Prayer for a Partner and Relationship Outcomes: The Moderating Role of Relationship Form (Marital vs. Nonmarital Relationship)

: The current study was designed to investigate the concurrent and temporal associations between petitionary-focused prayer for a partner and relationship outcomes such as relationship satisfaction, commitment and quality, romantic loneliness, and relationship disillusionment from the perspective of the moderating role of relationship form (marital vs. nonmarital relationship). This study utilized a cross-sectional and longitudinal design and examined 412 Polish adults aged 18–75 ( M = 35.10, SD = 9.50) at Time 1 and 190 participants after a four-month interval. The results revealed a moderation function of relationship form at T1 for relationship satisfaction at T1 and that more praying for a partner at T1 was associated with higher relationship satisfaction at T1 in marital relationships but not in nonmarital relationships. There was, however, a trend suggesting that more praying for a partner at T1 might be related to lower levels of romantic loneliness at T2 in marital relationships but not in nonmarital relationships. No interaction effects were observed for other relationship outcomes. These results suggest that the effects of praying for a partner may differ depending on relationship outcomes and that prayer may play a positive function with respect to relationship satisfaction in marital relationships in comparison to nonmarital relationships.


Introduction
Prayer represents a spiritual form of activity prevalent in all Abrahamic traditions (i.e., Judaism, Christianity, and Islam) and possesses parallels in other religious traditions, such as Buddhism or Hinduism (Fincham et al. 2010). Therefore, prayer may be considered an essential activity in most religions (Lambert et al. 2012). Petitionary prayer "invokes God's help in response to specific needs" (Fincham et al. 2010, p. 650) and includes requesting "God's help or intervention" (May et al. 2020(May et al. , p. 2992. Colloquial petitionary prayer rests upon using an individual's own language rather than a set of words or memorized prayer (Fincham et al. 2010;May et al. 2020). Petitionary prayer can take various forms, for instance, it can include (1) self-focused prayer (SFP) when an individual makes requests about one's protection or well-being (Fincham and Beach 2014;May et al. 2020), and (2) partner-focused petitionary prayer (PFPP) "in which there is an explicit focus on the needs of one's partner." (Fincham and Beach 2014, p. 587). In other words, partner-focused petitionary prayer involves praying for the person with whom an individual has the most important relationship (Lambert et al. 2012) and involves petitioning the Deity on behalf of one's partner's well-being, positive things for one's partner, protection over the partner, and helping the partner to achieve his/her goals (Fincham and Beach 2014).
In prior research, partner-focused prayer has been found to be concurrently related to relationship satisfaction  and prospectively (in a 7-week interval) with commitment, and this link was partially mediated through enhanced relationship satisfaction (Fincham and Beach 2014). Furthermore, prayer has also been shown to increase forgiveness (Lambert et al. 2010b) and reduce extradyadic romantic behaviors (infidelity) (Fincham et al. 2010), and a higher frequency of prayer was associated with lower alcohol consumption and problematic drinking behavior both cross-sectionally and longitudinally (Lambert et al. 2010a). Praying for a close relationship partner has also been demonstrated to be associated with more cooperative and forgiving behaviors toward one's partner (Lambert et al. 2013). The positive outcomes of praying in a relationship have been demonstrated not only for individual prayer for a partner but also for praying with and for a partner; specifically, praying for and with a partner has been found to be associated with greater trust in a partner (Lambert et al. 2012). Finally, in a recent study, May and colleagues (2020) investigated the role of daily partner-focused petitionary prayer in relationship communication and quality and cardiovascular functioning among 90 adults. The results of this study revealed that in a 4-week period, individuals who were engaged in partner-focused petitionary prayer improved their relationship quality and cardiovascular efficiency by improving their protective cardiovascular mechanisms (May et al. 2020).
The review of the literature pertaining to praying for a partner and relationship outcomes reveals two key notions. First, the majority of the prior research on religion and praying in the context of relationships has largely focused on marital unions (Aragoni et al. 2021;Henderson et al. 2018;Langlais and Schwanz 2017). This focus is understandable, taking into account that religious involvement enhances the chances of marrying and religious attendance, and the salience of religion in daily life is considered a key factor in making decisions regarding starting cohabitation before or within marriage (Mahoney 2010). At the same time, over a few years, unmarried couples have become a group of particular interest among researchers who are investigating religiosity and relationship outcomes (Aragoni et al. 2021). For instance, Langlais and Schwanz (2017) showed that relationship religiosity mediated the association between individual and partner religiosity and relationship satisfaction in dating relationships. Furthermore, various components of religious involvement have also been found to be positively linked with relationship satisfaction and expectations of marrying among nonmarried individuals (Henderson et al. 2018). Finally, in a very recent study, Aragoni and colleagues (2021) found that in nonmarital relationships, perceived religious differences and dyadic attendance were predictors of commitment, whereas perceived religious differences were related to relationship satisfaction and negative interactions. In this context, the gap in research on religiosity and unmarried relationships appears to require addressing the contemporary changes in the sociohistorical context of today's people's intimate relationships, which are marked by their greater diversity and complexity and various structural arrangements, such as cohabitation or living apart together (LAT) (e.g., Bühler and Nikitin 2020;Henderson et al. 2018). Moreover, these transformations are accompanied by the increased popularity of cohabitation and the relevance of nonmarital relationships for people's well-being (e.g., Henderson et al. 2018), and the diminished role of marriage, with "being married [to be] less likely to be considered a key part of adulthood" (Bühler and Nikitin 2020, p. 463). Despite this growing interest in unmarried unions, "surprisingly little is known about the role (if any) of religious factors in shaping the quality, stability, and trajectories of nonmarital unions" (Henderson et al. 2018(Henderson et al. , p. 1922, while religiosity has been suggested to play "some role in unmarried relationships, at least for a good number of people, and some of the findings from married couples can be extended to those who are not married" (Aragoni et al. 2021, p. 3).
Second, past research on prayer for a partner has mainly focused on outcomes such as relationship satisfaction and commitment, forgiveness, trust, infidelity, and marital conflict, while it has paid less attention to outcomes such as loneliness and marital/relationship disillusionment. The omission of these two outcomes appears to be quite surprising because loneliness is a common psychological problem that affects many individuals, and satisfying, high-quality relationships are one of the strongest factors that protect against loneliness (e.g., Mund and Johnson 2021). At the same time, studies devoted to religiosity and loneliness appear to be scarce, with at least two exceptions. Specifically, Ismail and Desmukh (2012) observed in a sample of 150 Pakistani Muslims a strong negative association between loneliness and religiosity measured in terms of religious gathering attendance, belief salience, and frequency of prayer. In turn, Pereira et al. (2014) noted the lack of a link between prayer for a partner and loneliness. With respect to marital/relationship disillusionment, this factor has been recognized in the literature as one of the factors related to stability vs. termination of marital and other relationships (Niehuis et al. 2021) that is defined as "an extreme version of disappointment [that] captures individuals' perception that their relationship has changed for the worse" (Niehuis et al. 2019, p. 210). In past research, greater disillusionment in cohabiting and married couples has been found to be related to a higher self-rated likelihood of relationship termination, even when relationship satisfaction, commitment, and length were controlled (Niehuis et al. 2015), and actor and partner effects of disillusionment temporally preceded individuals' depressive symptoms (see Niehuis et al. 2021). Because prior studies have well documented the link between marital/relationship disillusionment and several relationship outcomes, the inclusion of this construct that has not been-to the best of our knowledge-examined in research on praying and relationship outcomes appears to be an important task in this area.
In the context of the empirical evidence and theoretical framework elaborated above, the objective of the current study is to determine whether petitionary-focused prayer is concurrently and prospectively (within a 4-month interval) associated with diverse relationship outcomes and whether the form of a relationship (marital vs. nonmarital relationship) moderates the associations between prayer for a partner and relationship outcomes in a sample of Polish adults. For the current investigation, we examined conceptualizations of petitionary-focused prayer proposed by other researchers. We viewed it as a form of prayer in which individuals make requests regarding their partner's needs, well-being, positive things, God's protection, and achievement of goals (Fincham and Beach 2014).
Prior research has widely investigated the role of prayer for a partner in relationship outcomes such as relationship satisfaction, commitment, trust, or forgiveness, which could be termed positive relationship outcomes (e.g., Fincham and Beach 2014;Fincham et al. 2008;Lambert et al. 2010bLambert et al. , 2012, and for such negative outcomes as infidelity (Fincham et al. 2010) or alcohol consumption and problematic drinking behavior (Lambert et al. 2010a). However, little is known about the role of prayer for a partner in the context of other negative relationship outcomes, such as romantic loneliness and relationship disillusionment.
Moreover, past research has mainly focused on analyzing the function of prayer for a partner in marital relationships while omitting nonmarital relationships. Therefore, a critical step is exploring the moderating role of relationship form (marital vs. nonmarital relationship) in the link between prayer for a partner and positive and negative relationship outcomes among individuals both in marital and nonmarital relationships. The marital context (i.e., the degree to which marriage is considered a normative life event in a given peer group) affects individuals' subjective well-being (Wadsworth 2016). The unique combination of marital and religious aspects in Poland, where the current investigation was performed, provides an excellent opportunity to investigate the role of prayer for a partner in relationship outcomes in marital and nonmarital unions. First, an overwhelming number of Poles identify as religious, and more than 90% of Polish individuals belong to the Catholic Church, which makes Poland a highly homogeneous society regarding religious affiliation and ethnicity (Zarzycka et al. 2017;Żemojtel-Piotrowska et al. 2021), where about 70% declare praying at least once a week (Wendołowska and Czyżowska 2021). Moreover, in Poland, the Catholic religion has been acknowledged to strongly impact people's lives in domains such as marriage and other relationships (Zarzycka et al. 2017). Second, Poles firmly hold the view of achieving adulthood through heterosexual marriage and becoming parents (Fitzpatrick et al. 2013). This adherence to traditional forms of marital and family life is accompanied by a lower prevalence of cohabitating relationships than in other countries and a lower acceptance of cohabitation (Janicka and Szymczak 2019). This high value attached to traditional forms of marital and family life in Poland is also reflected in recent Polish census data, revealing that 55% of Poles desire to have a small family, including parents and children (CBOS 2019). This percentage has not been altered since 2013, while the preference for informal relationships declined between 2013 and 2019 from 5% to 3% (CBOS 2019).
To this end, based on prior research in these areas, we hypothesized that prayer for a partner at T1 would be positively associated with relationship satisfaction at T1 and T2, three dimensions of relationship commitment at T1 and T2 (bond with the partner, significance of the relationship, and concern for the partner's well-being), and three aspects of relationship quality at T1 and T2 (perceived support, relationship depth, and relationship conflict), whereas it would be negatively associated with romantic loneliness and relationship disillusionment at T1 and T2. Furthermore, we hypothesized that these associations might appear differently for married people than their nonmarried counterparts and that the relationship form (marital vs. nonmarital relationship) would moderate these associations. To be precise, we expected that the specific differences might be, in general, exacerbated or have a greater effect in the married group than in the unmarried group.

Plan of Analysis
First, prior to the main analysis, the descriptive statistics of major study variables stratified by relationship form, i.e., by marital and nonmarital relationships, were calculated using International Business Machines Corporation (IBM; New York, NY, USA) SPSS version 27. One-way ANOVA was applied to establish differences between individuals in marital and nonmarital relationships in terms of the variables tested. Next, the correlation matrix (r-Pearson) of synchronous, cross-lagged, and autoregressive correlations for continuous study variables was calculated using SPSS v. 27.
The main analysis consisted of a series of regression moderation analyses that aimed to establish whether the relationship form at T1 (marital vs. nonmarital relationship) moderated the association between prayer for a partner at T1 and relationship outcomes at both T1 and T2. In total, we conducted 18 moderation analyses in which we tested relationship satisfaction, relationship commitment measured in terms of bond with the partner, significance of the relationship, concern for the partner's well-being, relationship quality (assessed in terms of perceived support, relationship depth, and relationship conflict), romantic loneliness, and relationship disillusionment at T1 and T2 as outcomes of prayer for a partner at T1. The moderation analyses were conducted using PROCESS (Hayes 2013). The significance of the interaction effects was tested using the bootstrapping procedure. Unstandardized interaction effects were computed for each of the 50,000 bootstrapped samples, and a 95% confidence interval was computed. Religious participation (frequency of attending religious services) was included in the moderation analyses as a covariate.
Further, following the recommendations provided in the literature, we have reported ∆R 2 as an index of moderator effect size (Carte and Russell 2003). Finally, due to testing multiple outcomes, we controlled for the family wise error rate (FWER) (Vickerstaff et al. 2019) by employing Holm's method, which iteratively accepts and rejects hypotheses (Haynes 2013). It is a step-down method whereby the unadjusted p values are ranked from smallest to largest, and each unadjusted p-value is adjusted; next, each adjusted p-value is compared to the nominal significance level until a p-value higher than the significance level is noted, after which the method is concluded (Vickerstaff et al. 2019). The Bonferroni and Holm methods lead to the same and disjunctive power when analyzing multiple primary outcomes (Vickerstaff et al. 2019). The Holm-Bonferroni correction was calculated by employing an online calculator available at https://osf.io/yv5zr/ (accessed on 16 October 2021).

Descriptive Statistics
As Table 1 shows, individuals in marital and nonmarital relationships showed a significant difference in their reported prayer for a partner at T1 and T2, bond with the partner, significance of the relationship, concern for the partner's well-being, perceived support, conflict, and depth scores based on their relationship form. In general, married individuals reported higher levels of prayer for a partner at T1 and T2, bonding with the partner at T1 and T2, the significance of the relationship at T1 and T2, concern for the partner's well-being at T1 and T2, and a higher level of depth at T2 than respondents in nonmarital relationships did. Furthermore, married respondents experienced more perceived support and conflict at T1 than did individuals in nonmarital relationships. Married individuals also reported higher religious participation (frequency of attending religious services) at T1 and T2 than individuals in nonmarital relationships. Finally, the groups did not differ in reported relationship satisfaction at T1 and T2, romantic loneliness at T1 and T2, relationship disillusionment at T1 and T2, conflict at T2, or perceived support at T2. Table 2 shows the outcomes of moderation analyses assessing the moderating role of relationship form at T1 (marital vs. nonmarital relationship) in the concurrent and temporal associations between prayer for a partner at T1 and relationship outcomes at T1 and T2. For ease of interpretation, the results of the moderation analyses were stratified by T1 and T2. Due to testing multiple outcomes, the FWER was controlled by employing the Holm-Bonferroni procedure (Haynes 2013;Vickerstaff et al. 2019). As a result, the adjusted p value was considered and subsequently interpreted. In the case of each of the nine relationship outcomes, the covariate of religious participation was demonstrated to be nonsignificant.  As Table 2 displays, there was no main effect of prayer for a partner at T1 on relationship satisfaction at T1 (p Holm = 5.43); however, there was a main effect of relationship form at T1, and marital relationships at T1 were related to lower relationship satisfaction at T1 (t = −3.179, p Holm = 0.032). The interaction between prayer for a partner at T1 and relationship form at T1 was significantly associated with relationship satisfaction at T1 (p Holm = 0.036). Specifically, prayer for a partner at T1 was significantly associated with higher relationship satisfaction at T1 for individuals in marital relationships (t = 3.20, p = 0.002) but was not significant for individuals in nonmarital relationships (t = −0.04, p = 0.969). With respect to temporal associations between prayer for a partner at T1 and relationship satisfaction at T2, there were no significant main or interaction effects (see Table 2).

Moderation Analyses: Relationship Form as a Moderator of the Link between Praying for a Partner and Relationship Outcomes
With respect to the dimensions of relationship commitment, prayer for a partner at T1 was not significantly associated with a bond with the partner at T1 or T2; there were also no main effects of relationship form at T1 or interaction effects of prayer for a partner at T1 and relationship form at T1 on bond with the partner at T1 and T2. A similar pattern of associations with the lack of main and interaction effects was also noted for the significance of the relationship at T1 and T2. Next, the main effect was determined only for relationship form at T1, for concern for the partner's well-being at T2, and marital relationships at T1 were related to greater concern for the partner's well-being at T2 (t = 3.50, p Holm = 0.018).
With respect to the dimensions of relationship quality, prayer for a partner at T1 was not significantly associated with perceived relationship support at T1 or T2; however, relationship form at T1 was significantly related to relationship support at T1, and marital relationship at T1 was related to lower perceived relationship support at T1 (t = −3.15, p Holm = 0.030). At both assessments, interactions between prayer for a partner at T1 and relationship form at T1 were not significant. Furthermore, prayer for a partner at T1 was not significantly associated with relationship depth at T1 or T2. The relationship form at T1 was not related to relationship depth at T1 or T2, and at both assessments, the interaction effects were nonsignificant (p Holm = 0.649 at T1 and p Holm = 3.24 at T2). Prayer for a partner at T1 was also not significantly associated with relationship conflict at T1 or T2, but there was a main effect of relationship form at T1 on relationship conflict at T1, and marital relationships at T1 were related to higher relationship conflict at T1 (t = 3.23, p Holm = 0.001). The interaction effects were nonsignificant in both the first and second assessments.
Prayer for a partner at T1 was not significantly associated with romantic loneliness at T1 or T2; however, there was a main effect of relationship form at T1 on romantic loneliness at T1, and marital relationships at T1 were related to higher romantic loneliness (t = 3.00, p Holm = 0.042). Additionally, the interaction between prayer for a partner at T1 and relationship form at T1 was significant for romantic loneliness at T2 (p Holm = 0.034). At the same time, the conditional effect of the focal predictor (prayer for a partner at T1) at different values of the moderator (coded as 0-nonmarital relationships and 1-marital relationships) was nonsignificant for the nonmarital relationship group (effect = 0.17, SE = 0.11, t = 1.57, p = 0.118) and the marital relationship group (effect = −0.17, SE = 0.10, t = −1.67, p = 0.097).
Finally, prayer for a partner at T1 was not significantly associated with marital/relationship disillusionment at T1 or T2. There were also no significant effects of relationship form at T1 on relationship disillusionment at T1 and T2, and interactions between prayer for a partner at T1 and relationship form at T1 on marital/relationship disillusionment at T1 and T2 were nonsignificant (see Table 2).

Participants and Procedure
Participants were recruited through advertisements posted on Facebook that included a description of the goal of the study, informed consent, and a link to the questionnaires. Participants were compensated for their participation in the study; they received vouchers to a Polish online store worth 20 PLN at T1 and 25 PLN at T2. The dataset used in the research is publicly available on the Open Science Framework at the following: https://osf. io/7e8rk/?view_only=f38d13655deb43f8a3b0d4a33bb9c07d (accessed on 16 October 2021).
In total, 446 respondents were enrolled in the study; however, four respondents did not consent to the processing of their personal data, five respondents were not in a relationship, six respondents had been in a relationship for less than six months, five respondents were of other than Polish nationality, one respondent took the survey three times, one respondent took the survey two times, and six participants were in the same relationships, resulting in dependent data.
Hence, the final sample at Time 1 consisted of 412 participants, including 235 females (57%), 165 males (40%), and 12 persons who indicated other genders (3%) with a mean age of 35.10 years (SD = 9.50) (ranging from 18 to 75 years). For the relationship form, 187 participants (45.40%) were in nonmarital relationships, and 225 were in marital relationships (54.60%). Among the 412 eligible participants at T1, 220 participants (53.40%) indicated that they did not have the possibility to pray for their partner, whereas 192 participants (46.60%) indicated that they had the possibility to pray for their partner. Table 3 displays further characteristics of the sample at T1 and T2 as a function of relationship form.
As shown in Table 3, the comparisons between individuals in marital and non-marital relationships indicate significant differences with respect to the following sociodemographic variables: age at T1, F(1,410) = 69.27, p < 0.001, η 2 = 0.15 and T2, F(1,188) = 15.97, p < 0.001, η 2 = 0.08, with younger individuals in nonmarital relationships in comparison to married individuals; education at T1, Cramer's V = 0.32, p < 0.001 and T2, Cramer's V = 0.30, p < 0.001, with a higher number of individuals with basic and secondary education and university students in nonmarital relationships compared to married individuals who, in turn, are mainly characterized by possessing higher education; living arrangements at T1, phi = −0.43, p < 0.001 and T2, phi = −0.50, p < 0.001, with a higher number of individuals in nonmarital relationships who do not live with their partners in comparison to their married counterparts; having children at T1, phi = −0.50, p < 0.001 and T2, phi = −0.46, p < 0.001, with a higher number of individuals in nonmarital relationships who do not have children compared to married individuals; religion at T1, Cramer's V = 0.27, p < 0.001 and T2, Cramer's V = 0.27, p = 0.010, with a greater number of Roman Catholic followers among married individuals in contrast to individuals in nonmarital relationships; relationship length at T1, F(1,410) = 149.30, p < 0.001, η 2 = 0.27, and T2, F(1,188) = 61.62, p < 0.001, η 2 = 0.25, with a shorter length of nonmarital relationships compared to a longer duration of marital relationships. Among 412 eligible participants, 322 (78%) indicated their willingness to participate in a follow-up study four months later. In the second assessment, 190 individuals participated, whereas 222 individuals dropped out of the study (an attrition rate = 53.88%). The detailed characteristics of individuals at T1 and T2 as a function of attrition rate are provided in Table 4.
As shown in Table 4, the comparisons between individuals who participated in the first and second assessments and individuals who participated only in the first assessment revealed significant differences with respect to the following sociodemographic variables: a form of relationship, phi = 0.13, p = 0.009, with a higher number of individuals in nonmarital relationships dropping out of the study compared to married individuals; having children, phi = −0.10, p = 0.049, with a higher number of childless individuals dropping out of the study compared to individuals who had children; religion, Cramer's V = 0.22, p < 0.001, with a higher number of individuals who were not religious followers dropping out of the study compared to Roman Catholic individuals who remained in the study at T2; and higher religious participation (frequency of attending religious services) was observed among individuals who remained in the study, F(1,410) = 9.70, p = 0.002, η 2 = 0.07. However, the effect sizes for the indicated differences were small and medium (for religion). Finally, the comparisons between these two groups with respect to the major study variables showed that participants who were in the study at T1 and T2 reported a higher level of praying for a partner (M = 2.48, SD = 1.62) than did participants who were in the study only at T1 (M = 1.95, SD = 1.44), F(1,410) = 12.36, p < 0.001, η 2 = 0.03; however, this effect size was small.

Materials
The Partner-Focused Petitionary Prayer Scale (PFPPS; Fincham and Beach 2014) was used according to the present authors' Polish translation. This scale is a four-item measure of partner-focused petitionary prayer, assessing the frequency of engagement in the behavior depicted in each item. The four statements are rated on a 5-point Likert scale, ranging from 1 ("never") to 5 ("very frequently"). In the current study, the omega coefficient for the Polish version of the PFPP was = 0.98 at T1 and 0.99 at T2. The confirmatory factor analysis demonstrated the good fit of the Polish translation at T1, χ 2 (2) = 85.40, p < 0.001, CFI = 0.97, TLI = 0.91, RMSEA = 0.04, SRMR = 0.01.
Religious participation. To assess the frequency of attendance of religious services, as was done in the study by Fincham and Beach (2014), the following item was used: "How often do you attend religious services?". Responses were rated using a 4-point scale ranging from 1 (never, or almost never) to 4 (one or more times per week). Higher scores indicated more frequent attendance.
The Relationship Assessment Scale (RAS; Hendrick 1988) was used according to the Polish translation by Monfort et al. (2014). The RAS was designed to assess general relationship satisfaction. It consists of seven items rated on a 5-point scale, ranging from 1 (not well) to 5 (very well). In the current study, the omega coefficient for the RAS was 0.89 at T1 and 0.91 at T2.
The Commitment Questionnaire (CQ; Stanley and Markman 1992) in Polish was adapted by Janicka and Szymczak (2017). It is a 19-item instrument designed to assess commitment in married and cohabitating couples that is based on two constructs in the theory of commitment, i.e., (1) dedication, which is related to the desire to maintain the relationship and concerns the sense of a bond, and (2) constraint commitment, which is defined as the tendency to involve oneself in behaviors to maintain a relationship. The Polish version of the CQ consists of three subscales: bond with the partner, the significance of the relationship, and concern for the partner's well-being (Janicka and Szymczak 2017). All 19 statements are rated on a 7-point Likert scale from 1 (strongly disagree) to 7 (strongly agree). In the current study, the Omega coefficients were as follows: 0.90 for the bond with the partner, 0.84 for the significance of the relationship, and 0.90 for the concern for the partner's well-being subscale at T1. At T2, the omega coefficients were 0.92, 0.86, and 0.92, respectively.
The Social and Emotional Loneliness Scale for Adults (SELSA-S; DiTommaso et al. Adamczyk and DiTommaso (2014). It is a 15-item instrument to measure social, emotional, and romantic loneliness. Participants are asked to respond to statements using a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). In the current study, the romantic loneliness subscale was used, for which the omega coefficient was = 0.79 at T1 and 0.84 at T2.

2004) was used according to the Polish adaptation by
The Relationship Disillusionment Scale (RDS; Niehuis et al. 2015) was used according to the Polish adaptation by Niehuis and colleagues (2021). The RDS allows us to measure relational disillusionment among marital and cohabitating unions. It consists of 11 items rated on a 5-point Likert scale from 1 (strongly agree) to 5 (strongly disagree). In the current study, the omega coefficient for relationship disillusionment was = 0.95 at T1 and 0.95 at T2.
The Quality of Relationships Inventory (QRI; Pierce et al. 1991) was used according to the Polish adaptation by Suwalska- Barancewicz et al. (2015). The QRI measures perceived support, conflict, and the depth of the relationship. The first dimension concerns the extent to which an individual can rely on a partner for help in various situations. The second dimension pertains to the respondent's participation and involvement in a relationship and how positively he or she views it. In turn, the interpersonal conflict dimension reflects the extent to which an individual experiences anger and ambivalent feelings toward a partner. Participants rated the items on a 4-point Likert scale, ranging from 1 = not true to 4 = almost always true. In the current study, the omega coefficients were 0.87 for perceived support, 0.75 for relationship depth, and 0.91 for relationship conflict at T1. At T2, the omega coefficients were 0.85, 0.77, and 0.92, respectively.
Demographic Information. Respondents were asked to indicate their age, the gender they identify with the most, their place of residence, their highest educational level, whether they have a partner/spouse, what the form of their relationship is, the duration of their relationship with their current partner, whether they live with their partner in a joint household, whether they have a child/children, what their nationality is, and what religion they follow, with the option of indicating that they are not a follower of any religion.

Discussion
The main goals of this investigation were to determine whether prayer for a partner in the first assessment was concurrently and prospectively (in a four-month interval) associated with diverse relationship outcomes and to ascertain whether relationship form (marital vs. nonmarital relationship) moderates these associations. This study focused both on relationship outcomes (relationship satisfaction, commitment, and conflict) whose associations with praying for a partner have been examined in prior research (e.g., Fincham and Beach 2014;Fincham et al. 2008) and on those relationship outcomes that have been, to our knowledge, examined for the first time in reference to praying for a partner, i.e., bond with the partner, significance of the relationship, concern for the partner's well-being, perceived support, relationship depth, romantic loneliness, and relationship disillusionment. Because the performed moderation analyses provided numerous findings, in the current paper we have, however, focused on the essence of the present investigation, i.e., on the exploration of the concurrent and temporal associations between prayer for a partner and relationship outcomes and the moderating role of relationship form for these associations. By interpreting the obtained findings, we followed recommendations provided for moderation analyses; i.e., we interpreted main effects (i.e., effects of prayer for a partner) merely when the moderating effects were determined to be insignificant (Carte and Russell 2003), and we interpreted significant interaction effects as key coefficients even when the main effects of prayer for a partner and relationship form were determined in some cases to be insignificant since the purpose of the current study was to assess an interaction (see, for discussion, Bedeian and Mossholder 1994).
The concurrent association between prayer for a partner at the first assessment and relationship satisfaction at the first assessment was shown to hold true only for individuals in marital relationships but not for nonmarried individuals. The effect size of this moderation effect was, however, small and explained 2% of the variance in relationship satisfaction. This pattern of findings is congruent with previous studies showing that prayer for a partner was associated with higher satisfaction in relationships (e.g., Fincham and Beach 2014;Fincham et al. 2008). For instance, Fincham and colleagues (2008) observed, in a sample of 302 U.S. students (M age = 19.99), a positive, weak correlation between prayer for a partner and current relationship satisfaction (r = 0.15) and relationship satisfaction measured six weeks later (r = 0.14). A similar pattern of results was also observed in another study in which Fincham and Beach (2014) found, in a sample of 316 U.S. undergraduate students (M age = 19.05 and the mean duration of the relationship was 13.70 months), a moderate positive correlation (r = 0.20) between prayer for a partner and relationship satisfaction measured by the Couple Satisfaction Inventory (CSI). At the same time, our results provided evidence that the positive association between prayer for the partner and relationship satisfaction may pertain only to individuals in marital relationships. This means that for individuals in nonmarital unions, praying for a partner is not related to relationship satisfaction. Two factors might explain this finding.
First, in the current study, a higher percentage of married participants in T1 and T2 declared their belonging to the Roman Catholic religion compared to the lower percentage of participants in nonmarital relationships who belonged to the Roman Catholic religion. In addition, married participants at T1 and T2 reported a higher frequency of religious participation, i.e., a higher frequency of attending religious services, than their nonmarried counterparts. The higher rate of Roman Catholics among married individuals compared to individuals in informal relationships and higher religious participation among married individuals may imply higher religiosity and a more important role of religion in their lives than nonmarital individuals. Moreover, according to the Catholic Church, the presence of cohabiting unions offends the dignity of marriage, destroys the very concept of family, and weakens the meaning of fidelity, and cohabitation is contrary to moral law (John Paul 1981. Since Catholic teaching values marriages and strongly rejects cohabitation, Roman Catholic married individuals may experience enhanced positive relationship outcomes. In turn, individuals in informal relationships in Poland may experience religious and spiritual struggles concerning wrestling with attempts to follow moral principles and feelings of worry or guilt arising from perceived offenses committed (Zarzycka et al. 2020). Specifically, Roman Catholics in informal relationships may experience the discrepancy between Roman Catholic ideals of marriage and the disapproval of cohabitation in which they remain. In this specific religious and marital context of Poland, where marriage and other relationships are under the strong influence of the Catholic religion (Zarzycka et al. 2017;Żemojtel-Piotrowska et al. 2021) and where a lower prevalence of cohabitation and its lower acceptance rate compared to other countries is noted (Fitzpatrick et al. 2013;Janicka and Szymczak 2019), individuals in nonmarital relationships may not benefit from the practice of prayer to the same positive degree as individuals in marital relationships.
The second interaction effect of prayer for the partner at T1 and relationship form at T1 was observed for romantic loneliness at T2 after four months. To be precise, a higher level of prayer at T1 was related to a lower level of romantic loneliness after four months. This effect was small and explained 5% of the variance. At the same time, although the interaction between focal predictor X (prayer for the partner at T1) and moderator M (relationship form at T1) was statistically significant, we did not find evidence of a statistically significant conditional effect of X (prayer for the partner at T1) at any value of M (relationship form at T1 coded as 0-nonmarital relationships and 1-marital relationships) (for discussion, see Hayes 2013, p. 316). Therefore, the obtained results allow us only to claim that the effect of prayer for the partner at T1 depends linearly on the relationship form at T1. However, we are not able to indicate for which individuals (married or nonmarital individuals) the effect of prayer for the partner at T1 on romantic loneliness at T2 is different from zero (see Hayes 2013, p. 316). The statistics of the interaction effect, however, imply that this effect may be significant for married but not nonmarried individuals. To the best of our knowledge, the current study is among the few studies to examine the association between prayer for partners and the romantic loneliness experienced in relationships. Based on prior research investigating the role of prayer in a relationship, we expected that higher levels of praying for one's partner would be associated with lower levels of romantic loneliness. Prior research has shown that prayer for partners and religiosity, in general, may enhance relationship functioning (e.g., Beach et al. 2011;May et al. 2020). Moreover, praying in marriage has been considered to enable one to regain perspective, break a negative thought cycle, promote relaxation, and provide an opportunity for dialog with a supportive other ; furthermore, it appears to be advantageous due to engagement in such behaviors that enrich relationship functioning . In light of our findings, prayer for the partner also enhances relationship functioning by decreasing the levels of romantic loneliness reported after four months, which is a result of the experienced discrepancy between the desired and current relationship quality (e.g., Mund and Johnson 2021).
With respect to relationship commitment, we measured this construct in terms of three dimensions, namely, bonding with the partner, the significance of the relationship, and concern for the partner's well-being. Prior studies analyzing the link between prayer for a partner and relationship commitment found, for instance, in a sample taken from the U.S. with a total of 360 undergraduates, that prayer for one's partner was related to higher levels of dedication commitment , as well as to later dedication commitment (in a period of seven weeks) (Fincham and Beach 2014). In contrast, our results did not provide evidence for the associations between prayer for a partner and three aspects of relationship quality or for interaction effects. Next, in the current study, relationship quality was assessed in three domains, namely, perceived support, relationship conflict, and relationship depth. Prior studies have demonstrated that, in general, religiosity plays a crucial role in relationship maintenance for marriages (Langlais and Schwanz 2017) and for nonmarital and dating relationships (Aragoni et al. 2021;Henderson et al. 2018;Langlais and Schwanz 2017). Moreover, prayer for a partner has been recognized in past research as contributing to the improvement of relationship quality (e.g., Beach et al. 2011;May et al. 2020). Our results demonstrated, however, the lack of main and interaction effects of praying on three dimensions of relationship commitment, relationship quality, and relationship disillusionment. The nonsignificant effects of prayer for the partner and relationship form on relationship outcomes enumerated above, determined by employing the adjusted p values, may be explained by the existence of very small effects of prayer for the partner and relationship form for these relationship outcomes that were not able to be detected due to the current sample size at T1 (N = 412) and T2 (N = 190), as the sample size is one of the factors influencing the statistical power of moderation analyses (Aguinis 1995). Additionally, our results may imply the salient role of prayer for a partner for only selected relationship outcomes, i.e., for relationship satisfaction and romantic loneliness, but not for other relationship outcomes involving a bond with the partner, the significance of the relationship, and concern for the partner's well-being as a measure of relationship commitment, relationship quality, and relationship disillusionment.
Finally, in the current study, we did not observe any temporal associations between prayer for the partner and relationship outcomes assessed four months later, except for romantic loneliness. In prior research, Fincham and colleagues (2008) found that prayer for the partner predicted later relationship satisfaction (measured after six weeks) but not vice versa; however, this link appeared insignificant when later prayer for the partner was introduced in the model . Based on these results, Fincham and colleagues (2008) concluded that prayer may affect relationship satisfaction in periods shorter than 6-week intervals. In light of this notion, the lack of temporal associations determined in the current study between payer for the partner and relationship outcomes satisfaction may result from the four-month interval duration that occurred, which is a relatively long interval when trying to capture the effects of earlier praying for later relationship outcomes. Additionally, it is probable that the majority of interaction effects might not have been detected due to the unequal sample size across moderator-based subgroups at the second assessment (the subgroup of 117 married individuals was almost two times larger than the subgroup of 73 nonmarital individuals), which could violate the homogeneity of within-moderator-based subgroup error variances and significantly decrease the power of analysis (see Aguinis 1995).
There are some limitations that must be considered when interpreting the findings of the present study. First, the participants were recruited via Facebook; thus, individuals who do not use the internet and/or social media were excluded from participation, and such recruitment implies a self-selection bias (e.g., Strange et al. 2018). In addition, the recruitment via Facebook might have affected the age of participants since the largest groups of Facebook users in Poland involved individuals aged 25-34 years old (Chabrzyk et al. 2020). Future research might consider applying other forms of recruitment of participants. Second, the current investigation focused only on assessing the associations between prayer for a partner and relationship outcomes, and mental and physical health indicators have not been measured. Therefore, future research ought to extend the scope of the current investigation by the inclusion of various indicators of mental health such as depression, anxiety, subjective well-being, and physical health, for instance, in terms of physician ratings of individuals' functioning, individuals' self-reported health, or objective biological markers (e.g., blood pressure, cholesterol levels) (see for discussion Robles 2014;Robles et al. 2014) and by testing the mediation models in which relationship outcomes would be examined as mediators of the link between the prayer for a partner and health outcomes.
Third, although the current study employed both a cross-sectional methodological design and a longitudinal design, it faced a higher rate of attrition (53.88%) between the first and second assessments, which undermines the robustness of the current study's results. However, it may be worth noting that differences between participants who remained in both assessments and participants who dropped out at the second assessment concerned only a few aspects (a form of relationship, having children, religion, and praying for a partner), while the effect size for these differences was small and moderate in magnitude. Furthermore, although the interval between the two assessments was relatively short, it may have been too long to capture the effect of praying for the partner that may unfold in shorter periods . Thus, future research would benefit from employing shorter intervals between measurements. Fourth, the current study analyzed only the data gathered from one partner in a given couple, thereby hindering the possibility of investigating the effects of partner similarity in praying for a partner on relationship outcomes. Therefore, future studies should address these limitations by employing, for instance, a dyadic response surface analysis, which would allow the investigation of the effects of dyadic similarity of a given characteristic (e.g., Mund and Johnson 2021), such as praying for a partner, on relationship outcomes. Fifth, in the current investigation, we did not measure the object of petitionary prayer, and various intercession objects are likely to affect relationship outcomes differently. For instance, praying that a partner is successful in a job interview differs from praying for a partner with a terminal health diagnosis. Therefore, future research will benefit from assessing the object of the intercession of praying in relationships. Sixth, it must be emphasized that among religious individuals, a Christian religious affiliation was endorsed by many participants (37.60%); therefore, it is unclear to what extent the obtained results can be applied to non-Christian denominations. In addition, there remains a lack of a direct assessment of religiosity in the current investigation, as the assessment of the frequency of attendance of religious services employed in the current study may not be considered a comprehensive measure of religiosity. Thus, future research could utilize more diverse samples with a greater representation of non-Christian denominations and employ multidimensional measures of religiosity. Finally, the current investigation was performed during the COVID-19 pandemic, which has contributed to several mental health problems (Thomas and Barbato 2020). Therefore, it is reasonable to assume that the COVID-19 pandemic might have affected the experiences of the participants in the current study and thereby attuned them to the obtained results.

Conclusions
In conclusion, the obtained findings suggest that the effects of prayer for the partner depend on relationship form for two out of nine relationship outcomes analyzed in the current investigation, i.e., relationship satisfaction at the first assessment and romantic loneliness measured four months later. Furthermore, the small effect size of prayer for the partner and interactions between prayer for the partner and relationship form on relationship satisfaction and romantic loneliness suggest that prayer for the partner is only one of the factors related to relationship functioning (e.g., Henderson et al. 2018). Furthermore, the general lack of a temporal link between prayer for the partner and later relationship outcomes (measured in the current study over four months), except for romantic loneliness, appears to suggest that the effects of such prayer may occur at shorter intervals (see Fincham et al. 2008). Finally, the major conclusions pertain to the moderating role of prayer for partners in relationship satisfaction and romantic loneliness. Our findings suggest that prayer for the partner may be particularly beneficial for married individuals compared to individuals in nonmarital relationships. Hence, our findings emphasize the necessity of considering praying for the partner while taking into account the individuals' relationship context, i.e., whether the individuals are in marital or nonmarital relationships. Thus, religion and marital context appear to be important factors in understating the role of prayer for the partner in regard to relationship outcomes that may affect the mental and physical health of coupled individuals.
Author Contributions: K.A. and S.P. were involved in the study's conceptualization and design; S.P. contributed to data management; K.A. was involved in the statistical analyses; K.A. interpreted the data; K.A. and S.P. were involved in funding acquisition; K.A. and S.P. drafted the manuscript and read and approved the final manuscript. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare that there is no conflict of interest.