here cannot be identified separately from one another. So is unknownFootnote 9whether the cheated partner has knowledge of the breach of trust, which also makes it impossible to find out between cognitive dissonance, a changed assessment of the current relationship and the reaction of the cheated person or changed couple interaction as the cause of decreasing relationship satisfaction on the part of the perpetrator to differentiate an affair.

Do the relationships apply equally to women and men?

Finally, with regard to the connection between relationship satisfaction and infidelity, the question of differences between men and women should be asked. These are conceivable for several reasons: On the one hand, stricter norms generally apply to women with regard to sex outside of marriage (Pines and Friedman 1998), although a certain degree of gender alignment has already taken place (Sharpe et al. 2013). This can also be seen in Germany: Götsch (2015), for example, in a qualitative study among vocational school students, as well as earlier Bahne and Oswald (2005, p. 206) in a study with young people in Berlin, demonstrate a clear double standard with regard to permissive sexual behavior while In a representative survey of students, there were hardly any gender-specific differences in the assessment and practice of loyalty / infidelity (Böhm et al. 2016).

In addition, the existing literature suggests that female infidelity more often contains emotional components, while it is more often only sexual infidelity in men (Glass and Wright 1985; Træen et al. 2007).Footnote 10 For example, in one qualitative study, some men regard their affairs as more of a sport than affairs (Green et al. 2016), while in another qualitative study, women named primarily conflicts in the relationship and lack of attention from their partner as reasons for their infidelity (Jeanfreau et al. 2014).Footnote 11 It should be noted, however, that a more recent German study does not (no longer) find these gender-specific differences (Plack et al. 2010).

Another difference between men and women may be the detection rate of outside contacts. From the pairfam data it can be indirectly deduced that around two thirds of the women cheated and almost half of the men cheated know about the behavior of their partners: In the questionnaire, there is the question of whether someone has been unfaithful in the relationship with the answer options “yes, me”, “yes, my partner”, “yes, both me and my partner” and “no”. One can assume that the respondents answer the question to the best of their knowledge, since it is asked in the self-filling part of the interview. If one assumes that cheating occurs equally often among respondents and their partners (or more precisely: among male respondents and male partners of female respondents as well as female respondents and partners of male respondents), we can estimate from the aggregated information How high the proportion of known episodes of infidelity is: 63% of women know about a breach of trust, 44% of men.Footnote 12

These differences can have implications for the impact of relationship satisfaction on infidelity. According to Levinger (1976), the explanations given in the investment model for the (one-sided) exit from a relationship are insufficient. He recommends, in addition to rewards (attractions) also consider exogenous, such as social or cultural, influences or barriers in a model. Accordingly, stricter norms in the sense of a stricter loyalty requirement for women should mean that men are generally more likely to cheat than women, regardless of their relationship satisfaction. Corresponding gender differences in the prevalence of infidelity are known from the literature (Blow and Hartnett 2005b; Kröger 2010). Whether the applicable infidelity norms, and in particular the stricter norms for women, which take relationship satisfaction into account (and are particularly strict for unfaithful people in well-functioning relationships) is still unclear, but it is entirely plausible. In this case, the relationship between relationship satisfaction and infidelity could be more pronounced in women than in men, for example because men are generally allowed a certain amount of infidelity, while this applies to women only in relationships that are on the verge of failure or, more generally, in which they are very dissatisfied.

The second aspect of the gender-specific differences in motives and types of infidelity would also suggest a stronger connection between relationship satisfaction and infidelity for women, since, according to this argumentation, men use infidelity to satisfy sexual desires regardless of their relationship satisfaction, while women are emotionally involved and via fling prepare the exit from the unsatisfactory relationship.

The higher detection rate of infidelity among men, on the other hand, would have to influence their calculations in the other direction, if the perpetrators take the risk of being discovered into cheating in their decision (which, however, cannot necessarily be assumed). Particularly in the case of high relationship satisfaction, an increased risk of detection in men is likely to reduce the likelihood of infidelity, while it is accepted in the case of low relationship satisfaction.

It cannot be predicted which of the mechanisms will predominate, which is why we will forego a directed hypothesis.

We can also consider gender-specific differences with regard to the opposite direction, i.e. the effect of infidelity on relationship satisfaction. Due to the higher probability of being cleared up, in the case of infidelity by men, the reactions of the partners could tend to be the decisive factor for decreasing relationship satisfaction or, as a result, the change in relationship satisfaction for women after cheating could be lower because the partners learn less about the episode and thus the effects on the quality of the relationship is lower.

However, if the stricter infidelity norms for women are also taken into account in this direction of action, the consequence would be a stronger negative reaction of the cheated man and more relationship conflicts as well as a greater cognitive dissonance and, accordingly, a greater reduction in relationship satisfaction among women.

Which of these opposing effects outweighs or whether they cancel each other out can also be done here a priori cannot be clarified. In addition, it is unclear whether women and men react equally often to infidelity by separating. It could be that women due to higher exit barriers - for example because they fear material losses (see e.g. Leopold 2018) or because they feel obliged to keep an "intact" family for their children - more often after an episode of infidelity in one Relationship remains, while men with equally high losses in relationship satisfaction are more likely to opt for a separation. Selective breakups, however, can skew the relationship between infidelity and relationship satisfaction. A specific hypothesis on gender-specific effects is therefore also dispensed with here.

Literature review

Although empirical research on the relationship between relationship satisfaction and cheating has been around for decades (Blow and Hartnett 2005a, b), the current state of research leaves some questions unanswered. So far, with the exception of the study by Previti and Amato (2004) mentioned above, the effect of infidelity on relationship satisfaction or general relationship quality has mainly been investigated in qualitative studies or in clinical samples with a view to couples therapy (e.g. Olson et al. 2002; Atkins et al. 2005).

More empirical literature can be found on determinants of infidelity, such as relationship satisfaction, which, however, as we shall show below, has provided only poorly reliable results. In particular, the data situation is unfavorable, so that the majority of the studies are based on cross-sectional data. However, this cannot test whether a low level of relationship satisfaction leads to infidelity or vice versa.

For example, Treas and Giesen (2000) use cross-sectional analyzes to calculate determinants of infidelity for a group of American university students on the basis of data from the US National Health and Social Life Survey 1992 and take relationship satisfaction into account as a factor. Most studies report a connection between relationship satisfaction and infidelity, although - to emphasize once again - there can be no talk of a causal effect here.

The first longitudinal study was carried out by Drigotas et al. (1999) who used a survey of students at the beginning and at the end of a semester to investigate the relationship between characteristics such as relationship satisfaction at the first time of the survey and their infidelity during the semester. Even if both the theoretical model and the empirical approach are quite convincing, the reliability of the results is low due to the insufficient number of cases of only 74 students (14 of them men), the special sample and the short observation period of only two months.

A study with a larger longitudinal data set comes from Previti and Amato (2004), who consider three points in time in their model in order to differentiate the initial relationship quality, infidelity and separation over time. It is one of the few studies that investigates both directions of action. In this model, satisfaction with the marriage (only married couples are considered) no longer plays a role as soon as the relationship stability (i.e. the question of whether the respondent is already thinking about a separation or whether the couple has already talked about a separation) is checked . This can be interpreted in such a way that the relationship satisfaction only indirectly, mediated via the relationship stability, influences the risk of infidelity. Previti and Amato also find an effect of cheating on marriage satisfaction and thoughts of separation. However, the study has some weaknesses: the number of couples who report a case of infidelity is relatively low at 69. In addition, the question of infidelity (only about “extramarital sex” without considering other forms of infidelity) is asked in the face-to-face interview, which results in a distortion of the information in the sense of social desirability and thus an underestimation of infidelity should.

DeMaris (2009) uses the same data as Previti and Amato (2004) for an event data analysis over a period of 20 years and, similar to Previti and Amato, finds a statistically significant effect of relationship stability, but not relationship quality (measured in the previous wave) on the risk of breach of trust. However, since a number of indicators of relationship quality, such as violent conflict and frequency of time spent together, are included in the model, an effect of relationship quality could possibly be masked by multicollinearity.

Maddox Shaw et al. (2013) used data from unmarried couples over an observation period of 20 months to determine the determinants of infidelity. The analysis confirms the hypothesis that the higher the initial relationship satisfaction, the lower the risk of later infidelity. However, only bivariate models were calculated, which seems doubtful despite the within approach, which makes the control of time-constant features unnecessary.

Studies for Germany are rare. A study of risk factors for infidelity was carried out by Plack et al. (2010) and is based on non-representative cross-sectional data. The analysis of the effects of infidelity on partner satisfaction by Plack et al. (2008) is also based on cross-sectional data from a non-representative sample. Therefore, the results cannot be interpreted causally in the sense that a poor relationship quality causes an increased risk of infidelity.

In this way, we counter central points of criticism with our analyzes. We use a broad population sample from Germany, do not limit ourselves to married couples and analyze panel data, so we do justice to the temporal structure of the events.

Data and method

For the analysis, we use the data from the relationship and family panel pairfam Release 9.1 (Brüderl et al. 2018a), a longitudinal study with Germany-wide random samples of originally 13,891Footnote 13 Respondents from the three birth cohorts 1971 to 1973, 1981 to 1983 and 1991 to 1993. The respondents have been asked annually since 2008 on topics such as partner search, partnership quality, fertility and parenthood as well as intergenerational relationships.Footnote 14