Closeness Discrepancies in Couple Relationships: A Dyadic Response Surface Analysis

Abstract

Individuals can differ in the degree of closeness they desire in their romantic relationships: Some people may perceive their current level of closeness as just right, whereas others may feel not close enough or too close to their partners (referred to as negative and positive closeness discrepancy, respectively). This study (N = 1,177 individuals from 748 couples) examined the implications of closeness discrepancies for subjective relationship quality (SRQ) using dyadic response surface analysis. The analyses found evidence for linear, but not broad, closeness discrepancy effects: SRQ was lower for individuals reporting more negative closeness discrepancies and, independent of this actor effect, for individuals with partners who reported more negative closeness discrepancies. These results suggest that low levels of closeness paired with a strong desire for closeness can impair both partners’ relational well-being.

Keywords

closeness discrepancy couple relationships response surface analysis

The extent to which romantic partners feel close to each other is a key indicator of couples’ happiness and satisfaction. Numerous studies show that higher closeness is associated with higher relationship satisfaction and commitment (Agnew et al., 1998; Aron et al., 1992; Frost & Forrester, 2013), as well as lower probability of relationship break-up and cheating (Aron et al., 1992; Frost & Forrester, 2013; Le et al., 2010). According to self-expansion theory (Aron et al., 2013), closeness represents the degree to which individuals expand their self-concepts to incorporate their partners’ resources, identities, and perspectives. This so-called Inclusion of Other in the Self is supposed to help individuals fulfill their self-expansion motives aimed at, for instance, learning new things, becoming a better person, or gaining new perspectives. Including the partner in the self manifests in a greater sense of we-ness (Agnew et al., 1998), providing relational benefits throughout the course of the relationship, from initial attraction to long-term maintenance and stability of the relationship (for a literature overview, see Aron et al., 2013).

However, people can interindividually differ in how much closeness they typically desire and seek in their relationships (Hagemeyer, Neyer, et al., 2013). While some people feel as close to their partners as they would like, others may feel not close enough or even too close. Thus, to better understand closeness in couple relationships, it is important to relate partners’ actual experiences of closeness to their desired experiences of closeness. Previous research suggests that discrepancies between actual and desired closeness are detrimental to relationship satisfaction and well-being; individuals with matching actual and desired experiences of closeness, in turn, tend to report highest relationship quality (Frost et al., 2017; Frost & Forrester, 2013; Mashek & Sherman, 2004). The current study examined whether these findings can be replicated using response surface analysis (RSA; Edwards & Parry, 1993; Schönbrodt et al., 2018), which overcomes important limitations of approaches previously used to analyze closeness discrepancy effects (i.e., difference scores). Furthermore, we extend previous research by adopting a dyadic perspective, allowing us to examine how participants’ closeness discrepancies relate not only to their own but also to their partners’ relational well-being.

Implications of Closeness Discrepancies for Relationship Functioning

Actual and desired experiences of closeness can combine in three ways. First, people may perceive their closeness to their partners as just right. This is the case when actual experiences of closeness perfectly match the desired amount of closeness. Second, individuals can feel not close enough to their partners when their actual experiences fall below their desired amount of closeness (henceforth referred to as negative closeness discrepancies). Third, people can feel too close to their partners when their actual experiences of closeness exceed the degree of closeness they desire (henceforth referred to as positive closeness discrepancies).

Previous research on closeness discrepancies points to the general relational benefits of matching actual and desired experiences of closeness with the partner. A study by Frost and Forrester (2013) found that individuals with more matching actual and desired closeness levels report higher relationship satisfaction, higher commitment, and fewer break-up thoughts. In comparison, both greater negative and positive closeness discrepancies were associated with lower satisfaction and commitment, and more frequent break-up thoughts. Importantly, these effects of closeness discrepancies were independent of participants’ actual levels of closeness, suggesting that “Accounting for individuals’ actual experiences of closeness in romantic relationships is necessary but not sufficient in effectively explaining their relational well-being, stability, and mental health” (Frost & Forrester, 2013, p. 468). The few other studies on closeness discrepancies in couple relationships support this interpretation. Mashek and Sherman (2004), for example, found desiring less closeness to be associated with reduced relationship satisfaction, commitment, and partner-related passion. Moreover, sex-related closeness discrepancies have been linked to lower sexual satisfaction and lower orgasm frequency (Frost et al., 2017). Gamarel and Golub (2018), in contrast, showed that men who desire more closeness than they actually experience more readily engage in HIV risk prevention behavior.

The Ideal Standards Model (Fletcher & Simpson, 2000) provides insight into how closeness discrepancies can impair relationship quality. The model posits that individuals hold relatively stable ideal conceptions of themselves, their partners, and their relationships. These ideals function as guiding points against which individuals compare their actual relationship experiences. The degree to which actual experiences and desired ideals match can then inform about the quality of the relationship and partner, as well as reveal where adjustments may be needed. People should be most satisfied when their experiences reach their desired ideals, whereas dissatisfaction and unhappiness should result from a discrepancy between experiences and ideals.

Similar predictions can be made based on motivational set-point theories. The Zurich model of social motivation (Bischof, 1975), for instance, argues that people continuously compare their actual situation (e.g., how close they currently feel to their partners) against a relatively fixed set-point of experiences they typically require to feel satisfied (e.g., how much closeness they typically want). A discrepancy between actual experiences and the desired set-point is assumed to energize motivational processes aimed at decreasing the discrepancy and restoring a state of homeostasis. If the actual experience falls below the set-point, motivational appetence sets in. For example, a person should develop the need to establish more closeness to their partner when their actual experiences of closeness do not suffice to satisfy their desired closeness (i.e., a negative closeness discrepancy). In turn, motivational aversion will follow a positive discrepancy. When the actual experience of closeness exceeds the set-point, the person experiences the need to avoid closeness. Importantly, appetence and aversion come along with negatively valenced affect that likely rubs off on overall relationship evaluations. For instance, being not close enough to the partner (i.e., closeness appetence) due to physical or emotional distance may fuel feelings of longing and, in the long run, feelings of loneliness. Partner-related loneliness, in turn, is negatively linked to relationship quality (Mund & Johnson, 2020). Feeling too close to the partner (i.e., closeness aversion), on the other hand, may spark feelings of confinement, intrusion, and annoyance by the partner (Mashek et al., 2011; Mashek & Sherman, 2004), which may likewise impair relationship quality. In sum, we expected that individuals with negative or positive closeness discrepancies evaluate their relationships and partners in a worse light compared with individuals whose actual experiences meet their desired amount of closeness with their partners.

Closeness experiences are dyadic and interdependent at heart. Thus, to more comprehensively understand the implications of closeness discrepancies for relationship functioning, researchers may need to adopt a dyadic perspective and additionally investigate how a person’s relational well-being is affected by their partner’s closeness discrepancies. Such investigations are currently lacking. However, interpersonal closeness discrepancy effects are likely. On the one hand, individuals feeling not close enough to their partners (i.e., who perceive a negative closeness discrepancy) may demand more closeness from their partners and, as a consequence, engage in clinging (and potentially harmful) behavior. Feeling too close (i.e., perceiving a positive closeness discrepancy) may, on the other hand, propel individuals to distance themselves from their partners and put their own interests above their partner and the relationship, likely inducing dissatisfaction on the partner’s side (Hagemeyer et al., 2015; Hagemeyer & Neyer, 2012). Hence, we expected that individuals’ closeness discrepancies would demonstrate negative effects not only on their own but also on their partners’ relational well-being.

Analysis of Discrepancy Effects

Existing research on closeness discrepancies in couple relationships has exclusively operationalized closeness discrepancies via difference scores (Frost et al., 2017; Frost & Forrester, 2013; Gamarel & Golub, 2018). Difference scores are calculated by subtracting participants’ desired closeness levels from their actual closeness levels. Difference values greater than 0 (i.e., positive closeness discrepancies) thus indicate feeling too close, whereas difference scores below 0 (i.e., negative closeness discrepancies) indicate not feeling close enough to the partner. These difference scores are then used to predict relationship outcomes such as relationship satisfaction. To estimate the effects of negative and positive closeness discrepancies separately, participants’ difference scores can be subdivided into negative and positive scores and then analyzed with segmented regression (Keele, 2008). Frost and Forrester (2013) used this approach to demonstrate the negative implications of both negative and positive closeness discrepancies for relationship functioning and mental health. Negative effects on relationship outcomes have also been found in research using absolute difference scores, which assess the magnitude of closeness discrepancies in either direction (Frost et al., 2017; Gamarel & Golub, 2018). When analyzing closeness discrepancy effects, it is important to consider whether these effects are independent of the overall level of closeness (i.e., independent of whether the closeness discrepancy occurs at the lower or upper end of the closeness scale). Past research has done this by controlling for participants’ actual closeness levels in the prediction.

Difference scores have been criticized for their methodological and interpretative problems (Edwards, 2002; Edwards & Parry, 1993). Two major issues stand out. First, difference scores are an exact function of their two component variables (e.g., actual and desired closeness). As predictors, closeness difference scores do not explain any additional variance beyond actual and desired closeness; any effects of closeness difference scores capture nothing more than the simple effects of actual and desired closeness (Edwards, 2002). Rather, using closeness difference scores may actually confound the effects of the two components. In some cases, one component can drive the entire effect of the difference score, resulting in considerable issues with the interpretation of the results (for exemplary simulations, see Edwards, 2007). This problem becomes even more apparent when participants’ actual closeness levels are additionally controlled for, as was common practice in previous studies on closeness discrepancies in couples (Frost et al., 2017; Frost & Forrester, 2013; Gamarel & Golub, 2018). Controlling for actual closeness transforms the closeness difference score into a partialed measure of desired closeness (Edwards, 2002). As a consequence, any incremental effect of closeness difference scores will primarily reflect the contribution of desired closeness.

Second, using closeness difference scores as predictors implicitly constrains the effects of its two components (actual and desired closeness). Consider the following equation in which closeness difference scores are used as a predictor:

o u t c o m e = b_{0} + b_{1} closeness difference + ε

(1)

Because closeness difference scores are computed by subtracting desired from actual closeness, Equation 1 can be rewritten as follows:

\begin{array}{l} o u t c o m e = b_{0} + b_{1} (actual closeness - \\ desired closeness) + ε \\ = b_{0} + b_{1} actual closeness - \\ b_{1} desired closeness + ε \end{array}

(2)

Thus, combining actual and desired closeness into a difference score constrains their effects to be equal but opposite. For instance, if the effect of actual closeness in Equation 2 would be estimated as $+ . 50$ , the effect of desired closeness must amount to $- . 50$ . This implicit assumption of equal but opposite contributions is quite strict and may therefore fail to adequately describe the relevance of actual and desired closeness for relationship functioning.

There are additional shortcomings of difference scores that are beyond the scope of this article, such as reliability issues and dimensional reduction (for a detailed discussion, see Edwards, 2002; Edwards & Parry, 1993). Because of their many shortcomings, difference score models are nowadays considered inappropriate for analyzing discrepancy effects. Instead, RSA is increasingly recommended as the tool of choice for testing discrepancy hypotheses (Barranti et al., 2017; Edwards & Parry, 1993; Humberg, Nestler, & Back, 2018; Schönbrodt et al., 2018). RSA is based on polynomial regression, which overcomes several problems associated with difference scores. For instance, the effects of actual and desired closeness are no longer constrained to be equal and opposite, but can be freely estimated. Moreover, RSA can account for three-dimensional relationships between actual closeness, desired closeness, and relationship quality. This may uncover more complex associations than difference scores could capture.

The Present Study

The goal of the present study was to examine whether previous findings on the relevance of closeness discrepancies for couple relationships replicate when using RSA. Using dyadic data from a large sample of 1,177 individuals nested in 748 couples, we examined how different combinations between actual and desired levels of closeness relate to subjective relationship quality (SRQ). We first expected that overall higher levels of closeness, that is, higher levels of both actual and desired closeness, would positively contribute to relationship quality (Aron et al., 1992; Frost & Forrester, 2013; Hagemeyer, Neberich, et al., 2013; Hagemeyer, Neyer, et al., 2013; Le et al., 2010; Zygar et al., 2018). Beyond these main effects, relationship quality should also depend on the extent to which actual and desired closeness (mis)match. Based on previous research (Frost et al., 2017; Frost & Forrester, 2013; Mashek & Sherman, 2004), we expected individuals with greater positive or negative closeness discrepancies, that is, who desire more or less closeness than they actually experience, to report lower SRQ than individuals whose actual and desired closeness match. We took full advantage of our dyadic data and addressed these hypotheses at both the intrapersonal and interpersonal levels: Using dyadic RSA (Schönbrodt et al., 2018), we examined how individuals’ SRQ relates not only to their own but also to their partners’ actual-desired closeness combinations. In this regard, the current study presents an important extension of existing research, which has been limited to intrapersonal associations between closeness discrepancies and relationship outcomes.

Method

Information on Data and Materials

The data used in this study have been used in a previous publication of accuracy and bias in the perception of communal motive dispositions (Pusch et al., 2021). The data are available as a scientific-use file published along with release 12.0 of the German Family Panel Pairfam. Access to the data is managed by the Pairfam administration (for information, see https://www.pairfam.de/en/). For the current study, we set up a permanent online repository (https://osf.io/9njpz/) containing reproducible analysis scripts, materials and measures, and a codebook for all used variables.

Participants and Procedure

We used data from a research project on motivation in couple relationships that was carried out in cooperation with the German Family Panel (pairfam; Huinink et al., 2011). Pairfam is an ongoing panel study with a representative German sample of participants from three birth cohorts (born in 1971–1973, 1981–1983, and 1991–1993). Data for the current analyses were collected in two phases in 2016/2017. First, pairfam participants and their current partners were recruited via postal and electronic mail. Second, to gain a larger sample, additional participants of the same age range were recruited via online advertising, flyer distribution, university email lists, and websites for people interested in participating in psychological research (PsyWeb: www.psyweb.uni-muenster.de; Psytests: www.psytests.de).

Participation required being involved in a romantic relationship, being at least 18 years old, and speaking German fluently. The study was administered entirely online via questionnaires developed on the formr survey platform (Arslan et al., 2019). After couples registered for participation, each partner received a personalized link to the online questionnaire via email. Couples were instructed not to discuss the questionnaire before both partners had completed it (median completion time was 41 min). All participants who completed the questionnaire were compensated with personalized feedback on their partner-related motives and given the option to enter a draw of 55 shopping and travel vouchers worth a total of €9,000.

The questionnaire was completed by 1,216 participants. Following our preregistration (see https://osf.io/anvpj), we excluded 39 participants who were younger than 18 years $(n = 3),$ were in a same-sex relationship $(n = 35),$ or did not identify as either male or female $(n = 1) .$ The remaining final sample comprised 1,177 participants (56% female) nested in 748 couples. Participants reported a mean age of 31.48 years (SD = 7.88) and an average relationship duration of 7.55 years (SD = 6.76). Most participants held a high school degree (German Abitur; 70%) and were childless (68%) at the time of assessment.

The present study is the first to carry out dyadic analyses of closeness discrepancies in couple relationships using RSA. Formal power analyses were thus not feasible because we had no expected (co)variances and effect sizes to base them on. However, in planning this study, we aimed for a sample size large enough to detect between-couple effects of average size ( $r = . 21;$ Richard et al., 2003) with a power of at least 80%. This required at least 175 couples. The final sample exceeded this benchmark by far. The current analyses were conducted after data collection was finished.

Measures

Actual and desired closeness

The Inclusion of Other in the Self scale (IOS; Aron et al., 1992) was used to assess participants’ actual and desired closeness to their partners. Because the IOS is a single-item scale, we cannot estimate its reliability in the current study. However, previous research found the IOS scale to demonstrate good psychometric qualities such as high parallel-test reliability and strong convergent validity with related measures of closeness in couple relationships (e.g., Agnew et al., 1998; Aron et al., 1992, 1997); moreover, it has already been applied in previous research on closeness discrepancies (Frost et al., 2017; Frost & Forrester, 2013; Gamarel & Golub, 2018). The scale consists of seven figures depicting the self and the partner as increasingly overlapping circles, ranging from barely tangent to almost completely overlapping. The IOS scale was presented twice. In a first step, participants were asked to indicate which of the seven figures best represented how close they felt to their partners at the moment. This rating thus captured participants’ actual experiences of closeness to their partners. In a second step, participants were prompted to select the figure that best described how much closeness they desired in their relationship, thus assessing participants’ desired closeness to their partners. Both actual and desired closeness scores could range from 1 to 7 with higher scores indicating higher closeness.

Subjective relationship quality

SRQ is a multifaceted construct including, for instance, relationship satisfaction, sexual satisfaction, or commitment. These facets capture distinct but interrrelated aspects of relationship functioning: Although a person may evaluate certain aspects of their relationships differently, these evaluations all seem to be rooted in a general appraisal of the relationship (Fletcher et al., 2000; Hagemeyer et al., 2015). In the current study, we broadly assessed SRQ with the following six variables:¹

Relationship satisfaction was assessed with two items: “All in all, how satisfied are you with your relationship?” (answer options ranged from 0 = very dissatisfied to 10 = very satisfied) and “I am unhappy in my relationship” (reverse-scored; answer options ranged from 0 = not at all to 10 = absolutely). Participants’ also rated their sexual satisfaction on two items: “All in all, how satisfied are you with sexuality in your relationship?” (answer options ranged from 0 = very dissatisfied to 10 = very satisfied) and “My partner fulfills my sexual needs” (answer options ranged from 0 = not at all to 10 = absolutely). Loneliness in the relationship was measured with three items. On a 7-point scale (ranging from 1 = never to 7 = always), participants rated their agreement to the following statements: “I feel isolated from my partner,” “I feel lonely in my relationship,” and “I miss security and warmth in my relationship.” To assess perceived partner support, a single-item measure was used. On a scale ranging from 1 = never to 7 = always, participants rated their agreement to the following question: “When I have a problem, I can rely on my partner’s support.” Participants’ commitment was measured with two items: “Can you imagine your current partner being your partner for life?” and “In case of serious problems with my partner, I can imagine separating” (reverse-scored). Answers were given on a 5-point scale ranging from 1 = no, not at all to 5 = yes, absolutely. Finally, two items were used to measure participants’ self-disclosure toward their partners: “How often do you tell your partner what you’re thinking?” and “How often do you share your secrets and private feelings with your partner?” Self-disclosure frequency ratings were given on a 5-point scale ranging from 1 = never to 5 = always. Information on the sources and validity of the single items is provided in the Supplemental Materials (Table S2).

To compute overall SRQ scores, the six variables (loneliness was reversed) were first z-standardized using their respective means and standard deviations in the total sample, and then averaged. The six variables showed high internal consistency for both men (α = .81) and women (α = .85).² Confirmatory factor analyses confirmed that the variables loaded on a common factor; in models with acceptable fit to the data (men: $χ^{2} = 29.272$ , $d f = 9$ , $p = . 001$ , comparative fit index $[CFI] = . 967$ , root mean square error of approximation $[RMSEA] = . 066$ , standardized root mean square residual $[SRMR] = . 027$ ; women: $χ^{2} = 37.233$ , $d f = 9$ , $p < . 001$ , $CFI = . 973$ , $RMSEA = . 069$ , $SRMR = . 027$ ; Marsh et al., 2005), standardized loadings ranged from 0.52 to 0.77 for men and from 0.57 to 0.83 for women (see Table S1 in the Supplemental Materials).

Analysis Strategy

We tested our hypotheses with dyadic RSA (Edwards & Parry, 1993; Schönbrodt et al., 2018). RSAs are based on polynomial regression models in which an outcome is regressed on two predictor variables, their squares, and the product interaction between the linear predictors. In our model, participants’ SRQ was regressed on their actual closeness ( $b_{1}$ ) and its square ( $b_{3}$ ), their desired closeness ( $b_{2}$ ) and its square ( $b_{5}$ ), and the interaction between the two linear predictors ( $b_{4}$ ).³ In RSA, it is important for the two predictor variables to share a common and meaningful zero point. We therefore z-standardized actual and desired closeness using the grand mean ( $M_{grand} = 5.63$ ) and grand standard deviation ( $S D_{grand} = 1.24$ ) across both variables (Schönbrodt et al., 2018).

To account for the dyadic structure of the data, the polynomial regression model was extended to an Actor–Partner Interdependence Model (APIM; Kenny et al., 2006). The APIM is a well-established tool for the analysis of couple data. It accounts for dyadic nonindependence and allows the estimation of the unique effects of couple members’ predictor variables on their own outcomes (i.e., actor effects) and on their partners’ outcomes (i.e., partner effects). Running the polynomial regression analysis in an APIM framework thus allowed us to simultaneously predict participants’ SRQ by their own and their partners’ set of polynomial terms (Schönbrodt et al., 2018). In line with recent recommendations (Schönbrodt et al., 2018), we chose the APIM because we were interested in individual-level phenomena, namely, the associations of partners’ SRQ with their own and each other’s unique actual and desired closeness levels (Galovan et al., 2017).

Based on the results of the polynomial regression analysis, a response surface can be computed detailing how different combinations of actual and desired closeness relate to SRQ. Figure 1 shows an exemplary response surface plot. The estimated polynomial regression coefficients can be used to calculate the linear and quadratic term coefficients of two lines that describe the overall shape of the response surface: the line of congruence (LOC; the line where actual closeness = desired closeness) and the line of incongruence (LOIC; the line where actual closeness = −desired closeness). In the present study, a positive linear term coefficient of the LOC $(a_{1} = b_{1} + b_{2})$ would signify a positive additive main effect in that participants with higher actual and desired closeness report higher SRQ. A positive quadratic term coefficient of the LOC $(a_{2} = b_{3} + b_{4} + b_{5})$ would indicate that SRQ further increases in the area of extreme values of the two predictor variables. A positive linear term coefficient of the LOIC $(a_{3} = b_{1} - b_{2})$ would signify that SRQ is higher when participants rate their actual closeness to exceed their desired closeness; this would reflect in a response surface that is shifted along the LOIC to the lower right corner. A significantly positive quadratic term coefficient of the LOIC $(a_{4} = b_{3} - b_{4} + b_{5})$ would indicate that SRQ is higher the more actual and desired closeness mismatch, whereas a negative curvature would point to lower SRQ for greater mismatches between the two predictors.

Figure 1.

Example response surface.

We hypothesized that participants would report higher SRQ the higher their overall levels of actual and desired closeness and the more both actual and desired closeness match. Greater actual-desired closeness discrepancies, in turn, should be associated with lower SRQ. In RSA, this translates to a broad discrepancy effect as indicated by a specific constellation of response surface parameters (Humberg, Nestler, & Back, 2018; Nestler et al., 2019):

$a_{1}$ must be significantly positive,

$a_{4}$ must be significantly negative,

$a_{2}$ and $a_{3}$ must not significantly differ from 0.

In addition, a broad discrepancy effect also requires the LOC to equal the line with the highest predicted outcome values (i.e., the first principal axis). If all of the above conditions hold, a fifth response surface parameter ( $a_{5} = b_{3} - b_{5}$ ) can be computed to test whether the first principal axis lies on the LOC; the $a_{5}$ parameter must not significantly differ from 0 to corroborate a broad discrepancy effect (Nestler et al., 2019).

All analyses were carried out in the R statistical environment (R Core Team, 2020). Data preparation, handling, and visualization were done with the R-packages dplyr (Wickham et al., 2019), ggplot2 (Wickham, 2016), cowplot (Wilke, 2020), psych (Revelle, 2018), and RSA (Schönbrodt & Humberg, 2018). The dyadic polynomial regression models were fitted using the structural equation modeling package lavaan (Rosseel, 2012). We handled missing data by means of full information maximum likelihood procedures and calculated standard errors with a robust maximum likelihood estimator. Couples’ log-transformed relationship length was controlled as a dyadic covariate (couple members’ raw relationship length reports were highly correlated, $r = . 93$ ). Because we had no a priori reasons to expect sex differences in actor and partner effects, model paths were set equal between male and female participants. The assumption of indistinguishability between male and female participants was supported by omnibus tests of sex differences (Kenny et al., 2006): We specified path models in which all predictor and outcome variables were correlated and their means, variances, and covariances were set equal between male and female participants. All constrained models fitted the data well (for details, see Table S5 in the Supplemental Materials).

Results

Descriptive Statistics

Table 1 shows the means, standard deviations, and correlations of all variables. Both men and women, on average, desired more closeness than they actually experienced. As a consequence, the mean closeness discrepancy (i.e., the directional difference between actual and desired closeness) was negative for both men ( $M = - 0.49$ , $S D = 1.06$ ) and women ( $M = - 0.63$ , $S D = 1.13$ ), indicating that the average participant felt not close enough to their partner. However, as shown by the relatively large standard deviations, there was substantial variation between participants. Most participants (55.4%) reported no closeness discrepancies (i.e., identical ratings of actual and desired closeness); slightly fewer participants (39.5%) reported negative discrepancies (i.e., higher desired than actual closeness); and positive discrepancies (i.e., higher actual than desired closeness) were reported relatively rarely (5.1%). This distribution of closeness discrepancies is comparable to previous research (Frost & Forrester, 2013). Given the rare reports of positive closeness discrepancies in the current as well as previous studies, we refrain from drawing any inferences about the people reporting them; we report results for positive closeness discrepancies only for illustrative purposes.

Table 1.

Descriptive Statistics and Correlations.

Variables	M (SD)		Correlations
Variables	Men	Women	1	2	3	4	5	6	7	8	9	10
1. Actual closeness	5.47 (1.25)	5.24 (1.39)	.30***	.61***	.52***	.42***	−.57***	.44***	.47***	.49***	.65***	−.07
2. Desired closeness	5.96 (1.01)	5.87 (1.10)	.58***	.13**	.14***	.17***	−.18***	.15***	.18***	.19***	.23***	−.10**
3. Relationship satisfaction	8.07 (1.82)	8.18 (1.74)	.51***	.29***	.28***	.49***	−.64***	.51***	.50***	.49***	.80***	−.17***
4. Sexual satisfaction	6.81 (2.52)	7.24 (2.38)	.40***	.19***	.48***	.46***	−.51***	.34***	.33***	.40***	.67	−.22***
5. Loneliness	1.94 (1.06)	2.09 (1.13)	−.48***	−.21***	−.55***	−.54***	.37***	−.56***	−.48***	−.52***	−.82***	−.19***
6. Support	6.32 (0.96)	6.29 (1.10)	.45***	.30***	.50***	.38***	−.50***	.15**	.45***	.55***	.76***	−.17***
7. Commitment	4.04 (0.81)	3.98 (0.82)	.49***	.25***	.44***	.34***	−.36***	.39***	.27***	.39***	.70***	.03
8. Self-disclosure	3.76 (0.83)	3.95 (0.80)	.39***	.28***	.38***	.31***	−.33***	.42***	.28***	.24***	.74***	−.24***
9. SRQ	–0.02 (0.71)	0.02 (0.75)	.63***	.35***	.78***	.72***	−.76***	.74***	.65***	.64***	.39***	−.21***
10. Relationship length	7.55 (6.76)		−.05	−.03	−.17***	−.23***	.16***	−.13**	.15***	−.20***	−.17***

Note. $N$ = 1,177 individuals (from 748 couples). Correlations below the diagonal pertain to men, above the diagonal to women, and on the diagonal to within-couple correlations. Relationship length reports were averaged within couples. SRQ = subjective relationship quality. For 95% confidence intervals of the correlation coefficients, please see Table S4 in the Supplemental Materials.

$p$ < .01. *** $p$ < .001.

Hypothesis Testing

The dyadic polynomial regression model was used to predict whether SRQ fitted the data well, $χ^{2} = 25.610$ , $d f = 12$ , $p = . 012$ , $CFI = . 969$ , $RMSEA = . 039$ , $SRMR = . 019$ (Marsh et al., 2005). Result estimates are presented in Table 2 and illustrated in Figure 2. As can be seen on the left-hand side of the table, two polynomial actor effects reached statistical significance: Participants reported higher SRQ the higher they rated their actual closeness to their partners ( $b_{1}$ coefficient), but lower SRQ the more closeness they desired ( $b_{2}$ coefficient).

Table 2.

Results of the Dyadic Response Surface Analysis Predicting SRQ by Actor Participants’ and Their Partners’ Actual and Desired Closeness.

	Actor effects			Partner effects
Effects	Estimate ( $S E$ SE)	$p$ value	CI	Estimate ( $S E$ SE)	$p$ value	CI
Intercept	0.329 (0.045)	<.001	[0.241, 0.417]	0.329 (0.045)	<.001	[0.241, 0.417]
Actual closeness ( $b_{1}$ )	0.553 (0.053)	<.001	[0.450, 0.657]	0.196 (0.055)	<.001	[0.088, 0.304]
Desired closeness ( $b_{2}$ )	$- 0.137$ (0.052)	.008	[−0.239, −0.036]	−0.153 (0.054)	.005	[−0.260, −0.047]
Actual closeness² ( $b_{3}$ )	$- 0.045$ (0.031)	.138	[−0.105, 0.015]	−0.019 (0.027)	.474	[−0.072, 0.033]
Actual $\times$ Desired Closeness ( $b_{4}$ )	0.070 (0.054)	.195	[−0.036, 0.175]	0.001 (0.047)	.989	[−0.091, 0.092]
Desired closeness² ( $b_{5}$ )	$- 0.026$ (0.028)	.337	[−0.080, 0.028]	−0.057 (0.029)	.049	[−0.115, <0.001]
Response surface parameters
LOC_linear ( $a_{1}$ )	0.416 (0.032)	<.001	[0.353, 0.479]	0.043 (0.037)	.253	[−0.031, 0.116]
LOC_quadratic ( $a_{2}$ )	$- 0.002$ (0.032)	.947	[−0.066, 0.061]	−0.076 (0.026)	.003	[−0.126, −0.026]
LOIC_linear ( $a_{3}$ )	0.691 (0.099)	<.001	[0.496, 0.886]	0.350 (0.103)	.001	[0.148, 0.551]
LOIC_quardatic ( $a_{4}$ )	$- 0.141$ (0.095)	.135	[−0.327, 0.044]	−0.077 (0.086)	.369	[−0.246, 0.091]
$a_{5}$	$- 0.019$ (0.036)	.598	[−0.089, 0.051]	0.038 (0.036)	.284	[−0.032, 0.109]

Note. $N$ = 1,177 individuals (from 748 couples). The subscripts linear and quadratic refer to the linear and quadratic term coefficients, respectively, of the LOC and LOIC. The effect of the dyadic covariate relationship length (not displayed) was $b = - 0.182$ , $S E = 0.025$ , $p < . 001$ , CI = [ $- 0.232$ , $- 0.132$ ]. LOC = line of congruence; LOIC = line of incongruence; CI = confidence interval.

Figure 2.

Response surfaces for the prediction of subjective relationship quality (SRQ) by actual and desired closeness: (A) actor effects and (B) partner effects.

The response surface computed based on the polynomial actor effects (see Figure 2, Panel A) gave no indication of the hypothesized broad closeness discrepancy effect. Although there was a significantly positive additive main effect ( $a_{1}$ response surface parameter) and both the $a_{2}$ and $a_{5}$ response surface parameters did not reach significance, the $a_{3}$ response surface parameter was significantly positive, and the $a_{4}$ response surface parameter did not significantly differ from 0. Hence, the conditions for the hypothesized broad discrepancy effect were not met (Humberg, Nestler, & Back, 2018): There was no evidence that discrepancies between actual and desired closeness predict lower SRQ.

Beyond the intrapersonal actor associations, participants’ SRQ was significantly related to their partners’ ratings of actual and desired closeness closeness (see right-hand side of Table 2). Whereas actual closeness showed a significantly positive partner effect, the partner effect of desired closeness was negative. Again, however, there was no evidence for a broad discrepancy effect: Although the $a_{5}$ response surface parameter did not significantly differ from 0, there was no significantly positive additive main effect ( $a_{1}$ response surface parameter), both the $a_{2}$ and $a_{3}$ response surface parameters were significant, and the $a_{4}$ response surface parameter did not reach statistical significance.

Auxiliary Analysis: Linear Discrepancy Effects

In addition to estimating the full second-order polynomial regression model, we also tested whether a simpler linear APIM fits our data equally well. In this linear APIM, all higher order terms (quadratic effects and linear interaction effects) were constrained to zero so that solely the effects of participants’ actual and desired closeness on their own and their partners’ SRQ were freely estimated.

The linear APIM demonstrated good data fit ( $χ^{2} = 36.089$ , $d f = 18$ , $p = . 007$ , $CFI = . 958$ , $RMSEA = . 037$ , $SRMR = . 028$ ) that was not significantly worse than the fit of the full polynomial model, $χ^{2} = 10.765$ , $d f = 6$ , $p = . 096$ . The results of the linear APIM are presented in Table 3. Consistent with the results of our polynomial model, actual closeness showed a significantly positive actor effect on SRQ, whereas the effect of desired closeness was significantly negative. We additionally examined whether this effect pattern was consistent with a linear discrepancy effect, which we tested by following the condition-based regression approach developed by Humberg, Dufner, et al. (2018). Based on the estimates of the linear APIM, we first computed the parameter $a b s = | b_{1} - b_{2} | - | b_{1} + b_{2} |$ . This parameter was significantly positive, $b = 0.406$ , $p < . 001$ . Second, we subtracted $b_{2}$ from $b_{1}$ ; the result was significantly positive, $b = 0.841$ , $p < . 001$ .⁴ Together, these findings provide evidence for a positive linear discrepancy effect, indicating that actors report higher SRQ the less their actual closeness falls behind their desired closeness (i.e., the less negative and closer to zero their closeness discrepancies). Conversely, this means that participants reported lower SRQ the more negative their closeness discrepancies. Because of the small number of reported positive closeness discrepancies, the linear discrepancy effect cannot be meaningfully interpreted for the area of positive closeness discrepancies.

Table 3.

Results of the Linear APIM Predicting SRQ by Actor Participants’ and Their Partners’ Actual and Desired Closeness.

	Actor effects			Partner effects
Effects	Estimate ( $S E$ SE)	p value	CI	Estimate (SE)	p value	CI
Intercept	0.251 (0.036)	<.001	[0.181, 0.322]	0.251 (0.036)	<.001	[0.181, 0.322]
Actual closeness ( $b_{1}$ )	0.638 (0.038)	<.001	[0.564, 0.712]	0.212 (0.041)	<.001	[0.131, 0.294]
Desired closeness ( $b_{2}$ )	−0.203 (0.046)	<.001	[−0.293, −0.113]	−0.119 (0.046)	.009	[−0.208, −0.030]

Note. $N$ = 1,177 individuals (from 748 couples). The effect of the dyadic covariate relationship length (not displayed) was $b = - 0.184$ , $S E = 0.025$ , $p < . 001$ , CI = [ $- 0.233$ , $- 0.134$ ]. CI = confidence interval.

The partner effects estimated by the linear APIM were in the same direction as the actor effects, but smaller in magnitude. Again, we found evidence for a positive linear discrepancy effect ( $b (a b s) = 0.238$ , $p =. 004$ ; $b (b_{1} - b_{2}) = 0.331$ , $p < . 001$ ), indicating that participants’ negative closeness discrepancies were not only associated with their own, but also their partners’ SRQ. Thus, independent from their own closeness discrepancies, participants also reported lower SRQ the more negative closeness discrepancies were experienced by their partners.

The linear APIM is nested in the polynomial model used for the RSA as all parameters of the linear APIM (linear effects of actual and desired closeness) represent a subset of the parameters estimated in the polynomial model. Therefore, we additionally examined whether there is consistent evidence for positive linear discrepancy effects in the polynomial model. Regarding actor effects, the parameter $a b s$ amounted to 0.275 ( $p = . 004$ ) and the result of the difference $b_{1} - b_{2}$ was $0.691$ ( $p < . 001$ ); using the partner effects, an $a b s$ parameter of $0.307$ ( $p = . 002)$ was calculated, and $b_{1} - b_{2}$ resulted in $0.350$ ( $p = . 001$ ). Thus, there was converging evidence for positive linear discrepancy effects in the polynomial model.

Discussion

This study examined the relevance of closeness discrepancies for relationship quality. Based on previous research, we assumed that relationship functioning not only depends on how much closeness romantic partners actually experience in their relationships but also whether these experiences match the amount of closeness they desire. Previous findings suggest that actual-desired closeness discrepancies, indicated by either feeling not close enough (i.e., desiring more closeness than actually experienced) or too close (i.e., desiring less closeness than actually experienced) to one’s partner, impair relationship quality (Frost et al., 2017; Frost & Forrester, 2013; Gamarel & Golub, 2018; Mashek & Sherman, 2004). Individuals whose actual experiences of closeness matched their desired amount of closeness, in turn, were expected to report highest relationship quality. The goal of the current study was to examine whether these previous findings replicate when using RSA, an approach widely considered more appropriate to test discrepancy hypotheses (Barranti et al., 2017; Humberg, Nestler, & Back, 2018). Moreover, by running our analyses in a dyadic structural equation modeling framework, the current study was the first to examine whether SRQ is associated not only with individuals’ own but also with their partners’ closeness discrepancies.

Evidence for Linear But Not Broad Closeness Discrepancy Effects on Relationship Quality

Previous research on closeness discrepancies suggests that actual and desired closeness show broad discrepancy effects on SRQ. A broad closeness discrepancy effect means that both more positive and more negative closeness discrepancies are associated with lower SRQ. Conversely, people with matches between actual and desired closeness, especially at high levels of the two variables, should report highest SRQ. The dyadic RSA we employed in the current study allowed us to examine this hypothesized broad closeness discrepancy effect in a single analysis. In RSA, a broad discrepancy effect is indicated by a specific constellation of response surface parameters; importantly, these parameters must not be interpreted in isolation but in combination (Humberg, Nestler, & Back, 2018).

Unexpectedly, the constellation of response surface parameters estimated in the current study did not correspond to a broad discrepancy effect. Thus, our RSA failed to replicate the broad closeness discrepancy effects on relationship quality indicated by previous studies. Previous studies operationalized closeness discrepancies as difference scores (Frost et al., 2017; Frost & Forrester, 2013; Gamarel & Golub, 2018; Mashek & Sherman, 2004). Difference scores come with various methodological and interpretative problems, many of which RSA can overcome (e.g., Edwards & Parry, 1993; Humberg, Nestler, & Back, 2018; Schönbrodt et al., 2018). In our RSAs, the estimated broad discrepancy effect reflects, by method, the overall contribution of closeness discrepancies to SRQ, regardless of the direction of the discrepancy. The specific contributions of negative and positive closeness discrepancies cannot be differentiated from one another, but blend together in this broad discrepancy effect. Potential differences in these specific contributions, in addition to the limited data on positive closeness discrepancies, may therefore substantially reduce the chances of finding a significant broad closeness discrepancy effect on relationship quality. The current results therefore emphasize the importance of weighing alternative analytical approaches to the study of closeness discrepancies.

Although there was no evidence for a broad closeness discrepancy effect, a deeper inspection of the linear effect estimates of actual and desired closeness revealed a linear closeness discrepancy effect on relationship quality. Both in the polynomial regression model and in an equally well-fitting linear APIM that was fully nested in the polynomial regression model, actual closeness showed significantly positive effects and desired closeness showed significantly negative effects on relationship quality. Using Humberg and colleagues’ (2018) condition-based regression approach, we found that this effect pattern was consistent with a positive linear discrepancy effect. This positive linear discrepancy effect was meaningfully interpretable at least for negative closeness discrepancies, implying that individuals who experience less actual closeness than they desire (i.e., who report negative closeness discrepancies) are unhappier in their relationships. Individuals with less negative closeness discrepancies but more matching actual and desired closeness levels, in turn, seem to report higher SRQ. This finding corroborates the hypothesized relevance of negative closeness discrepancies for relationship functioning (Frost et al., 2017; Frost & Forrester, 2013). Feeling “not close enough” to one’s partner usually comes along with longing for the partner. Most people may miss their partners from time to time. However, when the partner is physically distant (e.g., if the couple has a long-distance relationship) or emotionally distant (e.g., if the partner shows low interest in shared activities) for an extended period of time, such feelings of longing may endure and affect one’s overall evaluation of the relationship. Indeed, negative closeness discrepancies may be particularly detrimental for personal well-being because they can, in most cases, not fully be regulated by the individual alone. Establishing more closeness is a couple-level task: One’s own closeness efforts (such as showing affection to the partner) need to be reciprocated by the partner for feelings of intimacy and connectedness to arise (Reis & Shaver, 1988).

The positive linear discrepancy effect is, from a statistical perspective, not limited to negative closeness discrepancies, but it could not be meaningfully interpreted for the region of positive closeness discrepancies due to the small number of participants who reported more actual closeness than they desired. The proportion of participants reporting positive closeness discrepancies was comparable to previous studies (e.g., Frost et al., 2017; Frost & Forrester, 2013), which corroborates that few people experience them. However, unlike previous studies, the current study found no evidence overall that positive closeness discrepancies are associated with relationship functioning. To clarify whether positive closeness discrepancies matter for relationship functioning, future studies with a larger number of people reporting positive closeness discrepancies are needed. One way to accomplish this could be selective sampling of individuals with certain traits likely associated with positive closeness discrepancies, such as avoidant attachment style (Hazan & Shaver, 1994).

Moreover, although relatively few people indicate positive closeness discrepancies in questionnaire reports, they may nonetheless experience them at some point in their daily lives. In a study by Mashek and Sherman (2004), the majority of participants admitted having felt too close to their partners sometime in the preceding 3 months. Closeness discrepancies are usually described as strong threats to one’s personal autonomy and identity that fuel feelings of dependence and confinement (Mashek et al., 2011). It seems likely that such feelings express in temporary drops of personal and relational well-being. However, it may be that most people are able to easily regulate positive closeness discrepancies with little individual effort (e.g., by distancing from the partner or pursuing individual activities) so that these positive closeness discrepancies do not lower their overall evaluation of the relationship. Extensive experience sampling studies asking couples to report their momentary actual and desired closeness and relationship functioning multiple times per day or week could address these important questions. Such studies may also examine whether the between-person effects of negative closeness discrepancies on relationship quality found in the present study apply to the within-person level.

Partner Effects of Closeness Discrepancies on Relationship Quality

Closeness is a central characteristic of interdependent dyads (Berscheid et al., 1989). Extensive research has shown that individuals’ relational well-being not only depends on their own characteristics and behaviors but also on their partners’ (Mund et al., 2016; Weidmann et al., 2016). Surprisingly, no previous study has examined the relevance of closeness discrepancies from a dyadic perspective. The current study was the first to investigate whether individuals’ reports of relationship quality are incrementally related to closeness discrepancies perceived by their partners.

Overall, the pattern of partner associations was similar to the pattern of actor associations. Although our RSA of the partner effects provided no evidence for a broad closeness discrepancy effect relationship quality, there were meaningful linear relationships: Participants’ actual closeness showed significantly positive effects, and their desired closeness showed significantly negative effects on their partners’ SRQ. These partner effects were smaller in magnitude than the actor effects, but likewise supported a positive linear discrepancy effect. Thus, it seems that individuals with more negative closeness discrepancies are not only unhappier with their relationships themselves but also have unhappier partners. This finding is particularly noteworthy because it provides first evidence for the interpersonal relevance of negative closeness discrepancies.

Future research is needed to better understand how interpersonal partner effects of closeness discrepancies manifest. More generally, “interpersonal [partner] effects need to be actuated through overt behavior” (Mund et al., 2016, p. 414). Thus, for the partner’s closeness discrepancies to take effect, they must express into overt behavior directed at the actor. Further studies are needed to identify the specific behaviors that closeness discrepancies express in. Currently, it is unknown whether or how people regulate closeness discrepancies. Examining the extent to which people regulate their closeness discrepancies internally and/or externally, and potential differences in regulation between negative and positive closeness discrepancies, presents an interesting task for future research. Due to the lack of theory, qualitative studies in which participants can freely describe how they experience and handle closeness discrepancies might be a first step to understanding the processes at play.

Limitations

Due of the cross-sectional data used in the present study, the findings do not allow for causal interpretations. The direction of the found associations between negative closeness discrepancies and SRQ remains unclear. For example, it might be possible that individuals experience greater matches between their desired and actual levels of closeness because they are more satisfied with the relationship in the first place. To enhance causal inference, future investigations may use experimental designs to examine the directional influence of closeness discrepancies on relationship quality or employ longitudinal assessments to examine the bidirectional transactions between negative closeness discrepancies and relationship variables over time.

Moreover, additional research is needed to examine whether the current findings generalize to couples from more diverse backgrounds. Participants in the current study were mostly middle-aged, in a cross-sex relationship, and highly educated. There is, however, research showing that couples with different socioeconomic status show different interaction patterns (Karney & Bradbury, 2020); whether socioeconomic status moderates the relevance of closeness for relationship functioning needs to be shown. There are other factors of interest, too. Whether couples share a household or live apart from each other, for instance, may considerably limit how close (or independent) partners can be. For instance, living apart likely makes pursuing individual goals and activities easier and thereby buffer potential negative consequences of positive closeness discrepancies. Living together, in turn, may greatly limit the possibilities of engaging in individual activities due to the higher objective closeness with the partner (Hagemeyer et al., 2015).

Finally, the current study focused on the relevance of closeness discrepancies for subjective evaluations of relationship quality. A compelling task for future research would be to test whether the current findings replicate with behavioral outcomes of closeness discrepancies (such as partner-directed behavior or physiological responses). Identifying specific intra- and interpersonal expressions of closeness discrepancies may contribute to elucidating the processes through which closeness discrepancies manifest in couples’ daily lives and thereby gain relevance for relationship functioning. In addition, future studies may also look at dyadic facets of relationship functioning such as conflict frequency, the frequency of face-to-face or digitally mediated contact, or the likelihood of future relationship dissolution. In a similar vein, it may also be worthwhile to focus not only on relational but also on personal well-being outcomes. Previous research, for instance, suggests that closeness discrepancies can contribute to mental health problems (Frost & Forrester, 2013). Additional outcomes of interest include, for instance, anger, stress, life satisfaction, or anxiety.

Conclusion

This study was the first to investigate whether previous findings on the implications of actual-desired closeness discrepancies for relationship functioning replicate when employing RSA, a method nowadays considered most appropriate to test discrepancy hypotheses. Unexpectedly, our RSA provided no evidence for broad closeness discrepancy effects. However, in combination, the linear effects of actual and desired closeness were consistent with a positive linear discrepancy effect that partially corroborated previous findings: People who lack closeness in their relationships seem to be least satisfied and happy, particularly when having a competing, strong desire for closeness at the same time. Participants whose desires for closeness considerably exceeded their actual experiences of closeness (i.e., negative closeness discrepancies) reported the lowest relationship quality. In contrast and disagreeing with previous research, there was no sufficient evidence that relationship quality is reduced by feeling too close to the partner (i.e., positive closeness discrepancies). Thus, what seems most relevant for relationship (dys)functioning is frustration due to lacking experiences of closeness paired with a strong desire for closeness.

Information on Preregistration

Data exclusion criteria (we reported all data exclusions in this article), as well as the dyadic RSA were preregistered; the auxiliary analysis (the linear APIM and the exploration of linear discrepancy effects) was not. Data were collected before the preregistration, and the study design and planned sample size were therefore not preregistered. Our preregistration can be found at https://osf.io/anvpj. We deviated from the preregistration in two respects. First, the current analyses used all variables measured in the study that are conceptually linked to SRQ. This allowed us to assess SRQ more broadly: in addition to the preregistered analysis of relationship satisfaction and commitment, we also analyzed sexual satisfaction, partner-related loneliness, perceived partner support, and self-disclosure (see the “Method” section for details). Our main analyses used an aggregate of the six SRQ variables. However, the results of supplemental analyses using the six variables as separate outcomes were largely consistent with the main findings. Second, to keep the manuscript succinct, we decided to not predict 1-year-changes in relationship outcomes.

Supplemental Material

sj-tex-1-psp-10.1177_01461672221113981 – Supplemental material for Closeness Discrepancies in Couple Relationships: A Dyadic Response Surface Analysis

Supplemental material, sj-tex-1-psp-10.1177_01461672221113981 for Closeness Discrepancies in Couple Relationships: A Dyadic Response Surface Analysis by Sebastian Pusch, Franz J. Neyer and Birk Hagemeyer in Personality and Social Psychology Bulletin

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by grants from the German Research Foundation to Birk Hagemeyer (HA 6884/2-1; HA 6884/2-2).

ORCID iD

Sebastian Pusch

Supplemental Material

Supplemental material is available online with this article.

Notes

References

Agnew

C. R.

Lange

P. A. M. V.

Rusbult

C. E.

Langston

C. A.

(1998). Cognitive interdependence: Commitment and the mental representation of close relationships. Journal of Personality and Social Psychology, 74(4), 939–954. https://doi.org/10.1037/0022-3514.74.4.939

Aron

E. N.

Smollan

(1992). Inclusion of other in the self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63(4), 596–612. https://doi.org/10.1037/0022-3514.63.4.596

Aron

Lewandowski

G. W. J.

Mashek

Aron

E. N.

(2013). The self-expansion model of motivation and cognition in close relationships. In Simpson

J. A.

Campbell

(Eds.), The Oxford handbook of close relationships (pp. 90–115). Oxford University Press.

Aron

Melinat

Aron

E. N.

Vallone

R. D.

Bator

R. J.

(1997). The experimental generation of interpersonal closeness: A procedure and some preliminary findings. Personality and Social Psychology Bulletin, 23(4), 363–377. https://doi.org/10.1177/0146167297234003

Arslan

R. C.

Walther

M. P.

Tata

C. S.

(2019). Formr: A study framework allowing for automated feedback generation and complex longitudinal experience-sampling studies using R. Behavior Research Methods, 52(1), 376–387. https://doi.org/10.3758/s13428-019-01236-y

Barranti

Carlson

E. N.

Côté

(2017). How to test questions about similarity in personality and social psychology research. Social Psychological and Personality Science, 8(4), 465–475. https://doi.org/10.1177/1948550617698204

Berscheid

Snyder

Omoto

A. M.

(1989). The relationship closeness inventory: Assessing the closeness of interpersonal relationships. Journal of Personality and Social Psychology, 57(5), 792–807. https://doi.org/10.1037/0022-3514.57.5.792

Bischof

(1975). A systems approach toward the functional connections of attachment and fear. Child Development, 46(4), 801–817. https://doi.org/10.2307/1128384

Edwards

J. R.

(2002). Alternatives to difference scores: Polynomial regression analysis and response surface methodology. In Drasgow

Schmitt

N. W.

(Eds.), Advances in measurement and data analysis (pp. 350–400). Jossey-Bass.

10.

Edwards

J. R.

(2007). Latent variable modeling in congruence research. Organizational Research Methods, 12(1), 34–62. https://doi.org/10.1177/1094428107308920

11.

Edwards

J. R.

Parry

M. E.

(1993). On the use of polynomial regression equations as an alternative to difference scores in organizational research. Academy of Management Journal, 36(6), 1577–1613. https://doi.org/10.2307/256822

12.

Fletcher

G. J. O.

Simpson

J. A.

(2000). Ideal standards in close relationships. Current Directions in Psychological Science, 9(3), 102–105. https://doi.org/10.1111/1467-8721.00070

13.

Fletcher

G. J. O.

Simpson

J. A.

Thomas

(2000). The measurement of perceived relationship quality components: A confirmatory factor analytic approach. Personality and Social Psychology Bulletin, 26(3), 340–354. https://doi.org/10.1177/0146167200265007

14.

Frost

D. M.

Forrester

(2013). Closeness discrepancies in romantic relationships: Implications for relational well-being, stability, and mental health. Personality and Social Psychology Bulletin, 39(4), 456–469. https://doi.org/10.1177/0146167213476896

15.

Frost

D. M.

McClelland

S. I.

Dettmann

(2017). Sexual closeness discrepancies: What they are and why they matter for sexual well-being in romantic relationships. Archives of Sexual Behavior, 46(8), 2353–2364. https://doi.org/10.1007/s10508-017-0960-2

16.

Galovan

A. M.

Holmes

E. K.

Proulx

C. M.

(2017). Theoretical and methodological issues in relationship research: Considering the common fate model. Journal of Social and Personal Relationships, 34(1), 44–68. https://doi.org/10.1177/0265407515621179

17.

Gamarel

K. E.

Golub

S. A.

(2018). Closeness discrepancies and intimacy interference: Motivations for HIV prevention behavior in primary romantic relationships. Personality and Social Psychology Bulletin, 45(2), 270–283. https://doi.org/10.1177/0146167218783196

18.

Hagemeyer

Neberich

Asendorpf

J. B.

Neyer

F. J.

(2013). (In)Congruence of implicit and explicit communal motives predicts the quality and stability of couple relationships. Journal of Personality, 81(4), 390–402. https://doi.org/10.1111/jopy.12016

19.

Hagemeyer

Neyer

F. J.

(2012). Assessing implicit motivational orientations in couple relationships: The Partner-Related Agency and Communion Test (PACT). Psychological Assessment, 24(1), 114–128. https://doi.org/10.1037/a0024822

20.

Hagemeyer

Neyer

F. J.

Neberich

Asendorpf

J. B.

(2013) The ABC of social desires: Affiliation, being alone, and closeness to partner. European Journal of Personality, 27. 442–457. https://doi.org/10.1002/per.1857

21.

Hagemeyer

Schönbrodt

F. D.

Neyer

F. J.

Neberich

Asendorpf

J. B.

(2015). When “together” means “too close”: Agency motives and relationship functioning in coresident and living-apart-together couples. Journal of Personality and Social Psychology, 109(5), 813–835. https://doi.org/10.1037/pspi0000031

22.

Hazan

Shaver

P. R.

(1994). Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22. https://doi.org/10.1207/s15327965pli0501_1

23.

Huinink

J. J.

Brüderl

Nauck

Walper

Castiglioni

Feldhaus

(2011). Panel analysis of intimate relationships and family dynamics (pairfam): Conceptual framework, and design. Zeitschrift für Familienforschung, 23, 77–101.

24.

Humberg

Dufner

Schönbrodt

F. D.

Geukes

Hutteman

van Zalk

M. H. W.

Denissen

J. J. A.

Nestler

Back

M. D.

(2018). Enhanced versus simply positive: A new condition-based regression analysis to disentangle effects of self-enhancement from effects of positivity of self-view. Journal of Personality and Social Psychology, 114(2), 303–322. https://doi.org/10.1037/pspp0000134

25.

Humberg

Nestler

Back

M. D.

(2018). Response surface analysis in personality and social psychology: Checklist and clarifications for the case of congruence hypotheses. Social Psychological and Personality Science, 10(3), 409–419. https://doi.org/10.1177/1948550618757600

26.

Karney

B. R.

Bradbury

T. N.

(2020). Research on marital satisfaction and stability in the 2010s: Challenging conventional wisdom. Journal of Marriage and Family, 82(1), 100–116. https://doi.org/10.1111/jomf.12635

27.

Keele

(2008). Semiparametric regression for the social sciences. John Wiley.

28.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. Guildford Press.

29.

Dove

N. L.

Agnew

C. R.

Korn

M. S.

Mutso

A. A.

(2010). Predicting nonmarital romantic relationship dissolution: A meta-analytic synthesis. Personal Relationships, 17(3), 377–390. https://doi.org/10.1111/j.1475-6811.2010.01285.x

30.

Marsh

H. W.

Hau

K.-T.

Grayson

(2005). Goodness of fit in structural equation models. In Maydeu-Olivares

McArdle

J. J.

(Eds.), Multivariate applications book series. contemporary psychometrics: A festschrift for Roderick P. McDonald (pp. 275–340). Lawrence Erlbaum.

31.

Mashek

D. J.

Israel

Aron

(2011). Wanting less closeness in romantic relationships. Basic and Applied Social Psychology, 33(4), 333–345. https://doi.org/10.1080/01973533.2011.614164

32.

Mashek

D. J.

Sherman

M. D.

(2004). Desiring less closeness with intimate others. In Mashek

D. J.

Aron

(Eds.), Handbook of closeness and intimacy (pp. 343–356). Lawrence Erlbaum.

33.

Mund

Finn

Hagemeyer

Neyer

F. J.

(2016). Understanding dynamic transactions between personality traits and partner relationships. Current Directions in Psychological Science, 25(6), 411–416. https://doi.org/10.1177/0963721416659458

34.

Mund

Johnson

M. D.

(2020). Lonely me, lonely you: Loneliness and the longitudinal course of relationship satisfaction. Journal of Happiness Studies, 22(2), 575–597. https://doi.org/10.1007/s10902-020-00241-9

35.

Nestler

Humberg

Schönbrodt

F. D.

(2019). Response surface analysis with multilevel data: Illustration for the case of congruence hypotheses. Psychological Methods, 24(3), 291–308. https://doi.org/10.1037/met0000199

36.

Pusch

Schönbrodt

F. D.

Zygar-Hoffmann

Hagemeyer

(2021). Perception of communal motives in couples: Accuracy, bias, and their associations with relationship length. Journal of Research in Personality, 91, Article 104060. https://doi.org/10.1016/j.jrp.2020.104060

37.

R Core Team. (2020). R: A language and environment for statistical computing. https://www.R-project.org/

38.

Reis

H. T.

Shaver

P. R.

(1988). Intimacy as an interpersonal process. In Duck

(Ed.), Handbook of personal relationships (pp. 367–389). John Wiley.

39.

Revelle

(2018). psych: Procedures for psychological, psychometric, and personality research. https://CRAN.R-project.org/package=psych

40.

Richard

F. D.

Bond

C. F.

Stokes-Zoota

J. J.

(2003). One hundred years of social psychology quantitatively described. Review of General Psychology, 7(4), 331–363. https://doi.org/10.1037/1089-2680.7.4.331

41.

Rosseel

(2012). Lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. https://doi.org/10.18637/jss.v048.i02

42.

Schönbrodt

F. D.

Humberg

(2018). RSA: An R package for response surface analysis (version 0.9.13). https://cran.r-project.org/package=RSA

43.

Schönbrodt

F. D.

Humberg

Nestler

(2018). Testing similarity effects with dyadic response surface analysis. European Journal of Personality, 32(6), 627–641. https://doi.org/10.1002/per.2169

44.

Weidmann

Ledermann

Grob

(2016). The interdependence of personality and satisfaction in couples. European Psychologist, 21(4), 284–295. https://doi.org/10.1027/1016-9040/a000261

45.

Wickham

(2016). ggplot2: Elegant graphics for data analysis. Springer. https://ggplot2.tidyverse.org

46.

Wickham

François

Henry

Müller

(2019). Dplyr: A grammar of data manipulation. https://CRAN.R-project.org/package=dplyr

47.

Wilke

C. O.

(2020). Cowplot: Streamlined plot theme and plot annotations for “ggplot2.” https://CRAN.R-project.org/package=cowplot

48.

Zygar

Hagemeyer

Pusch

Schönbrodt

F. D.

(2018). From motive dispositions to states to outcomes: An intensive experience sampling study on communal motivational dynamics in couples. European Journal of Personality, 32(3), 306–324. https://doi.org/10.1002/per.2145

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB