ArticlePDF Available

Abstract and Figures

Satisfaction with a romantic relationship often changes over time, and individuals differ in how satisfied they are in their relationship. However, no systematic review is available regarding the stability of individual differences in relationship satisfaction. Therefore, this meta-analysis synthesizes the available longitudinal data on rank-order stability of relationship satisfaction, as a function of age and relationship duration. Analyses were based on 148 samples including 153,396 participants reporting on their relationship over time. Mean age associated with the effect sizes ranged from 19 to 71 years, and mean relationship duration from 3 months to 46 years. On average, individual differences in relationship satisfaction were highly stable over time (r = .76, corrected for attenuation due to measurement error and based on an average time lag of 2.30 years). Rank-order stability varied systematically as a function of age, increasing from young to late adulthood with a slight decline during middle adulthood. Rank-order stability also varied as a function of relationship duration, increasing over the course of the relationship with a slight decline around 20 years of relationship duration. Moderator analyses suggested that relationship transitions shortly before Time 1 and sample type explained variance in rank-order stability. However, except for these two moderators, the pattern of findings was robust across all characteristics tested. In sum, this meta-analysis indicates that relationship satisfaction is a relatively stable construct, with lower stabilities in young adulthood and in the first years after beginning a relationship. This knowledge may stimulate future research on developmental processes within romantic relationships.
Content may be subject to copyright.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
1
Rank-Order Stability of Relationship Satisfaction:
A Meta-Analysis of Longitudinal Studies
Janina Larissa Bühler1,2 and Ulrich Orth1
1 Department of Psychology, University of Bern
2 Institute of Psychology, Heidelberg University
© 2022, American Psychological Association. This paper is not the copy of record and may
not exactly replicate the final, authoritative version of the article. Please do not copy or cite
without authors' permission. The final article will be available, upon publication, via its DOI:
10.1037/pspp0000430
Bühler, J. L. & Orth, U. (in press). Rank-order stability of relationship satisfaction: A meta-
analysis of longitudinal studies. Journal of Personality and Social Psychology.
http://dx.doi.org/10.1037/pspp0000430
Author Note
Janina Larissa Bühler https://orcid.org/0000-0003-3684-9682
Ulrich Orth https://orcid.org/0000-0002-4795-515X
Janina Larissa Bühler is now at the Department of Psychology, Johannes Gutenberg
University Mainz. This research was supported by Swiss National Science Foundation Grant
P2BSP1_188102 to Janina Larissa Bhler. Data, analysis script, and research materials are
available on the Open Science Framework (https://osf.io/n2z6b/).
Correspondence concerning this article should be addressed to Janina Larissa Bühler,
Department of Psychology, Johannes Gutenberg University Mainz, Binger Str. 14-16, 55122
Mainz, Germany. Email: jbuehler@uni-mainz.de.
The authors thank Samantha Krauss for her valuable help with coding.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
2
Abstract
Satisfaction with a romantic relationship often changes over time, and individuals differ in
how satisfied they are in their relationship. However, no systematic review is available
regarding the stability of individual differences in relationship satisfaction. Therefore, this
meta-analysis synthesizes the available longitudinal data on rank-order stability of
relationship satisfaction, as a function of age and relationship duration. Analyses were based
on 148 samples including 153,396 participants reporting on their relationship over time. Mean
age associated with the effect sizes ranged from 19 to 71 years, and mean relationship
duration from 3 months to 46 years. On average, individual differences in relationship
satisfaction were highly stable over time (r = .76, corrected for attenuation due to
measurement error and based on an average time lag of 2.30 years). Rank-order stability
varied systematically as a function of age, increasing from young to late adulthood with a
slight decline during middle adulthood. Rank-order stability also varied as a function of
relationship duration, increasing over the course of the relationship with a slight decline
around 20 years of relationship duration. Moderator analyses suggested that relationship
transitions shortly before Time 1 and sample type explained variance in rank-order stability.
However, except for these two moderators, the pattern of findings was robust across all
characteristics tested. In sum, this meta-analysis indicates that relationship satisfaction is a
relatively stable construct, with lower stabilities in young adulthood and in the first years after
beginning a relationship. This knowledge may stimulate future research on developmental
processes within romantic relationships.
Keywords: relationship satisfaction; rank-order stability; longitudinal studies; meta-
analysis
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
3
Rank-Order Stability of Relationship Satisfaction:
A Meta-Analysis of Longitudinal Studies
For most people, romantic relationships are among the closest and most important
relationships they experience in adulthood (Mund & Neyer, 2014), and satisfaction with the
romantic relationship is a predictor of health, longevity, and subjective well-being (Proulx et
al., 2007; Robles et al., 2014; Sbarra et al., 2011). As people go through life, their relationship
satisfaction does not remain constant but typically changes over time. However, although
mean-level change of relationship satisfaction has been meta-analyzed in previous research
(Bühler et al., 2021), rank-order stability of relationship satisfaction has not yet been
examined systematically. This is a critical limitation of the current state of knowledge because
information on both indices of stability and change is needed to gain a comprehensive
understanding of the development of relationship satisfaction across adulthood. Moreover,
knowledge about rank-order stability of relationship satisfaction may contribute to the
understanding of the nature of relationship satisfaction, by providing information about the
degree to which relationship satisfaction should be conceptualized as a trait-like construct.
Researchers have also debated about the most relevant time metric (i.e., age vs. relationship
duration) when studying the development of relationship satisfaction (Anderson et al., 2010).
Thus, it is essential to examine rank-order stability of relationship satisfaction both as a
function of age and as a function of relationship duration.
The goal of the present meta-analysis was to synthesize the available longitudinal data
on rank-order stability of relationship satisfaction within a given relationship to gain a robust
and precise picture of the stability of individual differences in relationship satisfaction.
Specifically, we sought to answer three questions: (a) What is the average rank-order stability
of relationship satisfaction? (b) Does rank-order stability of relationship satisfaction vary
across adulthood, as a function of age and as a function of relationship duration? (c) Does the
degree of stability differ across sample and methodological characteristics?
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
4
Rank-Order Stability of Relationship Satisfaction
Two central indices of stability must be considered when examining development
across adulthood: mean-level change and rank-order stability (Roberts & DelVecchio, 2000;
Roberts & Nickel, 2021; Roberts et al., 2006). When applied to the concept of relationship
satisfaction, indices of mean-level change reflect the average increase or decrease in
relationship satisfaction in a sample of individuals over time. If mean-level change is mapped
on a specific time metric (such as age or relationship duration), it is also referred to as index
of normative change. In contrast, rank-order stability of relationship satisfaction reflects the
stability of individual differences over time (i.e., high stability indicates that individuals in a
sample tend to keep the same rank on the construct over time). Thus, if rank-order stability is
high, then the relative position of individuals in the sample at a first assessment is a good
predictor of the relative position at a later assessment. Hence, the indices of mean-level
change and rank-order stability capture different aspects of developmental patterns in
psychological constructs (Caspi & Roberts, 1999; Robins, Fraley, et al., 2001). For example,
imagine three persons Heather, Tom, and Mary, of whom Heather has the highest and Mary
the lowest level of relationship satisfaction. If rank-order stability is high, then Heather will
still have the highest and Mary still the lowest level of relationship satisfaction at the next
assessmentirrespective of the mean level of their relationship satisfaction, which might
have changed for Heather, Tom, and Mary in a similar manner.
A recent meta-analysis examined the available data on mean-level change in
relationship satisfaction across adulthood, focusing on the role of age and relationship
duration (Bühler et al., 2021). The findings indicated that trajectories differed systematically
between the time metrics. Specifically, whereas the findings showed a U-shaped trend for age,
the pattern was more complex for relationship duration showing a decline in the first 10 years
of a relationship, followed by an increase over the next 10 years and again a decline after 20
years of relationship duration. However, as noted above, findings on mean-level change do
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
5
not provide any information about stability and change in the relative placements of
individuals (Block, 2014). For example, consider again a sample of individuals like Heather,
Tom, and Mary. All three of them may decrease in their relationship satisfaction over time
(i.e., the sample mean will decrease over time, indicating a lack of mean-level stability).
However, if Heather, Tom and Mary decrease by the same amount, their rank ordering will
remain exactly the same (indicating presence of rank-order stability). Moreover, even if the
individuals decrease by slightly different amounts (i.e., indicating individual differences in
change), it is possible that the rank order remains the same (indicating rank-order stability).
The example illustrates that change in mean levels is theoretically independent from change
in rank order.
More precisely, rank-order stability of relationship satisfaction concerns the question
to which degree relationship satisfaction should be conceptualized as a trait-like construct
(Fraley & Roberts, 2005). For example, intelligence is an individual-difference construct that
shows particularly high rank-order stability (Neisser et al., 1996). Also, the Big Five
personality traits and, to a somewhat lesser degree, self-esteem and life satisfaction are highly
stable constructs that are, consequently, considered personality traits (Lucas & Donnellan,
2007; Roberts & DelVecchio, 2000; Trzesniewski et al., 2003). A construct like mood, in
contrast, often changes quickly in response to the social environment (e.g., behavior of other
people) and in response to intrapersonal processes (e.g., expectancies), and is therefore
considered a state, not a trait. Based on a large meta-analytic dataset, Anusic and Schimmack
(2016) estimated the rank-order stability of several psychological constructs (i.e., personality
traits, self-esteem, life satisfaction, and affect), showing that rank-order stability decreases as
the time lag increases. At the same time, their findings on the longterm rank-order stability
(i.e., rank-order stability across long time intervals and corrected for measurement error)
showed that the stability coefficients asymptotically approached values of .83 for personality
traits, .56 for self-esteem, .52 for life satisfaction, and .42 for affect. Similar asymptotic
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
6
values have been reported in other studies on the longterm rank-order stability of these
constructs (Fraley & Roberts, 2005; Fujita & Diener, 2005; Kuster & Orth, 2013; Lucas &
Donnellan, 2007; Wagner et al., 2016). These findings imply that the rank-order stability of
psychological constructs is lower for longer time lags, but that rank-order stability does not
approach zero but nonzero values between 0 and 1, even across very long time lags (such as
several decades). The differing sizes of the asymptotic values (ranging from .42 to .83) also
suggest that psychological constructs differ in the degree to which they are trait-like.
There are two ways of how relationship satisfaction can be conceptualized (e.g.,
Fincham et al., 2018). On the one hand, relationship satisfaction is considered a construct
similar to life satisfaction. Empirical data suggest that relationship satisfaction is correlated
with life satisfaction at about medium size (.29 to .47; Be et al., 2013; Bühler et al., 2019).
Consequently, the rank-order stability of relationship satisfaction might be similar to that of
life satisfaction (Fujita & Diener, 2005; Lucas & Donnellan, 2007; Lykken & Tellegen, 1996;
Schimmack & Oishi, 2005). On the other hand, relationship satisfaction strongly correlates
with behavioral patterns in the relationship, such as communication styles. In fact, items of
both the Marital Adjustment Test (MAT; Locke & Wallace, 1959) and the Dyadic Adjustment
Scale (DAS; Spanier, 1976) correlate more strongly with communication factors than with
satisfaction factors (Funk & Rogge, 2007). Given that communication patterns are less stable
and often change over time (Johnson et al., 2021), rank-order stability of relationship
satisfaction might also be lower and closer to rank-order stability of state-like constructs.
Moreover, when assessing the rank-order stability of a psychological construct, such
as relationship satisfaction, it is essential to account for the time lag between assessments. As
noted above, theory and empirical findings clearly suggest that rank-order stability is often
large when the time lag is short (e.g., one year), but rank-order stability decreases as the time
lag increases (e.g., Ardelt, 2000; Fraley & Roberts, 2005; Kuster & Orth, 2013; Terracciano et
al., 2006). Specifically, as the time lag increases, rank-order stability typically levels off at
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
7
medium values, and this nonzero asymptote suggests that there is an enduring component of
individual differences in a given construct even across very long periods. Therefore, it is
crucial to account for the time lag between assessments when meta-analyzing rank-order
stability of relationship satisfaction.
Mechanisms Underlying the Rank-Order Stability of Relationship Satisfaction
Three types of mechanisms are relevant for explaining the rank-order stability of a
psychological construct: stochastic-contextual processes, person-environment transactions,
and developmental constancy factors (Fraley & Roberts, 2005). Below, we briefly review
these mechanisms and discuss how each of them applies to the rank-order stability of
relationship satisfaction. It should be noted that these mechanisms are not mutually exclusive
but jointly contribute to the rank-order stability of psychological constructs (Fraley &
Roberts, 2005).
First, all developmental processes are influenced by stochastic-contextual processes,
that is, by relatively random contextual factors, such as moving to a new place or meeting a
potential mate (Lewis, 1997, 1999, 2000a, 2000b). In fact, the statistical modeling by Fraley
and Roberts (2005) suggested that stochastic-contextual processes are needed to explain
individual differences in a psychological construct. More precisely, when the influence of
stochastic-contextual processes is ignored, individual differences in a construct would be
perfectly stable over time. Hence, stability and change in a psychological construct depend on
the stability of the context, and the degree of change and stability depends on how stable the
environmental conditions are.
Second, rank-order stability of psychological constructs also depends on person-
environment transactions, which means that individuals actively shape their environmental
conditions and that, simultaneously, these environmental conditions affect the individual
(Caspi & Bem, 1990; Caspi et al., 1989; Neyer & Asendorpf, 2001; Neyer et al., 2014). These
dynamic transactions between individuals and their environment (e.g., relationships) foster
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
8
consistency in individual differences. There are at least two transactive processes that are
relevant for explaining the rank-order stability of relationship satisfaction (Caspi & Bem,
1990). First, the way how individuals perceive their romantic relationship is influenced by
idiosyncratic social-cognitive biases (a transaction called reactive process). For example,
individuals who are satisfied with their relationship tend to perceive their partner and their
relationship through rose-colored glasses, which increases the likelihood of more positive
relationship experiences in the future. In contrast, individuals who are unhappy with their
relationship, show a negative bias in the perception of their relationship, which may cause
disappointment and relationship conflicts. In both cases, the person makes relationship
experiences that are congruent with their prior relationship satisfaction, which contributes to
the stability of individual differences in relationship satisfaction (e.g., Ickes et al., 1997;
Swann & Read, 1981). Second, individuals actively select themselves into environments,
including romantic relationships (a transaction called proactive process). For example,
individuals who believe that they are lovable and that others are trustworthy, tend to select
more supporting and trustworthy partners, which leads to more positive relationship
experiences (Erol & Orth, 2016; Sandra L. Murray et al., 2000). In contrast, individuals who
have more negative views of the self and others, will select untrustworthy partners, which will
cause more negative relationship experiences. In both cases, individuals tend to make
relationship experiences that are consistent with their pre-existing beliefs, which again
contributes to the stability of individual differences in relationship satisfaction. More
generally, the theoretical and statistical model by Fraley and Roberts (2005) suggested that
person-environment transactions amplify the degree of rank-order stability of psychological
constructs.
Third, rank-order stability is also influenced by developmental constancy factors (e.g.,
genetic predispositions and early formative experiences), which emphasize the role of latent
resiliency and vulnerability factors (Bowlby, 1973; McGue et al., 1993; Roberts & Caspi,
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
9
2003; Roberts & Wood, 2006). More precisely, developmental constancies may predispose
people to perceive themselves and their environments (e.g., their relationship and their
relationship partners) in a specific manner and to behave in ways that influence the quality of
their romantic relationships. For example, individuals who are high in neuroticism tend to
negatively interpret ambiguous relationship situations (Finn et al., 2013), which contributes to
lower relationship satisfaction. Such vulnerability and latent resiliency factors are a constant
influence on people’s perceptions and behavior in the relationship domain (McNulty, 2016),
which contributes to the stability of individual differences in relationship satisfaction. The
analyses by Fraley and Roberts (2005) suggested that developmental constancy factors are
needed to explain the typical pattern of rank-order stability of psychological constructs. In
fact, when constancy factors were omitted from their model, rank-order stability quickly
approached zero as the interval between assessments became longer.
Applied to romantic relationships, this reasoning suggests that the rank-order stability
of relationship satisfaction depends on the stability of all factors that influence the quality of
relationships, that is, (a) individual characteristics of the two partners, (b) characteristics of
the relationship, and (c) contextual factors outside of the relationship. For example, the
relationship science literature suggests that individual characteristics of the partners such as
emotional stability, conscientiousness, agreeableness, self-esteem, and attachment security
(i.e., low levels of both attachment-related anxiety and attachment-related avoidance)
significantly influence the quality of their relationship and, consequently, the partners’
satisfaction with their relationship (e.g., Erol & Orth, 2016; Li & Chan, 2012; McNulty, 2016;
Weidmann, Ledermann, et al., 2017). If individual differences in these factors are quite stable
over time (as suggested by empirical research; e.g., Fraley & Roberts, 2005; Kuster & Orth,
2013; Roberts & DelVecchio, 2000; Scharfe & Bartholomew, 1994), this suggests that
individual differences in relationship satisfaction will likewise be relatively stable. Similarly,
research suggests that individual characteristics, including emotional stability,
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
10
conscientiousness, agreeableness, self-esteem, and attachment security, contribute to
characteristics of the relationship (i.e., the pattern of relationship behavior that has evolved in
a specific relationship), such as patterns of communication, coping styles, responsiveness, and
conflict resolution (e.g., Campbell et al., 2005; S. L. Murray et al., 2000; Vater & Schröder-
Abé, 2015). Relationship patterns, in turn, influence people’s relationship satisfaction (e.g.,
Debrot et al., 2012; Gottman & Levenson, 1999; Vater & Schröder-Abé, 2015). Hence,
because individual differences in personality characteristics are relatively stable, patterns of
relationship behavior might also be relatively stable over time, which would further contribute
to rank-order stability of relationship satisfaction. In contrast, theory suggests that contextual
factors outside of the relationship may destabilize the relationship (Bodenmann, 1995; Hill,
1958; Randall & Bodenmann, 2009). For instance, the vulnerability-stress-adaptation model
(Karney & Bradbury, 1995; for a recent extension, see McNulty et al., 2021) emphasizes that
stressful life events (e.g., birth of a child) may impair adaptive processes within the
relationship (e.g., coping styles), which may compromise the relationship satisfaction of
couple members. Thus, samples that experienced potentially stressful changes in the
relationship context (e.g., samples of couples who had their first baby) might show lower
rank-order stability of relationship satisfaction than samples who did not experience
significant changes in the relationship context.
Does Rank-Order Stability of Relationship Satisfaction Vary Across Adulthood?
In addition to estimating the average rank-order stability of relationship satisfaction, it
is essential to understand how stability varies across adulthood. As noted above, in this meta-
analysis we examined rank-order stability of relationship satisfaction within a given
relationship as a function of age and relationship duration. Clearly, both time metrics are
strongly correlated: People of higher age, compared to people of younger age, are often in
relationships of longer duration, simply because they are older. Nevertheless, people separate
from their partner and begin a new romantic relationship across the entire period of adulthood
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
11
(Carr & Utz, 2020; Mehta et al., 2020). Moreover, while some of the mechanisms underlying
the rank-order stability of relationship satisfaction refer to age, others refer to relationship
duration. Therefore, it is important to separate the two time metrics, both conceptually and
empirically.
Rank-Order Stability of Relationship Satisfaction as a Function of Age
The developmental literature allows to derive hypotheses about how rank-order
stability of relationship satisfaction varies as a function of age. For instance, developmental
task theory posits that each life stage entails new developmental demands and societal
expectations (Erikson, 1968; Havighurst, 1972; see also Hutteman et al., 2014). Specifically, a
key developmental task in young adulthood (i.e., age 18 to 40 years) is to establish long-
lasting social relationships, including a committed romantic relationship (Ebner et al., 2006;
Heckhausen et al., 1989). At the same time, young adulthood is also characterized by
exploring different life paths and options, which sometimes implies leaving and entering
romantic relationships more readily (Arnett, 2000; Halpern-Meekin et al., 2013; Shulman &
Connolly, 2013).
In contrast, middle adulthood (i.e., age 40 to 65 years) involves the developmental
tasks of generativity and consolidation, expressed in caring for the next generation and
maintaining satisfactory social relationships, including marriage or a marriage-like
relationship (Erikson, 1968; McAdams, 2015). Moreover, in middle adulthood, individuals
usually develop an executive personality, which is, among other aspects defined by an
increase in mastery, competence, and control (Neugarten, 1968). This, in turn, increases the
capacity to handle multiple pressures and to cope successfully with difficult personal and
interpersonal experiences (Roberts & DelVecchio, 2000). Yet, middle adulthood is also a time
of potential crisis (Freund & Ritter, 2009; Levinson et al., 1976), resulting from the many
responsibilities in family, work, and community contexts, which may lead to stress, conflict,
and instability (Freund & Nikitin, 2012).
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
12
Finally, late adulthood (i.e., age 65 years and older) is often characterized by a greater
salience of loss-related issues (such as loss of beloved ones) and the perception of limited
remaining time (Carstensen et al., 1999; Charles & Carstensen, 2010). Therefore, key
developmental tasks in late adulthood involve the avoidance of, and adjustment to, losses and
the selective investment of time and energy into life domains (Ebner et al., 2006; Freund,
2008; Heckhausen et al., 1989; Ogilvie et al., 2001). As a result, older adults tend to focus
more strongly on present-oriented, rather than future-oriented, goals and invest more time and
energy in positive relationships with close others than in social interaction with acquaintances
(Carstensen et al., 1999; Frederick et al., 2017; Fung et al., 1999).
Taken together, the typical life situations and developmental tasks in young, middle,
and late adulthood suggest that individuals invest increasingly in establishing and maintaining
a romantic relationship as they go through life. This, in turn, should lead to more stability in
people’s relationship conditions. Given that romantic relationships differ substantially with
regard to their relationship quality (i.e., some will be fulfilling and satisfying for the partners,
whereas others will involve some level of conflict and be less satisfying), the developmental
trend towards consolidating and maintaining romantic relationships suggests that individual
differences in relationship satisfaction become more stable with age.
Rank-Order Stability of Relationship Satisfaction as a Function of Relationship Duration
Perspectives from relationship science allow to derive hypotheses on how rank-order
stability of relationship satisfaction varies as a function of relationship duration. Specifically,
the gradual disillusionment model (Huston et al., 2001; Huston & Houts, 1998) suggests that
baseline levels of relationship satisfaction decrease over the course of the relationship
(Diekmann & Mitter, 1984; Kurdek, 1998, 1999). The strongest decline often occurs over the
first 10 years of a relationship and, consequently, risk of separation peaks at around 10 years
after beginning a relationship (i.e., roughly corresponding to 7 years of marriage duration,
given that couples ususally have been together a few years before marrying; Kulu, 2014). This
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
13
implies that relationships of longer duration (10 years and more) predominantly represent the
“surviving” and more satisfied couples who have created a more stable relationship context
over time. As reviewed above, Fraley and Roberts’ (2005) model suggests that stable
environments contribute to higher rank-order stability of individual-difference constructs.
Therefore, we expected an increase in rank-order stability of relationship satisfaction over the
course of the relationship.
Moderators of Rank-Order Stability of Relationship Satisfaction
As noted above, individual factors and characteristics of the romantic relationship
might explain why rank-order stability of relationship satisfaction differs as a function of age
and relationship duration. Therefore, in this meta-analysis we also tested whether sample and
methodological characteristics are moderators of rank-order stability. For some of the
moderators, the literature allows to derive hypotheses about the significance and direction of
effects (i.e., living arrangement, marital status, presence of children, occurrence of
relationship transitions, and type of measure). For other moderators, however, no hypotheses
could be derived (e.g., ethnicity or gender). Nevertheless, to gain information about the
robustness and generalizability of the findings, we tested the full set of moderators.
Living Arrangements, Marital Status, and Presence of Children
People’s living arrangement, their marital status, and presence of children are sample
characteristics that may contribute to the explanation of rank-order stability of relationship
satisfaction. Specifically, couples who live in the same household, are married, and/or have
children might live in more stable relationship environments compared to couples who live in
separate households, are unmarried, and/or do not have children. Moreover, couples who live
in the same household, are married, and/or have children might encounter greater legal,
financial, and social barriers to separation, which may prevent relationship break-up even if
they are unsatisfied in their relationship (Rusbult, 1980, 1983). Together, these factors may
contribute to more stable individual differences in relationship satisfaction.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
14
Occurrence of Relationship Transitions
A sample characteristic that may contribute to lower rank-order stability of
relationship satisfaction is the occurrence of relationship transitions. Relationship transitions
generate new demands for individuals and couples, such as changes in family and work roles,
including new arrangements for household and work duties (Sanchez & Thomson, 1997).
Also, new parents often undergo stages of elevated stress and conflict (Doss et al., 2009).
These demands likely challenge the stable and consistent environment of the relationship
(Belsky & Rovine, 1990; Cast, 2004). Thus, the occurrence of relationship transitions may
destabilize the system of transactions between the person and the environment (Fraley &
Roberts, 2005).
Type of Measure
A methodological characteristic that may moderate the findings is the type of measure.
In general, measures of relationship satisfaction can be grouped into ad-hoc measures and
established measures. Established measures can be further classified into global satisfaction
measures, such as the Relationship Assessment Scale (RAS; Hendrick, 1988), and adjustment
measures, such as the DAS (Spanier, 1976). Global satisfaction measures rely on an
intrapersonal approach, reflecting people’s subjective evaluations of the relationship in
general. Adjustment measures, on the other hand, rely on an interpersonal or relationship
approach, reflecting typical patterns of interactions in the relationship, such as communication
and conflict styles (Fincham et al., 2018). Global satisfaction measures likely reflect the more
trait-like aspect of relationship satisfaction, while adjustment measures reflect the more state-
like aspect. Consequently, the use of global satisfaction measures should lead to greater
estimates of rank-order stability.
Sample Type, Ethnicity, Gender, and Baseline Mean of Relationship Satisfaction
We also tested for the moderating effects of sample type (i.e., nationally representative
vs. nonrepresentative), ethnicity, gender, and baseline mean of relationship satisfaction.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
15
Although we did not have directional hypotheses about the findings for these variables,
testing these variables provides important information about the robustness and
generalizability of the findings (see also Bühler et al., 2021; Orth et al., 2018). Specifically,
representative samples typically allow for more valid conclusions compared to
nonrepresentative samples, such as community samples and samples of college students (Orth
et al., 2018). Ethnicity and gender are key demographic characteristics that are of interest to
many researchers, so it is important to test whether meta-analytic findings differ between
ethnic groups and between women and men. For example, research suggests that dynamic
processes in romantic relationships vary by ethnicity (e.g., Orengo-Aguayo, 2015; see also
Karney & Bradbury, 2020). Finally, testing whether the meta-analytic findings hold across the
general level of relationship satisfaction in a sample (as indicated by the baseline mean) is
important, because some research suggests that baseline means in relationship satisfaction
may influence the dynamic processes in the relationship (e.g., Lavner et al., 2012). Thus,
testing these sample characteristics as moderators provides important information about the
generalizability of the findings.
The Present Research
The goal of this research was to synthesize the available longitudinal data on rank-
order stability of relationship satisfaction in adulthood. In the analyses, we will examine how
rank-order stability varies as a function of age and relationship duration. As noted above,
when estimating rank-order stability, it is essential to consider the time lag between
assessments. Therefore, we will conduct two sets of effect size analyses: without versus with
controlling for time lag (for similar procedures, see Roberts & DelVecchio, 2000;
Trzesniewski et al., 2003). The analyses that control for time lag will yield estimates of rank-
order stability as if all samples had the same time lag between assessments (by centering time
lag at the mean across effect sizes, i.e., 2.30 years). Finally, moderator analyses will provide
information about the robustness of the findings, by testing whether rank-order stability
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
16
differs across sample and methodological characteristics. It is important to note that some
potentially relevant moderators could not be examined in this meta-analysis, such as
personality variables, socioeconomic status, and sexual orientation (Chen & van Ours, 2018;
Conger et al., 2010; Karney & Bradbury, 1995). The reason is that (a) information on the
characteristics was not reported in most primary studies (i.e., personality variables), (b) the
information that was available was not comparable across most primary studies (i.e.,
socioeconomic status), or (c) the very low number of samples that provided information on
the characteristic would not have allowed for reliable conclusions (i.e., sexual orientation).
The present meta-analysis advances research on romantic relationships in several
ways. Although rank-order stability is a sample (or population) characteristic, knowledge
about how rank-order stability of relationship satisfaction changes as a function of age and
relationship duration, and about which individual and environmental factors moderate rank-
order stability, contributes to understanding the development of relationship satisfaction and,
more generally, relationships. For example, if the stability of individual differences is
particularly low in a specific developmental period (e.g., young adulthood), then this suggests
that the individual trajectories are more variable, and likely more malleable, compared to
developmental periods in which individual differences are very stable. Similarly, if the
stability of individual differences is relatively low in a specific relationship situation (e.g., in
the period after relationship transitions such as the transition to parenthood), then this
suggests that relationship interventions might be more impactful in this situation compared to
other situations in which rank-order stability is high. Moreover, this meta-analysis will allow
to evaluate whether rank-order stability of relationship satisfaction is of about similar size as
the rank-order stability of personality characteristics (such as the Big Five) or whether
individual differences in relationship satisfaction are less stable over time. If individual
differences in relationship satisfaction are as stable as individual differences in personality
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
17
characteristics, then this would suggest that relationship satisfaction within a given
relationship should be considered a trait-like construct.
Method
This meta-analysis used anonymized data and was therefore exempt from receiving
approval by the Ethics Committee of the authors’ institution (Faculty of Human Sciences,
University of Bern), in accordance with national law.
Transparency and Openness
We follow the Journal Article Reporting Standards (Appelbaum et al., 2018; Kazak,
2018) and describe how we obtained the samples included in the present meta-analysis. Data,
analysis script, and research materials (e.g., coding manual, information on study variables)
are available on the Open Science Framework (OSF, https://osf.io/n2z6b/). The design and
analyses of the present research were not pre-registered. Data were analyzed using R (R
Development Core Team, 2020), and the meta-analytic computations were conducted with the
metafor package (Viechtbauer, 2010).
The meta-analytic data set is based partially on data from another meta-analysis of
longitudinal studies on relationship satisfaction. Specifically, we used data on sample and
methodological characteristics that were also used in a meta-analysis on means and mean-
level change in relationship satisfaction (Bühler et al., 2021). For the present research, the
meta-analytic data set has been extended by including further studies that met the inclusion
criteria of the present meta-analysis and by coding the information required for meta-
analyzing rank-order stability of relationship satisfaction. With regard to effect size data and
analyses, there is no overlap between the present research and Bühler et al. (2021). For
reasons of completeness and clarity, we provide all relevant methodological information
below, even if information on some of the search and coding procedures is also reported in
Bühler et al. (2021).
Search and Selection Procedure
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
18
The flow diagram shown in Figure 1 summarizes the four steps of the search and
selection procedure in the present meta-analysis: identification, screening, eligibility, and
inclusion of studies.
Identification of Studies
We searched for English-language journal articles, books, book chapters, and
dissertations in the database PsycINFO to find relevant studies. The search was conducted on
September 9, 2019, and the following search terms were used: relationship satisfaction,
marital satisfaction, relationship quality, marital quality, dyadic adjustment, marital
adjustment, and marital relations. The search was restricted to empirical-quantitative and
longitudinal studies with non-clinical samples, by using the limitation options empirical
study, quantitative study, longitudinal study, and non-disordered population in PsycINFO.
The search yielded 1,207 potentially relevant studies, and two additional potentially relevant
studies were identified through other sources. Thus, the final data set consisted of 1,209
potentially relevant studies, including 53 dissertations.
1
1
Among the potentially relevant studies, the oldest study had been published in 1966. However, none of the
studies published before 2002 were included in the meta-analysis, for the following reasons. First, most of the
studies did not meet the inclusion criteria (e.g., because the studies were not longitudinal). Second, some studies
would have met the inclusion criteria, but the information on effect sizes provided in the study was insufficient
and although we contacted the authors of the studies, we did not receive the required information (e.g., because
the authors were no longer in academia). Third, some of the studies would have met the inclusion criteria but a
more recent study (i.e., published 2002 or later) using the same data was included because the more recent study
reported more complete information on sample characteristics and effect sizes.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
19
Figure 1
Flow Diagram of the Search and Selection Procedure
Note. The diagram has been adapted from Moher et al. (2009).
Screening
Studies identified through
database search (k= 1,207)
Identification
Studies assessed in full text
(k= 1,209)
Eligibility
Studies excluded (k= 1,123)
No romantic relationship (k= 373)
Not empirical-quantitative (k= 8)
No longitudinal design (k= 307)
Time lag < 2 months (k= 23)
Separation or widowhood (k= 2)
Clinical sample (k= 5)
Intervention study (k= 20)
No self-report measure (k= 143)
Measure not identical (k= 11)
Not sufficient information (k= 149)
Sample already included (k= 82)
Studies eligible
because all
required
information
provided in
study (k= 46)
Studies eligible
after receiving
additional
information from
authors via
e-mail
(k= 40)
Inclusion
Studies included in the meta-analysis
(k= 86)
Independent samples included in the meta-analysis
(k= 148)
Studies identified through other
sources (k= 2)
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
20
Screening and Eligibility of Studies
To decide on the eligibility of the studies, all studies were assessed in full text by the
first author or a second rater based on the inclusion criteria described below.
2
To decide
whether a study met the criteria for being included in the meta-analytic data set, the raters
followed standardized procedures. Studies were included if the following twelve criteria were
fulfilled: First, participants reported on a romantic relationship (i.e., measures of satisfaction
with non-romantic relationships, such as parent-child relationships, were not of interest).
Second, the study was empirical-quantitative. Third, the study used a longitudinal design (i.e.,
two or more assessments of the same sample). Fourth, the time lag between assessments was
2 months or more. More precisely, for the meta-analysis we used data from assessments that
were separated by at least 2 months; these assessments were coded as Time 1, Time 2, Time
3, etc. Fifth, the sample did not, as a whole, experience separation or widowhood. Sixth, the
sample was not clinical. Seventh, the study was not an intervention study. Eighth, relationship
satisfaction was assessed by self-report. Ninth, the measure of relationship satisfaction was
identical across assessments. Tenth, sufficient information was given to compute the effect
size. Eleventh, information on effect size data was consistent throughout the study. Twelfth,
the sample was not already included in the meta-analytic data set (specifically, when a sample
was used in more than one study, we selected the study that provided information on the
largest sample size or, if identical, the most comprehensive information on sample and effect
size data).
To obtain estimates on interrater agreement on eligibility, a random sample of 60
studies were rated by both raters, suggesting high interrater agreement on inclusion versus
exclusion of articles in the meta-analytic data set (i.e., 59 of 60 articles, resulting in κ = .92).
The diverging assessment was discussed until consensus was reached.
2
At the time of coding (i.e., October 2019 to February 2020), the qualifications of the raters were as follows:
The first author had a Ph.D. in Psychology and the second rater had a Masters degree in Psychology.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
21
Inclusion of Studies
Of the potentially relevant studies, 46 studies could be included immediately because
they provided sufficient information on the effect size (i.e., correlation coefficient between at
least two assessments) and information on mean age and/or mean relationship duration. In the
case of studies that met the criteria except for providing information on the effect size, mean
age, or mean relationship duration, we contacted the authors of the study with a request for
the missing information (if sufficient contact information was available in the study or could
be found elsewhere, e.g., on the website of the authors’ university). This procedure led to the
inclusion of 40 additional studies. In sum, the search procedures resulted in a total of 86
eligible studies, including 146 independent samples.
Coding Procedure
The 146 samples were coded by the first author or the second rater. The following data
were coded: year of publication, publication type, sample size, sample type, country,
ethnicity, sexual orientation, proportion of female participants, proportion of participants
living together with their partner in the same household, proportion of married participants,
proportion of participants with children, occurrence of a relationship transition (i.e., marriage
or birth of a child) between Time 1 and any of the following assessments (referred to as
transition), occurrence of a relationship transition (i.e., marriage or birth of a child) shortly
before Time 1 (referred to as post transition), type of relationship transition, time lag between
assessments, dyadic nature of sample, measure of relationship satisfaction, reliability
coefficient of relationship satisfaction averaged across assessments, range of scale (i.e., the
scale’s minimum possible score and the scale’s maximum possible score), mean of
relationship satisfaction at Time 1, mean age of participants at Time 1, mean relationship
duration at Time 1, year of Time 1 assessment, and correlation coefficient indicating the rank-
order stability of relationship satisfaction between assessments.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
22
In the case of dyadic female-male samples, we coded the data for female and male
participants separately to increase the power of moderator analyses testing for gender
differences in the findings. In all other cases of subsamples (e.g., young adults and middle-
aged adults), we coded the full sample.
3
If information on year of Time 1 assessment was not
reported, we estimated these data as follows: Year of Time 1 assessment = publication year
3 years time lag between the first and last measurement occasion of the study (based on the
assumption that studies are, on average, published 3 years after data collection has been
completed; for a similar procedure, see Orth et al., 2018). Moreover, for the effect size
analyses we needed participants’ mean age and mean relationship duration at the initial
assessment of each effect size interval (e.g., mean age and mean relationship duration at Time
4 for the effect size interval from Time 4 to Time 5). We used the information on mean age
and mean relationship duration at Time 1 and the information on time lags between
assessments to compute these values.
To obtain interrater agreement in this step of coding, a random sample of 40 studies
was rated by both coders, suggesting high interrater agreement, with κ = 1.00 for categorical
variables (except for one variable, see below) and r .99 for continuous variables. For sample
type, interrater agreement was κ = .90, resulting from one diverging assessment (one coding
was “community sample,” whereas the other was “college/university students”). All diverging
assessments were discussed until consensus was reached.
As described in Bühler et al. (2021), we used the following strategies to obtain data on
mean relationship duration if these data were missing in the study. First, we contacted the
authors of the study with a request for the information. This resulted in data on relationship
duration for 18 additional samples. Second, many studies provided information on proxies for
relationship duration, that is, duration of living together (15 samples) and/or marriage
3
The only exception was one study, in which we coded Israeli and German couples separately because the time
lag between assessments differed for the subsamples.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
23
duration (42 samples). On the basis of published data on how relationship duration, duration
of living together, and marriage duration are related, we used these proxies to obtain estimates
of relationship duration. More precisely, nationally representative data from Germany
(Schröder & Schmiedeberg, 2015; see also Heuveline & Timberlake, 2004) provide estimates
about the average time between beginning a relationship, moving in together, and marrying:
On average, people are in a relationship for 1.25 years before they move in together, and they
are in a relationship for 3.8 years before they marry. We used these estimates for creating an
overall relationship-duration variable. That is, if information on relationship duration was
missing, but information on duration of living together was available, we estimated
relationship duration by adding 1.25 years to the value of duration of living together, and if
information on marriage duration was available, we estimated relationship duration by adding
3.8 years to the value of marriage duration.
4
After using these procedures, information on
relationship duration was available for 106 samples.
Effect Size Measure
As effect size measure, we used the correlation between two assessments of
relationship satisfaction and included all available correlation coefficients that were reported
for the sample. However, if measures are not perfectly reliable, then the observed correlation
underestimates the true correlation. Because we were interested in estimates of the true rank-
order stability of relationship satisfaction, we corrected the correlations for attenuation
resulting from unreliability of the measures (Hunter & Schmidt, 1990, 2004, 2014). To obtain
the most accurate estimate of reliability of a given measure, we used the following strategies.
If available, we used the reliability coefficient of the measure as reported in the study (k =
4
As described, this procedure was based on estimates about the average time between beginning a relationship,
moving in together, and marrying. Therefore, we conducted sensitivity analyses by using 2 and 6 years (instead
of 1.25 and 3.8 years) as estimates of the average difference between relationship duration and duration of living
together, and relationship duration and marriage duration, respectively. The mean of the relationship-duration
variable used in the sensitivity analyses was 13.59 years (SD = 9.71, range = 0.2648.40). The results of the
sensitivity analyses are reported in Table S1 and showed that the pattern of findings was very similar to the
findings from the main analyses.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
24
134). If no reliability coefficient was provided in the study, we used one of the following two
methods: (a) When an established scale was employed (k = 4), we used the average reliability
coefficient of this measure as reported in other studies in the meta-analytic data set using the
same measure. (b) When a single item ad-hoc measure was employed (k = 10), we used .75 as
reliability estimate because the literature suggests that single item measures of constructs that
are highly schematizedi.e., constructs that can readily be reported by lay people, such as
relationship satisfactionoften have a reliability in the range of .70 to .80 (Lucas &
Donnellan, 2007; Robins, Hendin, et al., 2001; Woods & Hampson, 2005; for a similar
procedure, see Orth, 2018).
The disattenuated correlation coefficient is given by
𝑟′𝑖=𝑟𝑖
𝑟𝑒𝑙𝑖
,
where r’i is the disattenuated correlation coefficient in study i, ri is the observed
correlation coefficient in study i, and reli is the averaged reliability coefficient of the measure
in study i (Lipsey & Wilson, 2001). For the meta-analytic computations, the disattenuated
correlation coefficients were transformed to Fisher’s Zr values (Fisher, 1921; Hedges &
Olkin, 1985). After the meta-analytic computations, the effect size estimates were converted
back to the correlation metric.
Meta-Analytic Procedure
As noted above, for many samples the data set included more than one correlation
coefficient of relationship satisfaction, which yielded a multilevel data structure (i.e., effect
sizes nested in samples). To account for the multilevel structure in the meta-analytic
computations, we used the “rma.mv” function in the metafor package (Viechtbauer, 2010).
Following Lipsey and Wilson (2001), we used multilevel random-effects models to estimate
weighted mean effect sizes and multilevel mixed-effects models to test for moderators.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
25
When meta-analyzing correlation coefficients in the metric of Fisher’s Zr values, the
within-study variance is given by
𝑣𝑖= 1
𝑛𝑖 3 ,
where ni is the sample size of study i. However, when meta-analyzing correlation
coefficients that have been corrected for attenuation, the within-study variance is given by
𝑣′𝑖= 𝑣𝑖
𝑟𝑒𝑙𝑖
2 ,
where vi is the disattenuated within-study variance for study i, vi is the attenuated
within-study variance for study i (as given above), and reli is the averaged reliability
coefficient of the measure in study i (Lipsey & Wilson, 2001).
Results
Description of Studies
The meta-analytic data set included 148 samples, drawn from 86 studies and providing
402 effect sizes. In sum, the samples included 153,396 participants, and sample sizes ranged
from 32 to 84,711 (M = 1,050, SD = 7,001, Mdn = 179). An overview of the samples is given
in Table 1, and all effect sizes are reported at OSF (https://osf.io/n2z6b/).
26
Table 1
Descriptive Information for the Samples Included in the Meta-Analysis
Study
Sample
size
Mean age
T1
Mean
relationship
duration T1
Female
(in %)
Sample type
Country
Ethnicity
Measure
Reliability
Andres (2014)
153
34.00
12.00
100
Community
NLD
White
ENRICH
.80
Be et al. (2013), female
1,385
63.20
n.a.
100
National
GBR
n.a.
Ad-hoc
.80
Be et al. (2013), male
1,385
65.70
n.a.
0
National
GBR
n.a.
Ad-hoc
.71
Bikos et al. (2007)
32
38.63
n.a.
100
Community
TUR
White
KMSS
.93
Bloch et al. (2014), female
156
52.57
34.51
100
Community
USA
White
MAT/MRI
.85
Bloch et al. (2014), male
156
52.57
34.51
0
Community
USA
White
MAT/MRI
.80
Blumenstock and Papp (2017), female
373
24.30
4.30
100
Community
USA
Other
Ad-hoc
.88
Blumenstock and Papp (2017), male
373
26.50
4.30
0
Community
USA
Other
Ad-hoc
.85
Bodi et al. (2010)
389
35.04
10.87
100
Community
n.a.
White
RAS
.88
Bouchard (2014), female
151
28.00
6.25
100
Community
CAN
White
DAS
.91
Bouchard (2014), male
151
31.00
6.25
0
Community
CAN
White
DAS
.89
Bouchard et al. (2006), female
119
28.18
7.25
100
Community
CAN
n.a.
DAS
.78
Bouchard et al. (2006), male
119
30.40
7.25
0
Community
CAN
n.a.
DAS
.73
Bower et al. (2013), female
97
n.a.
7.81
100
Community
USA
White
DAS
.87
Bower et al. (2013), male
99
n.a.
7.81
0
Community
USA
White
DAS
.88
Brown et al. (2019)
88
54.98
n.a.
63
Community
USA
White
CSI
.84
Busby and Gardner (2008), female
275
26.29
4.25
100
Community
USA
White
REQ
.88
Busby and Gardner (2008), male
275
28.32
4.25
0
Community
USA
White
REQ
.88
Buyukcan-Tetik et al. (2017), female
195
29.97
6.51
100
Community
NLD
White
DAS
.87
Buyukcan-Tetik et al. (2017), male
195
32.91
6.51
0
Community
NLD
White
DAS
.87
Byers (2005)
87
37.70
12.20
62
Community
n.a.
White
GMRS
.96
27
Choi (2016)
2,078
31.35
7.45
100
Community
KOR
Asian
KMSS-R
.78
Christopher et al. (2015), female
96
29.34
3.52
100
Community
USA
White
MOQ
.92
Christopher et al. (2015), male
96
31.23
3.52
0
Community
USA
White
MOQ
.92
Crocker et al. (2017), Study 1
132
19.45
1.53
76
Community
USA
White
QMI
.96
DeMaris (2010)
704
35.63
16.86
62
National
USA
White
Ad-hoc
.87
Doohan et al. (2010), female
102
38.76
13.00
100
Community
USA
Other
MAT
.73
Doohan et al. (2010), male
102
41.20
13.00
0
Community
USA
Other
MAT
.74
Erol and Orth (2014), Study 2, female
6,115
40.30
15.90
100
National
USA
White
Ad-hoc
.75
Erol and Orth (2014), Study 2, male
6,115
43.00
15.90
0
National
USA
White
Ad-hoc
.75
Fallis et al. (2016), female
113
35.73
10.47
100
Community
CAN
White
QMI
.95
Fallis et al. (2016), male
113
37.96
10.47
0
Community
CAN
White
QMI
.95
Fincham and Beach (2007), female
84
41.10
n.a.
100
Community
USA
White
MAT
.90
Fincham and Beach (2007), male
84
43.30
n.a.
0
Community
USA
White
MAT
.90
Gao and Cummings (2019), female
237
37.82
16.80
100
Community
USA
White
MAT
.75
Gao and Cummings (2019), male
237
40.15
16.80
0
Community
USA
White
MAT
.75
Girme et al. (2014), Study 2, female
66
22.25
2.83
100
Community
NZL
n.a.
PRQC
.81
Girme et al. (2014), Study 2, male
66
22.25
2.83
0
Community
NZL
n.a.
PRQC
.78
Gray and Ozer (2019), female
325
29.60
n.a.
100
Community
USA
Other
CSI
.91
Gray and Ozer (2019), male
325
31.40
n.a.
0
Community
USA
Other
CSI
.91
Greving Mehall et al. (2009), female
157
30.59
9.37
100
Community
USA
White
MAT
.78
Greving Mehall et al. (2009), male
157
32.58
9.37
0
Community
USA
White
MAT
.72
Gustavson et al. (2016), female
238
46.00
n.a.
100
Community
NOR
White
RAS
.90
Gustavson et al. (2016), male
194
48.00
n.a.
0
Community
NOR
White
RAS
.90
Hagemeyer et al. (2013), female
547
39.40
11.40
100
Community
DEU
White
Ad-hoc
.75
Hagemeyer et al. (2013), male
547
41.60
11.40
0
Community
DEU
White
Ad-hoc
.75
Hakanen et al. (2011)
1,632
44.90
n.a.
72
Community
FIN
White
RSI
.96
Hammond and Overall (2014), female
88
21.08
2.58
100
Community
NZL
n.a.
PRQC
.88
28
Hammond and Overall (2014), male
88
22.73
2.58
0
Community
NZL
n.a.
PRQC
.88
Hernandez-Kane and Mahoney (2018)
67
29.85
5.05
67
Community
USA
White
KMSS
.97
Hsiao (2017)
614
42.83
n.a.
82
Community
TWN
Asian
Ad-hoc
.75
Impett et al. (2012), Partner 1
80
23.89
n.a.
n.a.
Community
CAN
Other
RSI
.90
Impett et al. (2012), Partner 2
76
23.79
n.a.
n.a.
Community
CAN
Other
RSI
.90
Ivanova (2016)
4,116
46.90
23.83
60
National
NLD
White
Ad-hoc
.95
Jayamaha and Overall (2015), Study 1
156
22.21
2.30
65
Student
NZL
n.a.
PRQC
.84
Jayamaha and Overall (2015), Study 2, female
174
22.43
3.00
100
Community
NZL
n.a.
PRQC
.85
Jayamaha and Overall (2015), Study 2, male
174
23.82
3.00
0
Community
NZL
n.a.
PRQC
.85
Jenkins et al. (2020), female
168
40.75
15.41
100
Community
USA
Black
CRDQ
.91
Jenkins et al. (2020), male
168
43.57
15.41
0
Community
USA
Black
CRDQ
.91
Jensen and Rauer (2015a), female
64
70.00
46.20
100
Community
USA
White
MSQO
.82
Jensen and Rauer (2015a), male
64
71.00
46.20
0
Community
USA
White
MSQO
.92
Jensen and Rauer (2015b)
67
20.80
3.03
100
Student
USA
White
IRQ
.94
Johnson and Anderson (2013), female
610
28.45
2.58
100
Community
USA
White
Ad-hoc
.90
Johnson and Anderson (2013), male
610
30.52
2.58
0
Community
USA
White
Ad-hoc
.90
Kanat-Maymon et al. (2016), Israeli, female
102
41.53
n.a.
100
Community
ISR
White
ENRICH
.79
Kanat-Maymon et al. (2016), Israeli, male
103
43.11
n.a.
0
Community
ISR
White
ENRICH
.79
Kanat-Maymon et al. (2016), German, female
209
39.16
n.a.
100
Community
DEU
White
ENRICH
.79
Kanat-Maymon et al. (2016), German, male
210
41.14
n.a.
0
Community
DEU
White
ENRICH
.79
Kerkhof et al. (2011), female
199
29.20
5.77
100
Community
NLD
White
DAS
.84
Kerkhof et al. (2011), male
199
32.07
5.77
0
Community
NLD
White
DAS
.86
Kluwer and Johnson (2007), female
262
28.80
6.46
100
Community
NLD
White
RSI
.89
Kluwer and Johnson (2007), male
262
31.20
6.46
0
Community
NLD
White
RSI
.89
Kouros (2011), female
296
37.84
14.38
100
Community
USA
White
MAT
.78
Kouros (2011), male
296
40.22
14.38
0
Community
USA
White
MAT
.78
Lavner and Bradbury (2010), female
232
25.50
4.13
100
Community
USA
Other
MAT
.78
29
Lavner and Bradbury (2010), male
232
27.00
4.13
0
Community
USA
Other
MAT
.78
LeBaron et al. (2014)
67
46.00
27.70
100
Community
USA
White
Ad-hoc
.87
Li et al. (2018), female
268
28.08
4.93
100
Community
CHN
Asian
QMI
.96
Li et al. (2018), male
268
29.59
4.93
0
Community
CHN
Asian
QMI
.94
Lickenbrock and Braungart-Rieker (2015), female
135
29.30
n.a.
100
Community
USA
White
MAT
.85
Lickenbrock and Braungart-Rieker (2015), male
135
30.79
n.a.
0
Community
USA
White
MAT
.85
Lin et al. (2017), female
141
39.84
n.a.
100
Community
TWN
Asian
KMSS
.96
Lin et al. (2017), male
141
42.06
n.a.
0
Community
TWN
Asian
KMSS
.96
Logan and Cobb (2013)
268
23.60
2.74
82
Student
CAN
Other
RAS
.87
Meltzer et al. (2014), Study 3, female
72
23.54
n.a.
100
Community
USA
White
QMI
.89
Meltzer et al. (2014), Study 3, male
72
24.92
n.a.
0
Community
USA
White
QMI
.89
Meltzer et al. (2013), female
169
23.40
n.a.
100
Community
USA
n.a.
QMI
.92
Meltzer et al. (2013), male
169
25.60
n.a.
0
Community
USA
n.a.
QMI
.92
Menéndez et al. (2011), female
108
27.12
n.a.
100
Community
ESP
n.a.
Ad-hoc
.81
Menéndez et al. (2011), male
79
28.73
n.a.
0
Community
ESP
n.a.
Ad-hoc
.81
Miller et al. (2003), female
168
21.00
n.a.
100
Community
USA
White
MOQ
.91
Miller et al. (2003), male
168
24.00
n.a.
0
Community
USA
White
MOQ
.91
Moen (2012), female
306
22.59
n.a.
100
Community
USA
White
KMSS
.86
Moen (2012), male
306
24.50
n.a.
0
Community
USA
White
KMSS
.87
Mund et al. (2015), Study 1
186
26.82
6.31
66
Community
DEU
White
RAS
.88
Mund et al. (2015), Study 2, female
2,124
31.35
9.38
100
National
DEU
White
RAS
.75
Mund et al. (2015), Study 2, male
2,124
34.16
9.38
0
National
DEU
White
RAS
.75
Naud et al. (2013), female
299
28.00
n.a.
100
Community
CAN
n.a.
DAS
.86
Naud et al. (2013), male
299
30.00
n.a.
0
Community
CAN
n.a.
DAS
.81
Nguyen et al. (2017), female
414
26.30
4.20
100
Community
USA
Hispanic
Ad-hoc
.70
Nguyen et al. (2017), male
414
27.90
4.20
0
Community
USA
Hispanic
Ad-hoc
.70
Niessen et al. (2018)
133
42.75
n.a.
31
Community
DEU
White
QMI
.97
30
Ogolsky et al. (2016), female
193
23.26
2.19
100
Community
USA
Other
MOQ
.75
Ogolsky et al. (2016), male
183
24.80
2.19
0
Community
USA
Other
MOQ
.75
Orth et al. (2015)
2,509
47.60
n.a.
40
Community
DEU
White
SRS
.93
Orth et al. (2012)
1,448
49.83
n.a.
57
Community
USA
White
RSS
.88
Padilla et al. (2018), female
246
39.38
19.12
100
Community
USA
Hispanic
CRDQ
.94
Padilla et al. (2018), male
246
41.94
19.12
0
Community
USA
Hispanic
CRDQ
.94
Paleari et al. (2005), female
124
43.80
22.60
100
Community
ITA
White
QMI
.96
Paleari et al. (2005), male
119
46.20
22.60
0
Community
ITA
White
QMI
.95
Parfitt et al. (2014), female
75
33.04
6.08
100
Community
GBR
White
DAS
.92
Parfitt et al. (2014), male
66
34.08
6.33
0
Community
GBR
White
DAS
.92
Parise et al. (2017), female
139
29.20
5.50
100
Community
ITA
White
QMI
.89
Parise et al. (2017), male
139
31.00
5.50
0
Community
ITA
White
QMI
.90
Peltz et al. (2018), female
249
35.00
10.50
100
Community
USA
White
CSI
.94
Peltz et al. (2018), male
249
36.00
10.50
0
Community
USA
White
CSI
.94
Reizer et al. (2014), Study 3, female
44
29.17
8.16
100
Community
ISR
n.a.
MAT
.84
Reizer et al. (2014), Study 3, male
44
30.34
8.16
0
Community
ISR
n.a.
MAT
.86
Roberson et al. (2015)
779
19.12
n.a.
81
Community
USA
White
CSI
.90
Robins et al. (2002)
214
21.00
2.08
63
National
NZL
White
Ad-hoc
.92
Ruffieux et al. (2014), female
162
40.40
14.60
100
Community
CHE
White
PFB
.92
Ruffieux et al. (2014), male
162
42.60
14.60
0
Community
CHE
White
PFB
.94
Sadikaj et al. (2015), female
93
27.89
4.22
100
Community
CAN
White
DAS
.93
Sadikaj et al. (2015), male
93
30.28
4.22
0
Community
CAN
White
DAS
.88
Schober (2012), female
5,624
33.55
7.94
100
Community
GBR
White
Ad-hoc
.80
Schober (2012), male
5,624
35.98
7.94
0
Community
GBR
White
Ad-hoc
.80
Sotskova et al. (2015), female
98
29.98
n.a.
100
Community
CAN
White
DAS
.91
Sotskova et al. (2015), male
98
32.03
n.a.
0
Community
CAN
White
DAS
.94
South et al. (2020), target
730
60.00
30.8
39
Community
USA
White
DAS
.82
31
South et al. (2020), spouse
551
60.00
30.8
61
Community
USA
White
DAS
.84
Sullivan et al. (2017)
86
22.50
n.a.
64
Community
USA
Other
RAS
.85
Sun et al. (2017), female
164
40.53
16.52
100
Community
USA
Black
CRDQ
.90
Sun et al. (2017), male
164
43.11
16.52
0
Community
USA
Black
CRDQ
.90
Szepsenwol et al. (2015), female
62
24.53
0.26
100
Community
ISR
n.a.
RAS
.82
Szepsenwol et al. (2015), male
62
25.87
0.26
0
Community
ISR
n.a.
RAS
.82
Tombeau Cost et al. (2018)
222
31.77
7.32
100
Community
CAN
n.a.
QMI
.81
Tremblay and Pierce (2011)
160
30.00
4.00
0
Community
CAN
White
DAS
.77
van den Troost et al. (2005), Study 3, female
386
45.00
25.80
100
National
NLD
White
Ad-hoc
.80
van den Troost et al. (2005), Study 3, male
386
47.50
25.80
0
National
NLD
White
Ad-hoc
.80
van Scheppingen et al. (2018)
84,711
29.59
n.a.
100
National
NOR
White
Ad-hoc
.88
Volling et al. (2015), female
241
31.60
9.57
100
Community
USA
White
IRQ
.76
Volling et al. (2015), male
241
33.20
9.57
0
Community
USA
White
IRQ
.75
Weidmann, Schönbrodt, et al. (2017), female
237
48.40
23.50
100
Community
CHE
White
RAS
.91
Weidmann, Schönbrodt, et al. (2017), male
237
50.70
23.50
0
Community
CHE
White
RAS
.91
Whiteman et al. (2007), female
188
36.74
17.36
100
Community
USA
White
CRDQ
.87
Whiteman et al. (2007), male
188
38.92
17.36
0
Community
USA
White
CRDQ
.87
Whitton et al. (2014)
748
25.71
3.14
65
Community
USA
White
DAS
.85
Zee and Weiss (2019)
678
46.41
20.70
50
National
USA
n.a.
Ad-hoc
.75
32
Note. Mean age and mean relationship duration are given in years. The column “Female” shows the percentage of female participants. T1 = Time 1. Reliability indicates
the reliability estimate of the relationship-satisfaction measure. “n.a.” indicates that data were not available. National = nationally representative. Country follows the
ISO31661 alpha3 codes: CAN = Canada; CHE = Switzerland; CHN = China; DEU = Germany; ESP = Spain; FIN = Finland; GBR = United Kingdom of Great
Britain and Northern Ireland; ITA = Italy; ISR = Israel; KOR = South Korea; NLD = Netherlands; NOR = Norway; NZL = New Zealand; USA = United States of
America; TUR = Turkey; TWN = Taiwan. Measures were as follows (including adaptations, subscales, and translations thereof): CRDQ = Couple Relationship Domains
Questionnaire (Huston et al., 1986); CSI = Couples Satisfaction Index (Funk & Rogge, 2007); DAS = Dyadic Adjustment Scale (Spanier, 1976); ENRICH = ENRICH
Marital Satisfaction (Fowers & Olson, 1993); GMRS = Global Measure of Relationship Satisfaction (Lawrance & Byers, 1998); IRQ = Intimate Relations Questionnaire
(Braiker & Kelley, 1979); KMSS = Kansas Marital Satisfaction Scale (Schumm et al., 1986); KMSS-R = Kansas Marital Satisfaction Scale Revised (Chung, 2004);
MAT = Marital Adjustment Test (Locke & Wallace, 1959); MOQ = Marital Opinion Questionnaire (Huston & Vangelisti 1991); MSQO = Marital Satisfaction
Questionnaire for Older Persons (Haynes et al., 1992); MRI = Marital Relationship Inventory (Burgess et al., 1971); PFB = Partnership Questionnaire
(Partnerschaftsfragebogen; Hahlweg, 1996); PRQC = Perceived Relationship Quality Components Inventory (Fletcher et al., 2000); QMI = Quality of Marriage Index
(Norton, 1983); RAS = Relationship Assessment Scale (Hendrick, 1988); REQ = Relationship Evaluation Questionnaire (Busby et al., 2001); RSI = Relationship
Satisfaction subscale from the Investment Model Scale (Rusbult et al., 1998); RSS = Relationship Satisfaction Scale (Gilford & Bengtson, 1979); SRS = Self-Report
Relationship Satisfaction (Schmitt et al., 1997); ad-hoc = measure constructed for the study (without a name).
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
33
Sample Characteristics
The publication year of the studies ranged from 2002 to 2019 (Mdn = 2014). Nine
percent were nationally representative samples, 89% were community samples, and 2% were
college samples. Almost all of the samples came from Western countries (96%), including the
United States (47%), Canada (12%), the Netherlands (7%), Germany (6%), New Zealand
(5%), Israel (4%), the United Kingdom (4%), Italy (3%), Switzerland (3%), Norway (2%),
Spain (1%), Finland (1%), and Turkey (1%). Only 4% were from Asian countries (i.e., China,
South Korea, and Taiwan). Data from African and South American countries were not
available. Of the samples, 78% were predominantly White (with “predominantly” defined as
at least 70% of the sample), 5% predominantly Asian, 3% predominantly Black, 3%
predominantly Hispanic/Latin America, and 11% belonged to another ethnicity or were
ethnically diverse. The mean proportion of female participants was 54% (range = 0% to
100%, SD = 47%, Mdn = 66%).
Participants’ mean age at Time 1 was 34.78 years (SD = 10.36), ranging from 19.20
years to 71.00 years, and their mean relationship duration at Time 1 was 11.06 years (SD =
9.28), ranging from 3 months to 46.20 years. Participants’ mean year of birth was 1969 (SD =
13.37), ranging from 1936 to 1995. Mean year of Time 1 assessment was 2004 (SD = 7.76),
ranging from 1980 to 2014. The mean proportion of participants who lived in the same
household as their partner was 82% (range = 11% to 100%, SD = 28%, Mdn = 100%), the
mean proportion of participants who were married was 79% (range = 0% to 100%, SD = 33%,
Mdn = 100%), and the mean proportion of participants who had children was 62% (range =
0% to 100%, SD = 44%, Mdn = 96%). As regards the occurrence of relationship transitions,
16% of the samples married or had a baby between Time 1 and any of the following
assessments, and another 16% of the samples married or had a baby shortly before Time 1
(i.e., on average, 6.72 months before Time 1).
Methodological Characteristics
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
34
The mean time lag between assessments was 2.30 years (SD = 3.09), ranging from 2
months (which was the minimal lag required to be included in this meta-analysis) to 19.83
years. In 83% of the samples, relationship satisfaction was assessed with an established
measure and in 17% with an ad-hoc measure. Of the studies with established measures, 46%
used a global satisfaction measure and 54% an adjustment measure. A detailed list of all
measures and their categorization into global satisfaction measures and adjustment measures
is reported in Table S2.
Preliminary Analyses
Testing for Influential Outliers
We used the “influence” command in the metafor package (Viechtbauer, 2010) to test
for influential outliers. In metafor, this command is not available for the “rma.mv” function,
so we used the “rma” function, which does not account for the multilevel data structure. The
results indicated that no sample qualified as potential outlier. Consequently, we used the full
data set in the remainder of the analyses.
Testing for Publication Bias
We did not expect publication bias to be a problem in this meta-analysis given that
most of the studies included did not focus on rank-order stability of relationship satisfaction
per se (i.e., most studies focused on other research questions but the statistics that were central
for this meta-analysis were reported in the studies).
For assessing publication bias, we used three methods. First, we examined the funnel
plot, which shows the relation between effect size and standard error and serves as a graphical
device to detect publication bias (Light & Pillemer, 1984; Rothstein et al., 2005; Sterne &
Egger, 2001; Sutton, 2009). The funnel plot had an asymmetric shape (Figure 2). Second, we
used Egger’s regression test, which statistically tests for asymmetry of the funnel plot (Egger
et al., 1997). In metafor, this test is not available for the “rma.mv” function, so we used the
“rma” function. The test was not significant, z = 0.755, p = .450. Third, we compared effect
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
35
Fisher's z Transformed Correlation Coefficient
Standard Error
0.2 0.133 0.067 0
0 0.5 1 1.5 2 2.5 3
sizes that were published in the studies (k = 163) with effect sizes that were not published in
the studies (but obtained from the authors upon request; k = 239), using a multilevel mixed-
effects meta-regression model. The results indicated that effect sizes did not significantly
differ from each other, QModel = 0.011, df = 1, p = .916. Overall, the findings from all three
methods suggest that there was no evidence for systematic publication bias. In particular,
comparing published versus unpublished effect sizeswhich might be the most direct test of
publication biasindicated no significant differences.
Figure 2
Funnel Plot Displaying the Relation Between the Effect Size (Fisher’s Z-Transformed
Disattenuated Correlation Coefficient) and the Standard Error of the Effect Size
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Fisher’s Z-transformed disattenuated correlation
coefficient
0.20 0.13 0.07 0.00
Standard error
36
Note. For effect size, the figure shows disattenuated correlation coefficients, indicating the rank-order stability of relationship satisfaction between two
assessments. The size of the points reflects the weights of each sample.
Mean relationship duration (years)
Mean age (years)
Effect size
Effect size
Effect size
20 30 40 50 60 70
Mean age (years)
Effect size
0.2 0.4 0.6 0.8 1.0
010 20 30 40
relationship duration (years)
Effect size
0.2 0.4 0.6 0.8 1.0
Time lag (years)
A Age
B Relationship duration
0 5 10 15 20
Time lag (years)
Effect size
0.2 0.4 0.6 0.8 1.0
0.2 0.4 0.6 0.8 1.0
0.2 0.4 0.6 0.8 1.0
0.2 0.4 0.6 0.8 1.0
20 30 40 50 60 70
0 10 20 30 40
0 5 10 15 20
C Time lag
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
37
Effect Size Analyses
Across samples, the weighted mean effect size was r = .766 with 95% CI [.732, .797].
When controlling for the effect of time lag, the weighted mean effect size was r = .763 with
95% CI [.728, .794]. Below, we report the effect size analyses for rank-order stability as a
function of age and relationship duration.
Rank-Order Stability as a Function of Age
Visual Overview. The scatterplot shown in Figure 3A provides an overview of the
relation between mean age and effect size. As the figure indicates, variability of effect sizes
was larger in young and middle adulthood compared to late adulthood. In addition, the
scatterplot shows that the effect sizes became larger with age.
Moreover, we used the locally estimated scatterplot smoothing (LOESS) curve
(Cleveland, 1978, 1981; Cleveland & Devlin, 1988) to gain further information about the
developmental pattern of rank-order stability as a function of age. Figure 4A shows the
LOESS curve with mean age as predictor and the disattenuated correlation coefficient as
outcome. The figure supports the increasing trend of rank-order stability as a function of age,
with a slight decline in rank-order stability between age 40 and 50 years.
38
Note. The figure shows disattenuated correlation coefficients, indicating the rank-order stability of relationship satisfaction between two assessments,
estimated with locally estimated scatterplot smoothing (LOESS) curves.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
39
Weighted Mean Effect Sizes in Age Groups. For the meta-analytic computations, we
constructed four age groups. For age 19 (19.20 years was the minimum age in the meta-
analytic data set) to 50 years, we constructed age groups with 10-year intervals (except for the
first group, which included 2 samples between 19 and 20 years): 1930 years, 3040 years,
and 4050 years. For age above 50 years, the number of samples was substantially lower.
Therefore, we constructed one age group from 5071 years (71 years was the maximum age
in the meta-analytic data set).
Table S3 reports descriptive information on time lag in the four age groups of the
effect size analyses (see values in the upper half of the table). As the table indicates, the mean
time lag ranged from 1.88 to 4.24 years across age groups, with the largest mean time lag in
the age group 4050 years.
We used the weighted mean effect size as best estimate of rank-order stability of
relationship satisfaction in each age group. Table 2 reports the meta-analytic estimates, both
without control of time lag and with control of time lag. In the analyses that controlled for
time lag, time lag was centered at the mean of the meta-analytic data set (i.e., 2.30 years).
Thus, with control of time lag, differences between age groups in time lag are statistically
controlled and, moreover, the meta-analytic estimates of rank-order stability refer to a time
lag of 2.30 years. The findings from both analyses indicated that effect sizes were larger with
higher age, except for the age group 4050 years, which showed a slightly lower effect size
compared to the age group 3040 years. Overall, however, the findings suggest that rank-
order stability of relationship satisfaction increased as a function of age, and this increase was
particularly pronounced in young adulthood.
40
Table 2
Estimates of Rank-Order Stability of Relationship Satisfaction as a Function of Age and Relationship Duration
Group
k
N
Weighted mean
effect size
95% CI
Q
Variances
σ12
σ22
Not controlled for time lag
Age
1930 years
157
96,232
.672
[.613, .724]
23737.5
.065
.128
3040 years
128
25,901
.804
[.751, .846]
8827.2
.185
.081
4050 years
66
28,251
.768
[.696, .825]
5139.7
.186
.064
5071 years
45
4,816
.921
[.865, .954]
6290.4
.134
.177
Relationship duration
05 years
79
7,623
.665
[ .597, .723]
1472.7
.070
.048
510 years
99
21,400
.802
[ .733, .854]
7326.9
.151
.163
1020 years
79
18,507
.811
[ .736, .866]
6884.1
.249
.073
2046 years
55
8,071
.856
[ .757, .917]
16828.9
.296
.176
Controlled for time lag
Age
2030 years
157
96,232
.656
[ .596, .708]
9659.4
.061
.115
3040 years
128
25,901
.795
[ .741, .838]
7937.4
.176
.075
4050 years
66
28,251
.786
[ .715, .841]
3333.4
.195
.054
5071 years
45
4,816
.926
[ .863, .961]
3005.2
.226
.138
Relationship duration
05 years
79
7,623
.556
[ .457, .641]
1375.9
.087
.030
510 years
99
21,400
.765
[ .686, .827]
6620.3
.140
.144
1020 years
79
18,507
.824
[ .761, .871]
4286.3
.195
.066
2046 years
55
8,071
.884
[ .785, .939]
8470.8
.370
.156
41
Note. Computations without control of time lag were made with multilevel random-effects models and computations with control of time lag
were made with multilevel mixed-effects meta-regression models. In the analyses with time lag as covariate, time lag was grand-mean
centered at 2.30 years. k = number of effect sizes; N = number of participants in each group. Weighted mean effect size = disattenuated
correlation coefficient, indicating the rank-order stability of relationship satisfaction between two assessments. CI = confidence interval; Q =
test statistic of the test for (residual) heterogeneity; σ12 = variance proportion attributable to the level of the grouping variable (i.e., between
samples); σ22 = variance proportion attributable to the level nested within the grouping variable (i.e., within samples). Values in bold are
significant at p < .05.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
42
Rank-Order Stability as a Function of Relationship Duration
Visual Overview. The scatterplot shown in Figure 3B provides an overview of the
relation between mean relationship duration and effect size. The variability of effect sizes was
larger in relationships with a duration of less than 20 years compared to relationships with a
duration of more than 20 years. The scatterplot also indicates that the effect sizes were larger
with longer relationship duration.
Again, we generated a LOESS curve to gain further information about the
developmental pattern of rank-order stability of relationship satisfaction as a function of
relationship duration. Figure 4B shows the LOESS curve with mean relationship duration as
predictor and the disattenuated correlation coefficient as outcome. The figure indicates an
increasing trend of rank-order stability as a function of relationship duration, with a marginal
decline after 20 years of relationship duration.
Weighted Mean Effect Sizes in Relationship-Duration Groups. For the meta-
analytic computations, we constructed four relationship-duration groups. For 0 to 10 years of
relationship duration, we constructed two groups with an interval of 5 years: 05 years and
510 years. For a relationship duration of more than 10 years, the number of samples was
lower. Therefore, we constructed one group from 10 to 20 years and one group from 20 to 46
years (46 years was the maximum relationship duration in the meta-analytic data set).
Table S3 reports descriptive information on time lag in the four relationship-duration
groups (see values in the lower half of the table). As the table indicates, the mean time lag
ranged from 1.36 to 3.30 years across relationship-duration groups, with the largest mean in
the group from 10 to 20 years.
Table 2 reports the meta-analytic estimates (both without control of time lag and with
control of time lag). Again, in the analyses that controlled for time lag, time lag was centered
at the mean of the meta-analytic data set (i.e., 2.30 years). In both sets of analyses, weighted
mean effect sizes were larger with longer relationship duration, suggesting that rank-order
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
43
stability of relationship satisfaction increases as a function of relationship duration. This
increase was particularly pronounced within the first ten years of the relationship.
Analyses With Time Lag
The scatterplot shown in Figure 3C provides an overview of the relation between time
lag and effect size. Furthermore, Figure S1 shows the frequency distribution of time lag
across samples. Although time lag ranged from 2 months to almost 20 years, most samples
had a time lag of less than 5 years. The correlation between time lag and effect size was .36
(p < .001), suggesting that rank-order stability was significantly lower over longer periods. To
further examine rank-order stability of relationship satisfaction depending on time lag, we
generated a LOESS curve. Figure 4C shows the LOESS curve with time lag as predictor and
the disattenuated correlation coefficient as outcome. The figure indicates a decreasing trend of
rank-order stability depending on time lag, suggesting that the curve levels off at an estimate
of about .55.
Meta-Regressions Testing Age and Relationship Duration Simultaneously
Analyses based on age and relationship-duration groups are a useful method for
describing the patterns of effects, but they might obscure the effects of continuous age and
relationship duration. Moreover, age and relationship duration were highly correlated with
each other (i.e., r = .97; see Table S4). Therefore, we tested multilevel mixed-effects meta-
regression models using both time metrics as continuous predictors, to disentangle the effects
of the two time metrics.
Specifically, we computed six models that differed systematically with regard to the
predictors included. Model 1 tested the linear effect of age. Model 2 tested the linear effect of
relationship duration. Model 3 tested the linear effects of age and relationship duration
simultaneously. Model 4 tested the linear and quadratic effects of age. Model 5 tested the
linear and quadratic effects of relationship duration. Finally, Model 6 tested the linear and
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
44
quadratic effects of age and relationship duration simultaneously. The findings from the meta-
regression models are reported in Table 3, with control of time lag.
The findings showed that both linear and quadratic age (Models 1 and 4) and linear
and quadratic relationship duration (Models 2 and 5) significantly predicted the effect sizes,
consistent with the conclusions from the analyses with age and relationship-duration groups
reported above. When linear age and linear relationship duration were tested simultaneously
(Model 3), age was no longer significant, but relationship duration remained significant.
Moreover, when quadratic age and quadratic relationship duration were included as additional
predictors (Model 6), linear relationship duration remained a significant predictor of the effect
sizes.
Overall, the meta-regression models allowed testing, with a different approach, rank-
order stability as a function of age and relationship duration. The findings indicate that age
and relationship duration significantly predicted rank-order stability of relationship
satisfaction, suggesting greater stability with higher age and with longer relationship duration.
Moreover, the results from the meta-regression models helped to disentangle the effect of age
and relationship duration, suggesting that relationship duration was the more dominant (and
significant) time metric for explaining rank-order stability of relationship satisfaction.
45
Table 3
Meta-Regression Predicting Effect Size from Age and Relationship Duration, Controlling for Time Lag
Model
Predictor
1
2
3
4
5
6
Linear age
.257 [.168, .347]
.140 [.463, .183]
.342 [.227, .457]
.208 [.544, .128]
Linear relationship
duration
.296 [.203, .390]
.440 [.096, .785]
.490 [.337, .643]
.710 [.301, 1.119]
Quadratic age
.070 [.130, .010]
.011 [.142, .119]
Quadratic relationship
duration
.114 [.186, .043]
.109 [.265, .046]
Note. k = 306. Computations were made with multilevel mixed-effects meta-regression models. Effect size = Fisher’s Z-transformed disattenuated
correlation coefficient, indicating the rank-order stability of relationship satisfaction between two assessments. The table shows unstandardized
regression coefficients, with 95% confidence intervals in brackets. To avoid numerically small estimates, age and relationship duration were rescaled by
the factor 101. Moreover, in the present analyses age was centered at 35 years, and relationship duration was centered at 10 years. Dash indicates that
this predictor is not included in the model. The analyses controlled for time lag, which was grand-mean centered at 2.30 years. Values in bold are
significant at p < .05.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
46
Moderator Analyses
Finally, we tested for moderators of the effect sizes, using multilevel mixed-effects
meta-regression models. The zero-order correlations between the moderators, age,
relationship duration, time lag, and effect size are provided in Table S4.
For the categorical moderator variables, we tested specific contrasts. As regards the
occurrence of transitions, we contrasted samples that married or had a baby between Time 1
and any of the following occasions (16%) with samples that had no transition between Time 1
and any of the following occasions (84%). We also contrasted samples that married or had a
baby shortly before Time 1 (16%) with samples that had no transition shortly before Time 1
(84%). As regards the measure of relationship satisfaction, we contrasted samples that were
assessed with an ad-hoc measure (17%) versus established scale (83%). Among those samples
that were assessed with an established scale, we contrasted global satisfaction scales (46%)
with adjustment scales (54%). For sample type, we contrasted nationally representative
samples (9%) with community and college samples (91%). For ethnicity, we contrasted
samples that were predominantly White/European (78%) with samples that had another
ethnicity (22%).
For the continuous variable baseline mean of relationship satisfaction, we had to take
into account that the primary studies used different measures to assess relationship
satisfaction, so that the observed Time 1 means were not directly comparable. Therefore, we
converted the Time 1 means into POMP scores to make them comparable across studies
(Cohen et al., 1999; see also Cerasoli, 2014). To compute POMP scores, we used the
following formula given by Cohen et al. (1999)
POMP = observed minimum
maximum minimum × 100,
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
47
where observed is the observed Time 1 mean of relationship satisfaction, minimum is
the scale’s minimum possible score, and maximum is the scale’s maximum possible score.
Thus, the POMP score reflects the Time 1 mean as percentage of the scale’s maximum
possible score, ranging from 0 to 100.
In the moderator analyses, we controlled for time lag and between-sample differences
in mean age and mean relationship duration. Controlling for time lag, mean age, and mean
relationship duration was essential given that effect sizes varied as a function of these
variables, as reported above (see also Table S4). Given the substantial number of tests in the
moderator analyses (i.e., 11 tests), we used the Bonferroni method and adjusted the
significance level to p < .0045 (i.e., dividing .05 by the number of tests).
The results of the moderator analyses are shown in Table 4 and indicated that the
occurrence of relationship transitions before Time 1 and sample type had significant
moderator effects. Specifically, samples who experienced (vs. did not experience) a
relationship transition shortly before Time 1 had a lower rank-order stability of relationship
satisfaction. Moreover, nationally representative samples had a lower rank-order stability of
relationship satisfaction than community or college samples. When we tested these two
moderators simultaneously, the coefficients were very similar to the coefficients from the first
step of the analyses and significant on the Bonferroni adjusted significance level.
5
None of the other study characteristics were significant, which strengthens confidence
in the robustness of the findings. That is, rank-order stability of relationship satisfaction did
not significantly differ by household shared with partner, marital status, presence of children,
the occurrence of relationship transitions between assessments, type of measure, ethnicity,
gender, and baseline mean of relationship satisfaction.
5
When testing the two moderators simultaneously, the coefficients were B = .280, SE = .090, p = .002
(occurrence of relationship transitions shortly before Time 1) and B = .490, SE = .154, p = .002 (sample type).
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
48
Table 4
Meta-Regression Models for Study Characteristics Predicting the Effect Size, Controlling for
Mean Age, Mean Relationship Duration, and Time Lag
Moderator
k
B
SE
p
Household shared with partner
141
.285
.244
.243
Married
275
.298
.169
.078
Presence of children
207
.196
.162
.228
Transition
306
.087
.115
.451
Post transition
306
.276
.091
.002
Measure (established)
306
.065
.125
.606
Measure (global)
249
.209
.105
.047
Sample type
306
.476
.153
.002
Ethnicity
291
.095
.120
.427
Female
306
.030
.107
.779
Baseline mean
265
.021
.009
.027
Note. Computations were made with multilevel mixed-effects meta-regression models. The
effect size used was the Fisher’s Z-transformed disattenuated correlation coefficient,
indicating the rank-order stability of relationship satisfaction between two assessments. Mean
age, mean relationship duration, and time lag were included as control variable in all models.
Mean age, mean relationship duration, time lag, and mean year of birth were grand-mean
centered prior to the analyses. Household shared with partner, married, presence of children,
and female were proportions. The following variables were dichotomous: transition (1 =
relationship transition between Time 1 and subsequent assessment, 0 = no relationship
transition between Time 1 and subsequent assessment ), post transition (1 = relationship
transition shortly before Time 1 , 0 = no relationship transition shortly before Time 1), the
two contrasts for measure of relationship satisfaction (1 = established scale, 0 = ad-hoc
measure; 1 = global satisfaction, 0 = adjustment measure), sample type (1 = nationally
representative sample, 0 = community or student sample), ethnicity (1 = predominantly
White, 0 = other). k = number of effect sizes.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
49
Discussion
The goal of this meta-analysis was to gain a robust and precise picture of rank-order
stability of relationship satisfaction across adulthood, as a function of age and relationship
duration. The meta-analytic data set was based on 148 independent samples, including 402
effect sizes from 153,396 participants. The mean age associated with the effect sizes ranged
from 19 to 71 years, and the mean relationship duration ranged from 3 months to 46 years.
The findings indicated that individual differences in relationship satisfaction within a given
relationship are relatively stable over time (average r = .76, corrected for attenuation due to
measurement error and based on an average time lag of 2.30 years). The findings also
suggested that rank-order stability of relationship satisfaction varied systematically as a
function of age and as a function of relationship duration. Specifically, rank-order stability of
relationship satisfaction increased from young to late adulthood, with a slight decrease in
middle adulthood. Moreover, rank-order stability of relationship satisfaction increased over
the course of the relationship, with a slight decline around 20 years of relationship duration.
In the analyses that examined age and relationship duration simultaneously, the findings
suggested that relationship duration was the more dominant time metric for explaining rank-
order stability of relationship satisfaction. Finally, the moderator analyses suggested that the
occurrence of relationship transitions shortly before Time 1 and sample type explained
variance in rank-order stability of relationship satisfaction. Except for these two moderators,
the pattern of findings was robust across study characteristics such as marital status, ethnicity,
and gender.
Rank-Order Stability of Relationship Satisfaction Across Adulthood
As noted in the Introduction, examining rank-order stability of relationship satisfaction
concerns the question to which degree relationship satisfaction should be conceptualized as a
trait-like construct (Fraley & Roberts, 2005). When comparing the present meta-analytic
estimate of rank-order stability of relationship satisfaction within a given relationship (r =
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
50
.76) with findings on other psychological constructs, it is important to note that this estimate
is disattenuated (i.e., corrected for unreliability of the measures used in individual studies)
and, therefore, comparison estimates should also be corrected for unreliability. Moreover,
comparison estimates should be based on a similar time lag (i.e., in the present research the
average time lag was 2.30 years). The data from Anusic and Schimmack (2016) suggest that
estimates of rank-order stability (corrected for measurement error and based on an average
time lag of 2.30) are about .88 for personality traits, .80 for self-esteem, .75 for life
satisfaction, and .88 for affect.
Thus, the present findings suggest that individual differences in relationship
satisfaction are less stable than individual differences in personality traits, self-esteem, and
affect, but as stable as individual differences in life satisfaction. Hence, similar to life
satisfaction, relationship satisfaction can be considered, to some degree, as a trait-like
construct. The relatively high rank-order stability of relationship satisfaction is particularly
interesting given that relationship satisfactionin contrast to life satisfactionis not a pure
characteristic of the individual, but is conceptually linked to the specific relationship
environment, including characteristics of the romantic partner and of the relationship with the
partner. It is likely that both individual predispositions (such as individual differences in
emotional stability, self-esteem, and secure attachment; e.g., McNulty, 2016) and relationship
characteristics (such as constructive communication patterns; e.g., Karney & Bradbury, 2020)
contribute to the relatively high rank-order stability of relationship satisfaction. Clearly, future
research using statistical models such as the STARTS model (Kenny & Zautra, 2001) is
needed to more directly estimate stable and unstable variance components in relationship
satisfaction (for an example of using the STARTS model, see Lucas & Donnellan, 2007).
Studies using the STARTS model could also contribute to examining the asymptote of rank-
order stability over long periods (Anusic & Schimmack, 2016; Fraley & Roberts, 2005;
Kuster & Orth, 2013; Schuerger et al., 1989).
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
51
In this research, we found that the pattern of how rank-order stability of relationship
satisfaction changed as a function of time lag was similar to that of other psychological
constructs. More precisely, we observed that the rank-order stability of relationship
satisfaction was significantly lower over longer periods (see Figure 4C). This decline could be
expected based on research on other psychological constructs, which generally show a
decrease in rank-order stability as the time lag increases (e.g. Fraley & Roberts, 2005). At the
same time, we found that the LOESS curve did not approach zero over long periods (e.g., 15
20 years), but leveled off at an estimate of about .55. This finding suggests that there is an
enduring component of individual differences in relationship satisfaction, again consistent
with general findings on the stability of individual differences in other psychological
constructs (e.g. Fraley & Roberts, 2005). To return to the example from the Introduction: The
findings imply that Heather, Tom, and Mary will likely maintain their relative standing on
relationship satisfaction at later occasions: Heather, who reports high relationship satisfaction
(compared to Tom and Mary) at the first assessment, is likely to report high relationship
satisfaction (compared to Tom and Mary) one year, five years, and even 15 years later. As
discussed in the Introduction, it is possible that the long-term stability of relationship
satisfaction can be explained by the long-term stability of the factors that influence couple
members’ relationship satisfaction. More specifically, the enduring component of individual
differences in relationship satisfaction could be explained by characteristics of the individuals
(e.g., personality characteristics such as the person’s attachment orientation), the relationship
(e.g., coping style of the couple), and the context (e.g., employment situation of the partners).
Thus, although rank-order stability of relationship satisfaction levels off at a large value when
assessed across long intervals, consistent with the notion of a relatively strong trait component
of the construct, rank-order stability is likely due to influences from the person and influences
from the relationship and context. As noted above, the present meta-analysis is based on data
from assessments within a given relationship. Therefore, future research should test whether
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
52
the longterm estimate of rank-order stability holds when relationship satisfaction is assessed
across different relationships of the same person.
At the same time, the present findings also showed that the degree of rank-order
stability was still substantially below unity, indicating that there is room for change (Roberts
& Nickel, 2021). To determine in which developmental periods change is more or less likely
to occur, we examined rank-order stability of relationship satisfaction as a function of age and
as a function of relationship duration.
Rank-Order Stability of Relationship Satisfaction as a Function of Age
The findings from both the LOESS curves and the group analyses indicated that,
overall, rank-order stability of relationship satisfaction increased as a function of age, with a
slight decline in middle adulthood. Specifically, the meta-analytic estimates from the group
analyses were .66 for 1930 years and .93 for 5071 years. The estimates were not
substantially altered when time lag between assessments was controlled for. The finding of
greater rank-order stability with higher age is in line with the theoretical perspectives
reviewed in the Introduction, according to which developmental tasks (Erikson, 1968;
Havighurst, 1972) and selective investment (Carstensen et al., 1999; Charles & Carstensen,
2010) contribute to increasing rank-order stability of relationship satisfaction across
adulthood.
The lower rank-order stability in young adulthood is consistent with research
indicating that young adults, and in particular emerging adults, explore different life-path
options, including romantic relationships (Arnett, 2000; Shulman & Connolly, 2013). Such
exploration may contribute to less stable relationship conditions (e.g., young adults often are
unmarried and do not live in a shared household; see also Table S4), which, in turn, could
contribute to lower rank-order stability of relationship satisfaction in this life stage. However,
in the present meta-analysis, being married and sharing a household did not moderate the
effect sizes (see Table 4). Thus, even if young adult couples are more often unmarried and
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
53
live more often in separate households, this likely does not explain the lower rank-order
stability in this developmental period.
Although the slight decline of rank-order stability in middle adulthood deviates from
the general upward trend across adulthood, this tentative drop may be related to specific
developmental challenges in midlife, which could affect romantic relationships in different
ways. For example, many middle adults (but not all) spend more time with work than with
leisure activities, and take on more generative, caring, and sometimes stressful social roles
(e.g., Hudson et al., 2019). These greater challenges might deplete their resources to invest in
the romantic relationship (Buck & Neff, 2012; Finkel & Campbell, 2001; Finkel et al., 2012;
Finkel et al., 2014), leading to a temporary decline in the rank-order stability of relationship
satisfaction in this developmental period. Nevertheless, we emphasize that the observed
decline in rank-order stability in middle adulthood was relatively small.
Finally, the relatively large stability of relationship satisfaction in late adulthood is
consistent with research suggesting that older adults invest more time and energy in positive
social relationships with close others, such as with their relationship partner (Carstensen et al.,
1999; Fredrickson & Carstensen, 1990; Fung et al., 1999). More investments, in turn, may
lead to more stable relationship conditions, contributing to greater rank-order stability of
relationship satisfaction in this life stage (Fraley & Roberts, 2005).
Rank-Order Stability of Relationship Satisfaction as a Function of Relationship Duration
The findings from both the LOESS curves and the group analyses indicated that rank-
order stability of relationship satisfaction generally increased as a function of relationship
duration, with a slight decline around 20 years of relationship duration. Specifically, the meta-
analytic estimates from the group analyses were .56 in the first 5 years of relationships and
.88 after a relationship duration of 20 years and longer. Again, the estimates were not
substantially altered when time lag was controlled for. Interestingly, the confidence intervals
were larger for couple members with long relationship duration (i.e., 35 years and more),
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
54
which may be due either to the lower number of samples who have been together for long
periods or to greater heterogeneity of rank-order stability in samples with long relationship
duration. Nevertheless, the overall trend suggested that rank-order stability generally
increases over the course of relationships. This finding is in line with the theoretical
perspectives reviewed in the Introduction, which suggested that in relationships that “survive”
the first years, relationship conditions tend to be more stable, which consequently leads to
increasing rank-order stability of relationship satisfaction as a function of relationship
duration (Diekmann & Mitter, 1984; Kulu, 2014; Kurdek, 1998, 1999).
There are at least two mechanisms that could explain this phenomenon. First, the
degree of rank-order stability depends on the stability of the context (see Fraley & Roberts,
2005). If a population as a whole (e.g., couples who have been together for 10 years or
longer) is embedded in relatively consistent environments, this contributes to greater rank-
order stability (Moss & Susman, 1980; Roberts & DelVecchio, 2000). By definition, in
relationships of long duration, romantic partners have spent a considerable amount of time
together and their relationship environment is likely characterized by accumulated
consistency and stability. Of course, this does not mean that individuals in relationships of
long duration do not experience any changes in their relationship environment, for example
due to diseases or transitioning into retirement (e.g., Specht et al., 2011). Yet, it seems that
these experiences do not generate substantial changes in the rank ordering of relationship
satisfaction over time. Moreover, couples who have spent a considerable amount of time
together often have invested many resources in their relationship, such as financial or
emotional resources. According to the investment model of relationships (Rusbult et al.,
1998), these investments (together with a relatively high level of relationship satisfaction and
low quality of alternatives) maintain relationship commitment and may contribute to greater
stability of the relationship. In addition, couples who are more satisfied in the relationship
anticipate that they will remain in the relationship and hence invest in more adaptive
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
55
relationship processes (e.g., constructive coping styles), strengthening their high level of
satisfaction. Similarly, less satisfied couples may not believe in continuing the relationship
and will invest much less in relationship-promoting behavior, which maintains their relatively
low level of satisfaction. Thus, these relationship processes may explain why individual
differences in relationship satisfaction are relatively stable over time.
Second, identity processes could also account for greater rank-order stability of
relationship satisfaction, combining psychological and environmental mechanisms (Roberts &
DelVecchio, 2000). Only recently, the concept of narrative identity has been applied to the
romantic relationship domain, illustrating that romantic relationshipssimilar to whole lives
(McAdams, 1995, 2013)can be represented by means of narrative identity (Bühler &
Dunlop, 2019). Specifically, the story that individuals form about their relationship reflects
their relationship identity. As argued in the context of individual lives, people with a strong
identity make decisions and choose life paths that are consistent with their personality and
their identity (see selective person-environment transactions; Caspi, 1998). Having a strong
sense of identity is also linked to continued investment in the chosen life path, which predicts
an increase in ego-resilience, that is, a better ability for adjusting to changes in the
environment and recovery from difficult situations (Pals, 1999). Moreover, a strong sense of
identity also exerts a filter function as to which information is perceived and processed by the
person. Finally, once a person’s identity has become known to others, they create reputations
about the person (Hogan & Roberts, 2000) and respond in a way that strengthens the person’s
personality and identity (see evocative person-environment transactions; Caspi, 1998).
Together, these identity processes may contribute to greater stability of individual differences
in psychological constructs (Roberts & DelVecchio, 2000).
Applied to the context of romantic relationships, this means that individuals with a
strong relationship identity might make decisions and choose relationship paths that are in
line with their relationship identity (e.g., seeing a couple therapist when relationship problems
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
56
arise). Having a strong relationship identity may also be linked to continuous investment in
the relationship, which contributes to a resilient and stable relationship. Moreover, people
with a strong relationship identity may selectively perceive and process information in a way
that strengthens their relationship identity (e.g., focusing on positive, rather than negative,
attributes of the relationship or the relationship partner). Finally, the reputation of the
relationship (e.g., “Max and Alex have been happily married ever since”) may feedback into
the relationship so that others react to the couple in a way that is consistent with their
established relationship identity. Clearly, the same relationship-identity mechanisms apply to
less satisfied couples. For instance, individuals who are less satisfied with their relationship
may make decisions that are in line with their relationship identity (e.g., “We do not fit
together anyway”) and may not consider, or even decline, to see a couple therapist when
problems arise.
At the same time, meta-analytic findings on mean levels suggest that relationship
satisfaction is, on average, relatively high over the course of the relationship, never falling
below 77% of the maximum possible (Bühler et al., 2021). These findings also correspond
with research based on latent class growth analysis and group-based modeling, indicating that
the majority of couples (around 7090%) remains fairly satisfied over time, while only a
subgroup of couples (around 1030%) experiences greater declines in relationship satisfaction
and potentially a relationship breakup (e.g., Anderson et al., 2010; Birditt et al., 2012; Foran
et al., 2013; Lavner & Bradbury, 2010; Lavner et al., 2012; Lorber et al., 2015). Thus, among
intact couples, even the less satisfied couple members are often relatively satisfied, which
may explain why they remain in the relationship over longer periods of time. In addition, as
described in the investment model, the level of relationship satisfaction is only one of the
factors that contributes to commitment and stability in the relationship. Hence, among the
intact, less satisfied couples there may be other factors, such as the investments that couple
members have already made, that motivate partners to stay together. Together, all of these
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
57
mechanisms may stabilize individual differences in relationship satisfaction and contribute to
greater rank-order stability of relationship satisfaction over the course of the relationship.
The Role of Age and Relationship Duration
As noted in the Introduction, age and relationship duration are generally highly
correlated, which was also the case in the present meta-analysis (r = .97; see Table S4). This
high correlation, however, is not a methodological artifact, but reflects the reality of most
couples: The older people are, the longer they have typically been in their relationship, simply
because they are older. Nevertheless, it is important to note that although the two time metrics
are highly correlated, they are not interchangeable (similar to height and weight, which are
strongly correlated but by no means the same construct; Fincham et al., 2018). Therefore, we
tested multilevel mixed-effects meta-regression models using both time metrics as continuous
predictors, with the goal of disentangling the effects of the two time metrics. The findings
suggested that relationship duration, rather than age, was the more dominant time metric for
explaining rank-order stability of relationship satisfaction within a given relationship. As
reviewed in the Introduction, Fraley and Roberts’ (2005) model suggests that stable
environments contribute to higher rank-order stability of individual-difference constructs. The
more dominant role of relationship duration for rank-order stability of relationship satisfaction
might be explained by the more stable relationship environments that longterm couples
usually experience. This, in turn, may result in more stable transactions between couple
members and their relationship environments (Fraley & Roberts, 2005). With longer
relationship duration, relationship partners also tend to become more congruent in personality
traits such as agreeableness, conscientiousness, and openness (Rammstedt & Schupp, 2008),
which may further stabilize the transactions between the partners and their environment over
the course of their relationship.
At the same time, we emphasize that it was not possible to unequivocally isolate the
unique effects of the time metrics because the time metrics are, as described above, inherently
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
58
linked to each other. Thus, in studies with romantic couples both age and relationship duration
should be examined simultaneously to account for both of their effects. Otherwise,
developmental processes in romantic relationships might mistakenly be attributed to one of
the time metrics, simply because the other time metric was not examined in the research.
Moderators of the Rank-Order Stability of Relationship Satisfaction
The pattern of findings was relatively robust across the moderators tested, with two
exceptions. First, in line with our hypotheses the occurrence of relationship transitions shortly
before Time 1 predicted lower rank-order stability of relationship satisfaction. Relationship
transitions, such as the birth of a child or marriage, might confront people with new demands
and challenges and might destabilize relationship environments (Doss et al., 2009; Fraley &
Roberts, 2005; Sanchez & Thomson, 1997). A destabilized relationship environment, in turn,
may lead to lower rank-order stability of relationship satisfaction. Second, nationally
representative samples had a lower rank-order stability of relationship satisfaction than
community samples and samples of college students. As noted in the Introduction, we had no
hypotheses about the moderating effect of sample type on the effect size. The present findings
suggest that romantic relationships in community and college samples may be more stable
than romantic relationships in nationally representative samples. However, more research is
needed for drawing stronger conclusions about the moderating effects of relationship
transitions and sample type on the rank-order stability of relationship satisfaction.
A general conclusion from the moderator analyses is that the findings were robust
across most sample and methodological characteristics. In other words, rank-order stability of
relationship satisfaction did not significantly differ in samples with different compositions
with regard to shared household, marital status, presence of children, ethnicity, and gender.
Moreover, the findings did not differ in samples that had versus had not undergone
relationship transitions between Time 1 and any of the following measurement occasions, in
samples with different measures and with different baseline means of relationship satisfaction.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
59
Limitations and Future Directions
Although this meta-analysis included data from 153,396 participants from 16
countries, one limitation is that the majority of samples were from Western countries (e.g.,
United States, Canada, the Netherlands, Germany). Given that the meta-analytic data set
included few studies from Asian countries and none from African and South American
countries, the present data did not allow testing whether the results hold outside of North
America and Europe. Moreover, nearly all samples consisted of participants from opposite-
sex relationships, and only one sample included participants from same-sex relationships.
Therefore, more primary studies are needed that examine couples from different cultural
backgrounds (Henrich et al., 2010) and individuals involved in diverse types of relationships,
including same-sex relationships (e.g., Chen & van Ours, 2018).
Furthermore, the meta-analytic data set might be selective to some degree because
samples were only included if they provided data on at least two measurement occasions
(thus, couples who separated after the first assessment did not provide the second assessment
needed for computing rank-order stability). To reduce concerns related to this limitation, we
used a relatively short time lag between assessments (i.e., 2 months). Nevertheless, these
“surviving” relationships might still represent relationships that are above average in their
commitment given that dissolving relationships often show higher levels of stress and
dissimilarity, both at baseline and over time (Finn et al., 2020). Therefore, the present
findings should be interpreted in light of these more stable relationship environments, and
samples of to-be-dissolved couples might show lower degrees of rank-order stability of
relationship satisfaction. Similarly, the present data allow conclusions about rank-order
stability of relationship satisfaction within a given relationship but cannot speak to rank-order
stability of relationship satisfaction across relationships. Research that follows individuals
within and across relationships would generate knowledge about such different degrees of
rank-order stability and, moreover, provide important knowledge on the trait-like nature of
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
60
relationship satisfaction across relationships (for cross-relationship research, see Johnson &
Neyer, 2019).
Another limitation is that the number of samples was relatively low for samples with a
mean age of above 50 years and with a relationship duration of above 20 years. Consequently,
the conclusions for rank-order stability of relationship satisfaction are less strong for these
developmental and relationship periods. In addition, there were no effect sizes available for
samples older than 71 years, which could be explained, at least partially, by health constraints
in this age group or by a higher likelihood that one of the partners was already deceased.
Future research is needed that specifically studies romantic relationships and relationship
trajectories in middle adulthood and especially late adulthood, including “gray divorces”
(Brown & Lin, 2012; Gloor et al., 2021). Finally, studying newlyweds in late adulthood
would contribute to lowering the correlation between age and relationship duration and would
provide further insights into the rank-order stability of relationship satisfaction as a function
of both age and relationship duration.
Conclusion
The present meta-analysis examined the rank-order stability of relationship
satisfaction, synthesizing data from 148 samples with more than 150,000 participants.
Overall, the findings showed that individual differences in relationship satisfaction are
relatively stable within a given relationship. The moderator analyses strengthened confidence
in this finding, given that rank-order stability was robust across most sample and
methodological characteristics. Moreover, the findings indicated that rank-order stability of
relationship satisfaction changes systematically as a function of age and of relationship
duration. For both time metrics, there was an overall increase of rank-order stability over
time. As regards age, rank-order stability increased from young to late adulthood, with a
slight decrease in middle adulthood. As regards relationship duration, rank-order stability
generally increased over the course of the relationship, with a slight decline around 20 years
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
61
of relationship length. When both time metrics were examined simultaneously, relationship
duration tended to be the more dominant time metric for explaining rank-order stability of
relationship satisfaction. Taken together, the present findings may stimulate future research
on the developmental processes that underlie stability and change of relationship satisfaction
across adulthood.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
62
References
References marked with an asterisk indicate studies included in the meta-analysis.
Anderson, J. R., Van Ryzin, M. J., & Doherty, W. J. (2010). Developmental trajectories of
marital happiness in continuously married individuals: A group-based modeling
approach. Journal of Family Psychology, 24(5), 587596.
https://doi.org/10.1037/a0020928
*Andres, M. (2014). Distress, support, and relationship satisfaction during military-induced
separations: A longitudinal study among spouses of Dutch deployed military
personnel. Psychological Services, 11(1), 2230. https://doi.org/10.1037/a003375022
Anusic, I., & Schimmack, U. (2016). Stability and change of personality traits, self-esteem,
and well-being: Introducing the meta-analytic stability and change model of retest
correlations. Journal of Personality and Social Psychology, 110(5), 766781.
https://doi.org/https://doi.org/10.1037/pspp0000066
Appelbaum, M., Cooper, H., Kline, R. B., Mayo-Wilson, E., Nezu, A. M., & & Rao, S. M.
(2018). Journal article reporting standards for quantitative research in psychology: The
APA Publications and Communications Board task force report. American
Psychologist, 73(1), 325. https://doi.org/10.1037/amp0000191
Ardelt, M. (2000). Still stable after all these years? Personality stability theory revisited.
Social Psychology Quarterly, 63(4), 392405. https://doi.org/10.2307/2695848
Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through
the twenties. American Psychologist, 55(5), 469480. https://doi.org/10.1037/0003-
066x.55.5.469
*Be, D., Whisman, M. A., & Uebelacker, L. A. (2013). Prospective associations between
marital adjustment and life satisfaction. Personal Relationships, 20(4), 728739.
https://doi.org/10.1111/pere.12011
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
63
Belsky, J., & Rovine, M. (1990). Patterns of marital change across the transition to
parenthood: Pregnancy to three years postpartum. Journal of Marriage and the
Family, 52(1), 519. https://doi.org/10.2307/352833
*Bikos, L. H., Çiftçi, A., Güneri, O. Y., Demir, C. E., Sümer, Z. H., Danielson, S., DeVries,
S., & Bilgen, W. A. (2007). A repeated measures investigation of the first-year
adaptation experiences of the female expatriate spouse living in Turkey. Journal of
Career Development, 34(1), 526. https://doi.org/10.1177/0894845307304063
Birditt, K. S., Hope, S., Brown, E., & Orbuch, T. (2012). Developmental trajectories of
marital happiness over 16 years. Research in Human Development, 9(2), 126144.
https://doi.org/10.1080/15427609.2012.680844
*Bloch, L., Haase, C. M., & Levenson, R. W. (2014). Emotion regulation predicts marital
satisfaction: More than a wives' tale. Emotion, 14(1), 130144.
https://doi.org/10.1037/a0034272
Block, J. (2014). Lives through time. Psychology Press.
*Blumenstock, S. M., & Papp, L. M. (2017). Sexual distress and marital quality of
newlyweds: An investigation of sociodemographic moderators. Family Relations: An
Interdisciplinary Journal of Applied Family Studies, 66(5), 794808.
https://doi.org/10.1111/fare.12285
Bodenmann, G. (1995). A systemic-transactional conceptualization of stress and coping in
couples. Swiss Journal of Psychology, 54(1), 3449.
*Bodi, O., Mikula, G., & Riederer, B. (2010). Long-term effects between perceived justice in
the division of domestic work and women's relationship satisfaction: Testing for
moderation effects. Social Psychology, 41(2), 5765. https://doi.org/10.1027/1864-
9335/a000009
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
64
*Bouchard, G. (2014). The quality of the parenting alliance during the transition to
parenthood. Canadian Journal of Behavioural Science/Revue canadienne des sciences
du comportement, 46(1), 2028. https://doi.org/10.1037/a003125920
*Bouchard, G., Boudreau, J., & Hébert, R. (2006). Transition to parenthood and conjugal life:
Comparisons between planned and unplanned pregnancies. Journal of Family Issues,
27(11), 15121531. https://doi.org/10.1177/0192513X06290855
*Bower, D., Jia, R., Schoppe-Sullivan, S. J., Mangelsdorf, S. C., & Brown, G. L. (2013).
Trajectories of couple relationship satisfaction in families with infants: The roles of
parent gender, personality, and depression in first-time and experienced parents.
Journal of Social and Personal Relationships, 30(4), 389409.
https://doi.org/10.1177/0265407512457656
Bowlby, J. (1973). Attachment and loss: Vol. II. Separation: Anxiety and anger. Basic Books.
*Brown, B. J., Robinson, D., Jensen, J. F., Seedall, R. B., Hodgson, J., & Norton, M. C.
(2019). Will improving my marriage improve my sleep? Journal of Couple &
Relationship Therapy, 18(2), 85103. https://doi.org/10.1080/15332691.2017.1417938
Brown, S. L., & Lin, I.-F. (2012). The gray divorce revolution: Rising divorce among middle-
aged and older adults, 19902010. The Journals of Gerontology: Series B, 67(6), 731
741. https://doi.org/10.1093/geronb/gbs089
Buck, A. A., & Neff, L. A. (2012). Stress spillover in early marriage: The role of self-
regulatory depletion. Journal of Family Psychology, 26(5), 698708.
https://doi.org/10.1037/a0029260
Bühler, J. L., & Dunlop, W. L. (2019). The narrative identity approach and romantic
relationships. Social and Personality Psychology Compass, 13(4), Article e12447.
https://doi.org/10.1111/spc3.12447
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
65
Bühler, J. L., Krauss, S., & Orth, U. (2021). Development of relationship satisfaction across
the life span: A systematic review and meta-analysis. Psychological Bulletin, 147(10),
10121053. https://doi.org/http://dx.doi.org/10.1037/bul0000342
Bühler, J. L., Weidmann, R., Nikitin, J., & Grob, A. (2019). A closer look at life goals across
adulthood: Applying a developmental perspective to content, dynamics, and outcomes
of goal importance and goal attainability. European Journal of Personality, 33(3),
359384. https://doi.org/10.1002/per.2194
*Busby, D. M., & Gardner, B. C. (2008). How do I analyze thee? Let me count the ways:
Considering empathy in couple relationships using self and partner ratings. Family
Process, 47(2), 229242. https://doi.org/10.1111/j.1545-5300.2008.00250.x
*Buyukcan-Tetik, A., Campbell, L., Finkenauer, C., Karremans, J. C., & Kappen, G. (2017).
Ideal standards, acceptance, and relationship satisfaction: Latitudes of differential
effects. Frontiers in Psychology, 8, 1691. https://doi.org/10.3389/fpsyg.2017.01691
*Byers, E. S. (2005). Relationship satisfaction and sexual satisfaction: A longitudinal study of
individuals in long-term relationships. Journal of Sex Research, 42(2), 113118.
https://doi.org/10.1080/00224490509552264
Campbell, L., Simpson, J. A., Boldry, J., & Kashy, D. A. (2005). Perceptions of Conflict and
Support in Romantic Relationships: The Role of Attachment Anxiety. Journal of
Personality and Social Psychology, 88(3), 510531.
https://doi.org/https://doi.org/10.1037/0022-3514.88.3.510
Carr, D., & Utz, R. L. (2020). Families in later life: A decade in review. Journal of Marriage
and Family, 82(1), 346363. https://doi.org/10.1111/jomf.12609
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory
of socioemotional selectivity. American Psychologist, 54(3), 165181.
https://doi.org/10.1037/0003-066x.54.3.165
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
66
Caspi, A. (1998). Personality development across the life course. In W. Damon & N.
Eisenberg (Eds.), Handbook of child psychology: Social, emotional, and personality
development (pp. 311388). John Wiley & Sons Inc.
Caspi, A., & Bem, D. J. (1990). Personality continuity and change across the life course. In L.
A. Pervin (Ed.), Handbook of personality: Theory and research (pp. 549575).
Guilford Press.
Caspi, A., Bem, D. J., & Elder, G. H., Jr. (1989). Continuities and consequences of
interactional styles across the life course. Journal of Personality, 57(2), 375406.
https://doi.org/10.1111/j.1467-6494.1989.tb00487.x
Caspi, A., & Roberts, B. W. (1999). Personality continuity and change across the life course.
In Handbook of personality: Theory and research (pp. 300326). Guilford Press.
Cast, A. D. (2004). Well-being and the transition to parenthood: An identity theory approach.
Sociological perspectives, 47(1), 5578. https://doi.org/10.1525/sop.2004.47.1.55
Cerasoli, C. P., Nicklin, J. M., & Ford, M. T. (2014). Intrinsic motivation and extrinsic
incentives jointly predict performance: A 40-year meta-analysis. Psychological
Bulletin, 140(4), 9801008. https://doi.org/10.1037/a0035661
Charles, S. T., & Carstensen, L. L. (2010). Social and emotional aging. Annual Review of
Psychology, 61(1), 383409. https://doi.org/10.1146/annurev.psych.093008.100448
Chen, S., & van Ours, J. C. (2018). Subjective Well-being and Partnership Dynamics: Are
Same-Sex Relationships Different? Demography, 55(6), 22992320.
https://doi.org/10.1007/s13524-018-0725-0
*Choi, E. (2016). Marital satisfaction and maternal depressive symptoms among Korean
mothers transitioning to parenthood. Journal of Family Psychology, 30(4), 516521.
https://doi.org/10.1037/fam0000178
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
67
*Christopher, C., Umemura, T., Mann, T., Jacobvitz, D., & Hazen, N. (2015). Marital quality
over the transition to parenthood as a predictor of coparenting. Journal of Child and
Family Studies, 24(12), 36363651. https://doi.org/10.1007/s10826-015-0172-0
Cleveland, W. S. (1978). Robust locally weighted regression and smoothing scatterplots.
Journal of the American Statistical Association 74(368), 829836.
https://doi.org/10.1080/01621459.1979.10481038
Cleveland, W. S. (1981). LOWESS: A Program for smoothing scatterplots by robust locally
weighted regression. The American Statistician, 35(1), 54.
https://doi.org/10.2307/2683591
Cleveland, W. S., & Devlin, S. J. (1988). Locally weighted regression: An approach to
regression analysis by local fitting. Journal of the American Statistical Association,
83(403), 596−610. https://doi.org/10.1080/01621459.1988.10478639
Cohen, P., Cohen, J., Aiken, L. S., & West, S. G. (1999). The problem of units and the
circumstance for POMP. Multivariate Behavioral Research, 34(3), 315346.
https://doi.org/10.1207/s15327906mbr3403_2
Conger, R. D., Conger, K. J., & Martin, M. J. (2010). Socioeconomic status, family processes,
and individual development. Journal of Marriage and Family, 72(3), 685704.
https://doi.org/10.1111/j.1741-3737.2010.00725.x
*Crocker, J., Canevello, A., & Lewis, K. A. (2017). Romantic relationships in the ecosystem:
Compassionate goals, nonzero-sum beliefs, and change in relationship quality. Journal
of Personality and Social Psychology, 112(1), 58-75.
https://doi.org/10.1037/pspi0000076
Debrot, A., Cook, W. L., Perrez, M., & Horn, A. B. (2012). Deeds matter: Daily enacted
responsiveness and intimacy in couples' daily lives. Journal of Family Psychology,
26(4), 617627. https://doi.org/https://doi.org/10.1037/a0028666
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
68
*DeMaris, A. (2010). The 20-year trajectory of marital quality in enduring marriages: Does
equity matter? Journal of Social and Personal Relationships, 27(4), 449471.
https://doi.org/10.1177/0265407510363428
Diekmann, A., & Mitter, P. (1984). A comparison of the “sickle function” with alternative
stochastic models of divorce rates. In A. Diekmann & P. Mitter (Eds.), Stochastic
modelling of social processes (pp. 123153). Academic Press.
*Doohan, E.-A. M., Carrere, S., & Riggs, M. L. (2010). Using relational stories to predict the
trajectory toward marital dissolution: The oral history interview and spousal feelings
of flooding, loneliness, and depression. Journal of Family Communication, 10(1), 57
77. https://doi.org/10.1080/15267430903396401
Doss, B. D., Rhoades, G. K., Stanley, S. M., & Markman, H. J. (2009). The effect of the
transition to parenthood on relationship quality: An 8-year prospective study. Journal
of Personality and Social Psychology, 96. https://doi.org/10.1037/a0013969
Ebner, N. C., Freund, A. M., & Baltes, P. B. (2006). Developmental changes in personal goal
orientation from young to late adulthood: From striving for gains to maintenance and
prevention of losses. Psychology and Aging, 21(4), 664678.
https://doi.org/10.1037/0882-7974.21.4.664.
Egger, M., Davey Smith, G., Schneider, M., & Minder, C. (1997). Bias in meta-analysis
detected by a simple, graphical test. British Medical Journal, 315(7109), 629634.
https://doi.org/10.1136/bmj.315.7109.629
Erikson, E. H. (1968). Identity: Youth and crisis. Norton.
*Erol, R. Y., & Orth, U. (2014). Development of self-esteem and relationship satisfaction in
couples: Two longitudinal studies. Developmental Psychology, 50(9), 22912303.
https://doi.org/10.1037/a0037370
Erol, R. Y., & Orth, U. (2016). Self-esteem and the quality of romantic relationships.
European Psychologist, 21(4), 274283. https://doi.org/10.1027/1016-9040/a000259
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
69
*Fallis, E. E., Rehman, U. S., Woody, E. Z., & Purdon, C. (2016). The longitudinal
association of relationship satisfaction and sexual satisfaction in long-term
relationships. Journal of Family Psychology, 30(7), 822831.
https://doi.org/10.1037/fam0000205
*Fincham, F. D., & Beach, S. R. H. (2007). Forgiveness and marital quality: Precursor or
consequence in well-established relationships? The Journal of Positive Psychology,
2(4), 260286. https://doi.org/10.1080/17439760701552360
Fincham, F. D., Rogge, R., & Beach, S. R. H. (2018). Relationship satisfaction. In A. L.
Vangelisti & D. Perlman (Eds.), The Cambridge handbook of personal relationships
(pp. 422436). Cambridge University Press.
https://doi.org/10.1017/9781316417867.033
Finkel, E. J., & Campbell, W. K. (2001). Self-control and accommodation in close
relationships: An interdependence analysis. Journal of Personality and Social
Psychology, 81(2), 263277. https://doi.org/10.1037/0022-3514.81.2.263
Finkel, E. J., DeWall, C. N., Slotter, E. B., McNulty, J. K., Pond, R. S., Jr., & Atkins, D. C.
(2012). Using I3 theory to clarify when dispositional aggressiveness predicts intimate
partner violence perpetration. Journal of Personality and Social Psychology, 102(3),
533549. https://doi.org/10.1037/a0025651
Finkel, E. J., Hui, C. M., Carswell, K. L., & Larson, G. M. (2014). The suffocation of
marriage: Climbing Mount Maslow without enough oxygen. Psychological Inquiry,
25(1), 141. https://doi.org/10.1080/1047840x.2014.863723
Finn, C., Johnson, M. D., & Neyer, F. J. (2020). Happily (n)ever after? Codevelopment of
romantic partners in continuing and dissolving unions. Developmental Psychology,
56(5), 1022-1028. https://doi.org/10.1037/dev0000897
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
70
Finn, C., Mitte, K., & Neyer, F. J. (2013). The relationship-specific interpretation bias
mediates the link between neuroticism and satisfaction in couples. European Journal
of Personality, 27(2), 200212. https://doi.org/10.1002/per.1862
Fisher, R. A. (1921). On the “probable error” of a coefficient of correlation deduced from a
small sample. Metron, 1, 132.
Foran, H. M., Hahlweg, K., Kliem, S., & O'Leary, K. D. (2013). Longitudinal patterns of
relationship adjustment among German parents. Journal of Family Psychology, 27(5),
838843. https://doi.org/10.1037/a0034183
Fraley, R. C., & Roberts, B. W. (2005). Patterns of continuity: A dynamic model for
conceptualizing the stability of individual differences in psychological constructs
across the life course. Psychological Review, 112(1), 6074.
https://doi.org/10.1037/0033-295x.112.1.60
Frederick, D. A., Lever, J., Gillespie, B. J., & Garcia, J. R. (2017). What keeps passion alive?
Sexual satisfaction is associated with sexual communication, mood setting, sexual
variety, oral sex, orgasm, and sex frequency in a national U.S. Study. Journal of Sex
Research, 54(2), 186201. https://doi.org/10.1080/00224499.2015.1137854
Fredrickson, B. L., & Carstensen, L. L. (1990). Choosing social partners: How old age and
anticipated endings make people more selective. Psychology and Aging, 5(3), 335
347. https://doi.org/10.1037/0882-7974.5.3.335
Freund, A. M. (2008). Successful aging as management of resources: The role of selection,
optimization, and compensation. Research in Human Development, 5(2), 94106.
https://doi.org/10.1080/15427600802034827
Freund, A. M., & Nikitin, J. (2012). Junges und mittleres Erwachsenenalter. In W. Schneider
& U. Lindenberger (Eds.), Entwicklungspsychologie (pp. 259282). Beltz.
Freund, A. M., & Ritter, J. O. (2009). Midlife crisis: A debate. Gerontology, 55(5), 582591.
https://doi.org/10.1159/000227322
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
71
Fujita, F., & Diener, E. (2005). Life satisfaction set point: Stability and change. Journal of
Personality and Social Psychology, 88(1), 158164.
https://doi.org/https://doi.org/10.1037/0022-3514.88.1.158
Fung, H. H., Carstensen, L. L., & Lutz, A. M. (1999). Influence of time on social preferences:
Implications for life-span development. Psychology and Aging, 14(4), 595604.
https://doi.org/10.1037/0882-7974.14.4.595
Funk, J. L., & Rogge, R. D. (2007). Testing the ruler with item response theory: increasing
precision of measurement for relationship satisfaction with the Couples Satisfaction
Index. Journal of Family Psychology, 21(4), 572583. https://doi.org/10.1037/0893-
3200.21.4.572
*Gao, M., & Cummings, E. M. (2019). Understanding parent-child relationship as a
developmental process: Fluctuations across days and changes over years.
Developmental Psychology, 55(5), 10461058. https://doi.org/10.1037/dev0000680
*Girme, Y. U., Overall, N. C., & Faingataa, S. (2014). “Date nights” take two: The
maintenance function of shared relationship activities. Personal Relationships, 21(1),
125149. https://doi.org/10.1111/pere.12020
Gloor, S., Gonin-Spahni, S., Znoj, H., & Perrig-Chiello, P. (2021). Repartnering and
trajectories of life satisfaction after separation and divorce in middle and later life.
Journal of Social and Personal Relationships.
https://doi.org/10.1177/02654075211009594 (Advance online publication)
Gottman, J. M., & Levenson, R. W. (1999). Rebound from marital conflict and divorce
prediction. Family Process, 38(3), 287292. https://doi.org/10.1111/j.1545-
5300.1999.00287.x
*Gray, J. S., & Ozer, D. J. (2019). Comparing two models of dyadic change: Correlated
growth versus common fate. Social Psychological and Personality Science, 10(7),
957965. https://doi.org/10.1177/1948550618799066
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
72
*Greving Mehall, K., Spinrad, T. L., Eisenberg, N., & Gaertner, B. M. (2009). Examining the
relations of infant temperament and couples' marital satisfaction to mother and father
involvement: A longitudinal study. Fathering, 7(1), 2348.
https://doi.org/10.3149/fth.0701.23
*Gustavson, K., Roysamb, E., Borren, I., Torvik, F. A., & Karevold, E. (2016). Life
satisfaction in close relationships: Findings from a longitudinal study [Peer
Reviewed]. Journal of Happiness Studies: An Interdisciplinary Forum on Subjective
Well-Being, .17(3), pp. https://doi.org/10.1007/s10902-015-9643-7
*Hagemeyer, B., Neberich, W., 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), 390402.
https://doi.org/10.1111/jopy.12016
*Hakanen, J. J., Peeters, M. C. W., & Perhoniemi, R. (2011). Enrichment processes and gain
spirals at work and at home: A 3-year cross-lagged panel study. Journal of
Occupational and Organizational Psychology, 84(1), 830.
https://doi.org/10.1111/j.2044-8325.2010.02014.x
Halpern-Meekin, S., Manning, W. D., Giordano, P. C., & Longmore, M. A. (2013).
Relationship churning in emerging adulthood: On/off relationships and sex with an ex.
Journal of Adolescent Research, 28(2), 166188.
https://doi.org/10.1177/0743558412464524
*Hammond, M. D., & Overall, N. C. (2014). Endorsing benevolent sexism magnifies
willingness to dissolve relationships when facing partner-ideal discrepancies. Personal
Relationships, 21(2), 272287. https://doi.org/10.1111/pere.12031
Havighurst, R. J. (1972). Developmental tasks and education (3rd ed.). McKay.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
73
Heckhausen, J., Dixon, R. A., & Baltes, P. B. (1989). Gains and losses in development
throughout adulthood as perceived by different adult age groups. Developmental
Psychology, 25(1), 109121. https://doi.org/10.1037/0012-1649.25.1.109
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Academic Press.
Hendrick, S. S. (1988). A generic measure of relationship satisfaction. Journal of Marriage
and the Family, 50(1), 9398. https://doi.org/10.2307/352430
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences, 33(23), 6183.
https://doi.org/10.1017/s0140525x0999152x
*Hernandez-Kane, K. M., & Mahoney, A. (2018). Sex through a sacred lens: Longitudinal
effects of sanctification of marital sexuality. Journal of Family Psychology, 32(4),
425434. https://doi.org/10.1037/fam0000392
Heuveline, P., & Timberlake, J. M. (2004). The role of cohabitation in family formation: The
United States in comparative perspective. Journal of Marriage and Family, 66(5),
12141230. https://doi.org/10.1111/j.0022-2445.2004.00088.x
Hill, R. (1958). Generic features of families under stress. Social Casework, 39(23), 139150.
https://doi.org/10.1177/1044389458039002-318
Hogan, R., & Roberts, B. W. (2000). A socioanalytic perspective on person-environment
interaction. In W. B. Walsh, K. H. Craik, & H. Price (Eds.), Person-environment
psychology: New directions and perspectives (2nd ed., pp. 123). Lawrence Erlbaum
Associates Publishers.
*Hsiao, Y.-L. (2017). Longitudinal changes in marital satisfaction during middle age in
Taiwan. Asian Journal of Social Psychology, 20(1), 2232.
https://doi.org/10.1111/ajsp.12161
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
74
Hudson, N. W., Lucas, R. E., & Donnellan, M. B. (2019). Healthier and happier? A 3-year
longitudinal investigation of the prospective associations and concurrent changes in
health and experiential well-being [Peer Reviewed]. Personality and Social
Psychology Bulletin, .45(12), pp. https://doi.org/10.1177/0146167219838547
30975027
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias
in research findings. Sage.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias
in research findings (2nd ed.). Sage.
Hunter, J. E., & Schmidt, F. L. (2014). Methods of meta-analysis: Correcting error and bias
in research findings (3rd ed.). Sage.
Huston, T. L., Caughlin, J. P., Houts, R. M., Smith, S. E., & George, L. J. (2001). The
connubial crucible: Newlywed years as predictors of marital delight, distress, and
divorce. Journal of Personality and Social Psychology, 80(2), 237252.
https://doi.org/10.1037/0022-3514.80.2.237
Huston, T. L., & Houts, R. M. (1998). The psychological infrastructure of courtship and
marriage: The role of personality and compatibility in romantic relationships. In T. N.
Bradbury (Ed.), The developmental course of marital dysfunction (pp. 114151).
Cambridge University Press. https://doi.org/10.1017/CBO9780511527814.006
Hutteman, R., Hennecke, M., Orth, U., Reitz, A. K., & Specht, J. (2014). Developmental tasks
as a framework to study personality development in adulthood and old age. European
Journal of Personality, 28(3), 267278. https://doi.org/10.1002/per.1959
Ickes, W., Snyder, M., & Garcia, S. (1997). Personality influences on the choice of situations.
In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp.
166198). Academic Press.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
75
*Impett, E. A., Kogan, A., English, T., John, O., Oveis, C., Gordon, A. M., & Keltner, D.
(2012). Suppression sours sacrifice: Emotional and relational costs of suppressing
emotions in romantic relationships. Personality and Social Psychology Bulletin, 38(6),
707720. https://doi.org/10.1177/0146167212437249
*Ivanova, K. (2016). Relationship satisfaction of the previously married: The significance of
relationship specific costs and rewards in first and repartnering unions. Journal of
Social and Personal Relationships, 33(5), 559580.
https://doi.org/10.1177/0265407515583942
*Jayamaha, S. D., & Overall, N. C. (2015). Agents' self‐esteem moderates the effectiveness of
negative‐direct partner regulation strategies. Personal Relationships, 22(4), 738761.
https://doi.org/10.1111/pere.12108
*Jenkins, A. I., Fredman, S. J., Le, Y., Sun, X., Brick, T. R., Skinner, O. D., & McHale, S. M.
(2020). Prospective associations between depressive symptoms and marital
satisfaction in Black couples. Journal of Family Psychology, 34(1), 1223.
https://doi.org/10.1037/fam0000573
*Jensen, J. F., & Rauer, A. J. (2015a). Marriage work in older couples: Disclosure of marital
problems to spouses and friends over time. Journal of Family Psychology, 29(5), 732
743. https://doi.org/10.1037/fam0000099
*Jensen, J. F., & Rauer, A. J. (2015b). Young adult females' relationship work and its links to
romantic functioning and stability over time. Journal of Social and Personal
Relationships, 33(5), 687708. https://doi.org/10.1177/0265407515588221
*Johnson, M. D., & Anderson, J. R. (2013). The longitudinal association of marital
confidence, time spent together, and marital satisfaction. Family Process, 52(2), 244
256. https://doi.org/10.1111/j.1545-5300.2012.01417.x
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
76
Johnson, M. D., Lavner, J. A., Mund, M., Zemp, M., Stanley, S. M., Neyer, F. J., Impett, E.
A., Rhoades, G. K., Bodenmann, G., Weidmann, R., Bühler, J. L., Burriss, R. P.,
Wünsche, J., & Grob, A. (2021). Within-couple associations between communication
and relationship satisfaction over time. Personality and Social Psychology Bulletin, 1
16. https://doi.org/10.1177/01461672211016920
Johnson, M. D., & Neyer, F. J. (2019). (Eventual) stability and change across partnerships.
Journal of Family Psychology, 33(6), 711721. https://doi.org/10.1037/fam0000523
*Kanat-Maymon, Y., Sarid, O., Mor, Y., Mirsky, J., & Slonim-Nevo, V. (2016). Gaps
between immigrant spouses in host country language proficiency: Longitudinal effects
on marital satisfaction. The Journal of Psychology, 150(6), 793808.
https://doi.org/10.1080/00223980.2016.1196159
Karney, B. R., & Bradbury, T. N. (1995). Assessing longitudinal change in marriage: An
introduction to the analysis of growth curves. Journal of Marriage and the Family,
57(4), 10911108. https://doi.org/10.2307/353425
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),
100116. https://doi.org/10.1111/jomf.12635
Kazak, A. E. (2018). Editorial: Journal article reporting standards. American Psychologist,
73(1), 12. https://doi.org/10.1037/amp0000263
Kenny, D. A., & Zautra, A. (2001). Traitstate models for longitudinal data. In L. M. Collins
& A. G. Sayer (Eds.), New methods for the analysis of change (pp. 243263).
American Psychological Association. https://doi.org/https://doi.org/10.1037/10409-
008
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
77
*Kerkhof, P., Finkenauer, C., & Muusses, L. D. (2011). Relational consequences of
compulsive Internet use: A longitudinal study among newlyweds. Human
Communication Research, 37(2), 147173. https://doi.org/10.1111/j.1468-
2958.2010.01397.x
*Kluwer, E. S., & Johnson, M. D. (2007). Conflict frequency and relationship quality across
the transition to parenthood. Journal of Marriage and Family, 69(5), 495517.
https://doi.org/10.1111/j.1741-3737.2007.00434.x
*Kouros, C. D. (2011). Transactional relations between marital functioning, spouses'
depressive symptoms and children's adjustment [Doctoral dissertation, University of
Notre Dame]. University of Notre Dame, Hesburgh Libraries.
https://curate.nd.edu/usage/79407w64d4h
Kulu, H. (2014). Marriage duration and divorce: The seven-year itch or a lifelong itch?
Demography, 51(3), 881893. https://doi.org/10.1007/s13524-013-0278-1
Kurdek, L. A. (1998). The nature and predictors of the trajectory of change in marital quality
over the first 4 years of marriage for first-married husbands and wives. Journal of
Family Psychology, 12(4), 494510. https://doi.org/10.1037/0893-3200.12.4.494
Kurdek, L. A. (1999). The nature and predictors of the trajectory of change in marital quality
for husbands and wives over the first 10 years of marriage. Developmental
Psychology, 35(5), 12831296. https://doi.org/10.1037/0012-1649.35.5.1283
Kuster, F., & Orth, U. (2013). The long-term stability of self-esteem: Its time-dependent
decay and nonzero asymptote. Personality and Social Psychology Bulletin, 39(5),
677690. https://doi.org/10.1177/0146167213480189
*Lavner, J. A., & Bradbury, T. N. (2010). Patterns of change in marital satisfaction over the
newlywed years. Journal of Marriage and Family, 72(5), 11711187.
https://doi.org/10.1111/j.1741-3737.2010.00757.x
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
78
Lavner, J. A., Bradbury, T. N., & Karney, B. R. (2012). Incremental change or initial
differences? Testing two models of marital deterioration. Journal of Family
Psychology, 26(4), 606616. https://doi.org/10.1037/a0029052
*LeBaron, C. D. L., Miller, R. B., & Yorgason, J. B. (2014). A longitudinal examination of
women's perceptions of marital power and marital happiness in midlife marriages.
Journal of Couple & Relationship Therapy, 13(2), 93113.
https://doi.org/10.1080/15332691.2013.852492
Levinson, D. J., Darrow, C. M., Klein, E. B., Levinson, M. H., & McKee, B. (1976). Periods
in the adult development of men: Ages 18 to 45. The Counseling Psychologist, 6(1),
2125. https://doi.org/10.1177/001100007600600105
Lewis, M. (1997). Altering fate: Why the past does not predict the future. Guilford Press.
Lewis, M. (1999). On the development of personality. In L. A. Pervin & O. P. John (Eds.),
Handbook of personality: Theory and research (pp. 327346). Guilford Press.
Lewis, M. (2000a). Issues in the study of personality development. Psychological Inquiry,
12(2), 6783. https://doi.org/10.1207/S15327965PLI1202_02
Lewis, M. (2000b). Models of development. In D. Cervone & W. Mischel (Eds.), Advances in
personality science (pp. 153176). Guilford Press.
Li, T., & Chan, D. K. S. (2012). How anxious and avoidant attachment affect romantic
relationship quality differently: A meta‐analytic review. European Journal of Social
Psychology, 42(4), 406419. https://doi.org/https://doi.org/10.1002/ejsp.1842
*Li, X., Cao, H., Zhou, N., Ju, X., Lan, J., Zhu, Q., & Fang, X. (2018). Daily communication,
conflict resolution, and marital quality in Chinese marriage: A three-wave, cross-
lagged analysis. Journal of Family Psychology, 32(6), 733742.
https://doi.org/10.1037/fam0000430
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
79
*Lickenbrock, D. M., & Braungart-Rieker, J. M. (2015). Examining antecedents of infant
attachment security with mothers and fathers: An ecological systems perspective.
Infant Behavior and Development, 39, 173187.
https://doi.org/10.1016/j.infbeh.2015.03.003
Light, R. J., & Pillemer, D. B. (1984). Summing up: The science of reviewing research.
Harvard University Press.
*Lin, W.-F., Li, T.-S., & Chen, L. H. (2017). Sacrifice in a supportive marriage: Social
support as a moderator buffers the negative effects of sacrifice in marriage. Journal of
Happiness Studies: An Interdisciplinary Forum on Subjective Well-Being, 18(2), 575
589. https://doi.org/10.1007/s10902-016-9738-9
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Sage.
Locke, H. J., & Wallace, K. M. (1959, Vol.21). Short marital-adjustment and prediction tests:
Their reliability and validity. Marriage & Family Living, 21(3), 251255.
https://doi.org/10.2307/348022
*Logan, J. M., & Cobb, R. J. (2013). Trajectories of relationship satisfaction: Independent
contributions of capitalization and support perceptions. Personal Relationships, 20(2),
277293. https://doi.org/10.1111/j.1475-6811.2012.01408.x
Lorber, M. F., Erlanger, A. C. E., Heyman, R. E., & O'Leary, K. D. (2015). The honeymoon
effect: Does it exist and can it be predicted? Prevention Science, 16(4), 550559.
https://doi.org/10.1007/s11121-014-0480-4
Lucas, R. E., & Donnellan, M. B. (2007). How stable is happiness? Using the STARTS model
to estimate the stability of life satisfaction. Journal of Research in Personality, 41(5),
10911098. https://doi.org/https://doi.org/10.1016/j.jrp.2006.11.005
Lykken, D., & Tellegen, A. (1996). Happiness is a stochastic phenomenon [Peer Reviewed].
Psychological Science, 7(3), 186189. https://doi.org/10.1111/j.1467-
9280.1996.tb00355.x
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
80
McAdams, D. P. (1995). What do we know when we know a person? Journal of Personality,
63(3), 365396. https://doi.org/10.1111/j.1467-6494.1995.tb00500.x
McAdams, D. P. (2013). The psychological self as actor, agent, and author. Perspectives on
Psychological Science, 8, 272295. https://doi.org/10.1177/1745691612464657
McAdams, D. P. (2015). Three lines of personality development: A conceptual itinerary.
European Psychologist, 20(4), 252264. https://doi.org/10.1027/1016-9040/a000236
McGue, M., Bacon, S., & Lykken, D. T. (1993). Personality stability and change in early
adulthood: A behavioral genetic analysis. Developmental Psychology, 29(1), 96109.
https://doi.org/10.1037/0012-1649.29.1.96
McNulty, J. K. (2016). Highlighting the contextual nature of interpersonal relationships.
Advances in Experimental Social Psychology, 54, 247315.
https://doi.org/10.1016/bs.aesp.2016.02.003
McNulty, J. K., Meltzer, A. L., Neff, L. A., & Karney, B. (2021). How both partners’
individual differences, stress, and behavior predict change in relationship satisfaction:
Extending the VSA model. Proceedings of the National Academy of Sciences of the
United States of America, 118(27), e2101402118.
https://doi.org/10.1073/pnas.2101402118
Mehta, C. M., Arnett, J. J., Palmer, C. G., & Nelson, L. J. (2020). Established adulthood: A
new conception of ages 30 to 45. American Psychologist, 75(4), 431444.
https://doi.org/10.1037/amp0000600
*Meltzer, A. L., McNulty, J. K., Jackson, G. L., & Karney, B. R. (2014). Sex differences in
the implications of partner physical attractiveness for the trajectory of marital
satisfaction. Journal of Personality and Social Psychology, 106(3), 418428.
https://doi.org/10.1037/a0034424
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
81
*Meltzer, A. L., Novak, S. A., McNulty, J. K., Butler, E. A., & Karney, B. R. (2013). Marital
satisfaction predicts weight gain in early marriage. Health Psychology, 32(7), 824
827. https://doi.org/10.1037/a0031593
*Menéndez, S., Hidalgo, M. V., Jiménez, L., & Moreno, M. C. (2011). Father involvement
and marital relationship during transition to parenthood: Differences between dual and
single-earner families. The Spanish Journal of Psychology, 14(2), 639647.
https://doi.org/0.5209/rev_SJOP.2011.v14.n2.12
*Miller, P. J. E., Caughlin, J. P., & Huston, T. L. (2003). Trait expressiveness and marital
satisfaction: The role of idealization processes. Journal of Marriage and Family,
65(4), 978995. https://doi.org/10.1111/j.1741-3737.2003.00978.x
*Moen, D. A. (2012). Newlywed to established marriage: A longitudinal study of early risk
and protective factors that influence marital satisfaction (Publication No. 12-2011)
[Doctoral dissertation, Utah State University]. Utah State University, Merrill-Cazier
Library. https://digitalcommons.usu.edu/etd/1127
Moss, H. A., & Susman, E. J. (1980). Longitudinal study of personality development. In J. O.
G. Brim & J. Kagan (Eds.), Constancy and change in human development (pp. 530
595). Harvard University Press.
*Mund, M., Finn, C., Hagemeyer, B., Zimmermann, J., & Neyer, F. J. (2015). The dynamics
of self-esteem in partner relationships. European Journal of Personality, 29(2), 235
249. https://doi.org/10.1002/per.1984
Mund, M., & Neyer, F. J. (2014). Treating personality-relationship transactions with respect:
Narrow facets, advanced models, and extended time frames. Journal of Personality
and Social Psychology, 107(2), 352368. https://doi.org/10.1037/a0036719
Murray, S. L., Holmes, J. G., & Griffin, D. W. (2000). Self-esteem and the quest for felt
security: How perceived regard regulates attachment processes. Journal of Personality
and Social Psychology, 78(3), 478498. https://doi.org/0.1037//0022-3514.78.3.478
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
82
Murray, S. L., Holmes, J. G., & Griffin, D. W. (2000). Self-esteem and the quest for felt
security: How perceived regard regulates attachment processes. Journal of Personality
and Social Psychology, 78(3), 478-498. https://doi.org/10.1037//0022-3514.78.3.478
*Naud, C., Lussier, Y., Sabourin, S., Normandin, L., Clarkin, J. F., & Kernberg, O. F. (2013).
How attachment and excessive self-sacrificing depressive dynamics are related to
couple relationship satisfaction over time. Couple and Family Psychology: Research
and Practice, 2(1), 1433. https://doi.org/10.1037/a0031705
Neisser, U., Boodoo, G., Bouchard, T. J., Jr., Boykin, A. W., Brody, N., Ceci, S. J., Halpern,
D. F., Loehlin, J. C., Perloff, R., Sternberg, R. J., & Urbina, S. (1996). Intelligence:
Knowns and unknowns. American Psychologist, 51(2), 77101.
https://doi.org/10.1037/0003-066x.51.2.77
Neugarten, B. L. (1968). The awareness of middle age. In B. L. Neugarten (Ed.), Middle age
and aging (pp. 9398). University of Chicago Press.
Neyer, F. J., & Asendorpf, J. B. (2001). Personality-relationship transaction in young
adulthood. Journal of Personality and Social Psychology, 81(6), 11901204.
https://doi.org/10.1037/0022-3514.81.6.1190
Neyer, F. J., Mund, M., Zimmermann, J., & Wrzus, C. (2014). Personality-relationship
transactions revisited. Journal of Personality, 82(6), 539550.
https://doi.org/10.1111/jopy.12063
*Nguyen, T. P., Williamson, H. C., Karney, B. R., & Bradbury, T. N. (2017). Communication
moderates effects of residential mobility on relationship quality among ethnically
diverse couples. Journal of Family Psychology, 31(6), 753764.
https://doi.org/10.1037/fam0000324
*Niessen, C., Müller, T., Hommelhoff, S., & Westman, M. (2018). The impact of preventive
coping on business travelers' work and private life. Journal of Organizational
Behavior, 39(1), 113127. https://doi.org/10.1002/job.2215
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
83
Ogilvie, D. M., Rose, K. M., & Heppen, J. B. (2001). A comparison of personal project
motives in three age groups. Basic and Applied Social Psychology, 23(3), 207215.
https://doi.org/10.1207/s15324834basp2303_7
*Ogolsky, B. G., Surra, C. A., & Monk, J. K. (2016). Pathways of commitment to wed: The
development and dissolution of romantic relationships. Journal of Marriage and
Family, 78(2), 293310. https://doi.org/10.1111/jomf.12260
Orengo-Aguayo, R. E. (2015). Mexican American and other Hispanic couples’ relationship
dynamics: A review to inform interventions aimed at promoting healthy relationships.
Marriage & Family Review, 51(7), 633667.
https://doi.org/https://doi.org/10.1080/01494929.2015.1068253
Orth, U., Erol, R. Y., & Luciano, E. C. (2018). Development of self-esteem from age 4 to 94
years: A meta-analysis of longitudinal studies. Psychological Bulletin, 144(10), 1045
1080. https://doi.org/10.1037/bul0000161
*Orth, U., Maes, J., & Schmitt, M. (2015). Self-esteem development across the life span: A
longitudinal study with a large sample from Germany. Developmental Psychology,
51(2), 248259. https://doi.org/10.1037/a0038481
*Orth, U., Robins, R. W., & Widaman, K. F. (2012). Life-span development of self-esteem
and its effects on important life outcomes. Journal of Personality and Social
Psychology, 102(6), 12711288. https://doi.org/10.1037/a0025558
*Padilla, J., McHale, S. M., De Jesus, S. A. R., Updegraff, K. A., & Umana-Taylor, A. J.
(2018). Longitudinal course and correlates of parents' differential treatment of siblings
in Mexican-origin families. Family Process, 57(4), 979995.
https://doi.org/10.1111/famp.12328
*Paleari, F. G., Regalia, C., & Fincham, F. (2005). Marital quality, forgiveness, empathy, and
rumination: A longitudinal analysis. Personality and Social Psychology Bulletin,
31(3), 368378. https://doi.org/10.1177/0146167204271597
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
84
Pals, J. L. (1999). Identity consolidation in early adulthood: Relations with ego-resiliency, the
context of marriage, and personality change. Journal of Personality, 67(2), 295329.
https://doi.org/10.1111/1467-6494.00057
*Parfitt, Y., Ayers, S., Pike, A., Jessop, D., & Ford, E. (2014). A prospective study of the
parentbaby bond in men and women 15 months after birth. Journal of reproductive
and infant psychology, 32(5), 441456.
https://doi.org/10.1080/02646838.2014.956301
*Parise, M., Manzi, C., Donato, S., & Iafrate, R. (2017). Free to love? The role of intrusive
parenting for young adult children’s romantic relationship quality. Journal of
Prevention & Intervention in the Community, 45(3), 187201.
https://doi.org/10.1080/10852352.2016.1198127
*Peltz, J. S., Rogge, R. D., & Sturge-Apple, M. L. (2018). Transactions within the family:
Coparenting mediates associations between parents’ relationship satisfaction and the
parentchild relationship. Journal of Family Psychology, 32(5), 553564.
https://doi.org/10.1037/fam0000413
Proulx, C. M., Helms, H. M., & Buehler, C. (2007). Marital quality and personal well-being:
A meta-analysis. Journal of Marriage and Family, 69(3), 576593.
https://doi.org/10.1111/j.1741-3737.2007.00393.x
R Development Core Team. (2020). R: A language and environment for statistical computing.
R Foundation for Statistical Computing. https://www.R-project.org
Rammstedt, B., & Schupp, J. (2008). Only the congruent survive--Personality similarities in
couples. Personality and Individual Differences, 45(6), 533535.
https://doi.org/10.1016/j.paid.2008.06.007
Randall, A. K., & Bodenmann, G. (2009). The role of stress on close relationships and marital
satisfaction. Clinical Psychology Review, 29(2), 105115.
https://doi.org/10.1016/j.cpr.2008.10.004
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
85
*Reizer, A., Ein-Dor, T., & Shaver, P. (2014). The avoidance cocoon: Examining the
interplay between attachment and caregiving in predicting relationship satisfaction.
European Journal of Social Psychology, 44(7), 774786.
https://doi.org/10.1002/ejsp.2057
*Roberson, P. N., Fish, J. N., Olmstead, S. B., & Fincham, F. D. (2015). College adjustment,
relationship satisfaction, and conflict management: A cross-lag assessment of
developmental “spillover”. Emerging Adulthood, 3(4), 244-254.
https://doi.org/10.1177/2167696815570710
Roberts, B. W., & Caspi, A. (2003). The cumulative continuity model of personality
development: Striking a balance between continuity and change in personality traits
across the life course. In R. M. Staudinger & U. Lindenberger (Eds.), Understanding
human development: Lifespan psychology in exchange with other disciplines. Kluwer
Academic.
Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits
from childhood to old age: A quantitative review of longitudinal studies.
Psychological Bulletin, 126, 325. https://doi.org/10.1037/0033-2909.126.1.3
Roberts, B. W., & Nickel, L. B. (2021). Personality development across the life course: A
neo-socioanalytic perspective. In O. P. John & R. W. Robins (Eds.), Handbook of
personality: Theory and research (pp. 259-283). Guilford.
Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in
personality traits across the life course: A meta-analysis of longitudinal studies.
Psychological Bulletin, 132(1), 125. https://doi.org/10.1037/0033-2909.132.1.1
Roberts, B. W., & Wood, D. (2006). Personality development in the context of the neo-
socioanalytic model of personality. In D. K. Mroczek & T. D. Little (Eds.), Handbook
of personality development (pp. 1139). Erlbaum.
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
86
*Robins, R. W., Caspi, A., & Moffitt, T. E. (2002). It's not just who you're with, it's who you
are: Personality and relationship experiences across multiple relationships. Journal of
Personality, 70(6), 925964. https://doi.org/10.1111/1467-6494.05028
Robins, R. W., Fraley, R. C., Roberts, B. W., & Trzesniewski, K. H. (2001). A longitudinal
study of personality change in young adulthood. Journal of Personality, 69(4), 617
640. https://doi.org/https://doi.org/10.1111/1467-6494.694157
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem:
Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale.
Personality and Social Psychology Bulletin, 27(2), 151161.
https://doi.org/0146167201272002
Robles, T. F., Slatcher, R. B., Trombello, J. M., & McGinn, M. M. (2014). Marital quality and
health: A meta-analytic review. Psychological Bulletin, 140(1), 140187.
https://doi.org/10.1037/a0031859
Rothstein, H. R., Sutton, A. J., & Borenstein, M. (2005). Publication bias in meta-analysis:
Prevention, assessment and adjustments. John Wiley & Sons.
https://doi.org/10.1002/0470870168.ch1
*Ruffieux, M., Nussbeck, F. W., & Bodenmann, G. (2014). Long-term prediction of
relationship satisfaction and stability by stress, coping, communication, and well-
being. Journal of Divorce & Remarriage, 55(6), 485501.
https://doi.org/10.1080/10502556.2014.931767
Rusbult, C. E. (1980). Commitment and satisfaction in romantic associations: A test of the
investment model. Journal of Experimental Social Psychology, 16(2), 172186.
https://doi.org/10.1016/0022-1031(80)90007-4
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
87
Rusbult, C. E. (1983). A longitudinal test of the investment model: The development (and
deterioration) of satisfaction and commitment in heterosexual involvements. Journal
of Personality and Social Psychology, 45(1), 101117. https://doi.org/10.1037/0022-
3514.45.1.101
Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998). The Investment Model Scale:
Measuring commitment level, satisfaction level, quality of alternatives, and
investment size. Personal Relationships, 5(4), 357387.
https://doi.org/10.1111/j.1475-6811.1998.tb00177.x
*Sadikaj, G., Rappaport, L. M., Moskowitz, D., Zuroff, D. C., Koestner, R., & Powers, T.
(2015). Consequences of interpersonal spin on couple-relevant goal progress and
relationship satisfaction in romantic relationships. Journal of Personality and Social
Psychology, 109(4), 722737. https://doi.org/10.1037/pspp0000043
Sanchez, L., & Thomson, E. (1997). Becoming mothers and fathers: Parenthood, gender, and
the division of labor. Gender & Society, 11, 747772.
https://doi.org/10.1177/089124397011006003
Sbarra, D. A., Law, R. W., & Portley, R. M. (2011). Divorce and death: A meta-analysis and
research agenda for clinical, social, and health psychology. Perspectives on
Psychological Science, 6(5), 454474. https://doi.org/10.1177/1745691611414724
Scharfe, E., & Bartholomew, K. (1994). Reliability and stability of adult attachment patterns.
Personal Relationships, 1(1), 2343. https://doi.org/https://doi.org/10.1111/j.1475-
6811.1994.tb00053.x
Schimmack, U., & Oishi, S. (2005). The influence of chronically and temporarily accessible
information on life satisfaction judgments. Journal of Personality and Social
Psychology, 89(3), 395406. https://doi.org/https://doi.org/10.1037/0022-
3514.89.3.395
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
88
*Schober, P. S. (2012). Paternal child care and relationship quality: A longitudinal analysis of
reciprocal associations. Journal of Marriage and Family, 74(2), 281296.
https://doi.org/10.1111/j.1741-3737.2011.00955.x
Schröder, J., & Schmiedeberg, C. (2015). Effects of relationship duration, cohabitation, and
marriage on the frequency of intercourse in couples: Findings from German panel
data. Social Science Research, 52, 7282.
https://doi.org/10.1016/j.ssresearch.2015.01.009
Schuerger, J. M., Zarrella, K. L., & Hotz, A. S. (1989). Factors that influence the temporal
stability of personality by questionnaire. Journal of Personality and Social
Psychology, 56(5), 777783. https://doi.org/https://doi.org/10.1037/0022-
3514.56.5.777
Shulman, S., & Connolly, J. (2013). The challenge of romantic relationships in emerging
adulthood: Reconceptualization of the field. Emerging Adulthood, 1(1), 2739.
https://doi.org/10.1177/2167696812467330
*Sotskova, A., Woodin, E. M., & Gou, L. H. (2015). Hostility, flooding, and relationship
satisfaction: Predicting trajectories of psychological aggression across the transition to
parenthood. Aggressive Behavior, 41(2), 134148. https://doi.org/10.1002/ab.21570
*South, S. C., Boudreaux, M. J., & Oltmanns, T. F. (2020). The impact of personality
disorders on longitudinal change in relationship satisfaction in long-term married
couples. Journal of Personality Disorders, 34(4), 439-458.
https://doi.org/10.1521/pedi_2019_33_392
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of
marriage and similar dyads. Journal of Marriage and the Family, 38(1), 1528.
https://doi.org/10.2307/350547
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
89
Specht, J., Egloff, B., & Schmukle, S. C. (2011). Stability and change of personality across
the life course: The impact of age and major life events on mean-level and rank-order
stability of the Big Five. Journal of Personality and Social Psychology, 101(4), 862
882. https://doi.org/10.1037/a0024950
Sterne, J. A., & Egger, M. (2001). Funnel plots for detecting bias in meta-analysis: guidelines
on choice of axis. Journal of Clinical Epidemiology, 54(10), 10461055.
https://doi.org/10.1016/s0895-4356(01)00377-8
*Sullivan, T. J., Feinstein, B. A., Marshall, A. D., & Mustanski, B. (2017). Trauma exposure,
discrimination, and romantic relationship functioning: A longitudinal investigation
among LGB young adults. Psychology of Sexual Orientation and Gender Diversity,
4(4), 481490. https://doi.org/10.1037/sgd0000249
*Sun, X., McHale, S. M., Crouter, A. C., & Jones, D. E. (2017). Longitudinal links between
work experiences and marital satisfaction in african american dual-earner couples.
Journal of Family Psychology, 31(8), 10291039. https://doi.org/10.1037/fam0000381
Sutton, A. J. (2009). Publication bias. In H. Cooper, L. V. Hedges, & J. C. Valentine (Eds.),
The handbook of research synthesis and meta-analysis (pp. 435452). Russell Sage
Foundation.
Swann, W. B., & Read, S. J. (1981). Self-verification processes: How we sustain our self-
conceptions. Journal of Experimental Social Psychology, 17(4), 351372.
https://doi.org/10.1016/0022-1031(81)90043-3
*Szepsenwol, O., Mizrahi, M., & Birnbaum, G. E. (2015). Fatal suppression: The detrimental
effects of sexual and attachment deactivation within emerging romantic relationships.
Social Psychological and Personality Science, 6(5), 504512.
https://doi.org/0.1177/1948550615572636
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
90
Terracciano, A., Costa, P. T., Jr., & McCrae, R. R. (2006). Personality plasticity after age 30.
Personality and Social Psychology Bulletin, 32(8), 9991009.
https://doi.org/10.1177/0146167206288599
*Tombeau Cost, K., Jonas, W., Unternaehrer, E., Dudin, A., Szatmari, P., Gaudreau, H.,
Kennedy, J., Atkinson, L., Steiner, M., Meaney, M., & Fleming, A. (2018). Maternal
perceptions of paternal investment are associated with relationship satisfaction and
breastfeeding duration in humans. Journal of Family Psychology, 32(8), 10251035.
https://doi.org/10.1037/fam0000468
*Tremblay, S., & Pierce, T. (2011). Perceptions of fatherhood: Longitudinal reciprocal
associations within the couple. Canadian Journal of Behavioural Science, 43(2), 99
110. https://doi.org/10.1037/a0022635
Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. (2003). Stability of self-esteem
across the life span. Journal of Personality and Social Psychology, 84(1), 205220.
https://doi.org/https://doi.org/10.1037/0022-3514.84.1.205
*van den Troost, A., Vermulst, A. A., Gerris, J. R., & Matthijs, K. (2005). The Dutch marital
satisfaction and communication questionnaire: A validation study. Psychologica
Belgica, 45(3), 185206. https://doi.org/10.5334/pb-45-3-185
*van Scheppingen, M. A., Denissen, J. J. A., Chung, J. M., Tambs, K., & Bleidorn, W.
(2018). Self-esteem and relationship satisfaction during the transition to motherhood.
Journal of Personality and Social Psychology, 114(6), 973991.
https://doi.org/10.1037/pspp0000156
Vater, A., & Schröder-Abé, M. (2015). Explaining the link between personality and
relationship satisfaction: Emotion regulation and interpersonal behaviour in conflict
discussions. European Journal of Personality, 29(2), 201215.
https://doi.org/10.1002/per.1993
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
91
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of
Statistical Software, 36(3), 148. https://doi.org/10.18637/jss.v036.i03
*Volling, B. L., Oh, W., Gonzalez, R., Kuo, P. X., & Yu, T. (2015). Patterns of marital
relationship change across the transition from one child to two. Couple and Family
Psychology: Research and Practice, 4(3), 177197.
https://doi.org/10.1037/cfp0000046
Wagner, J., Ludtke, O., & Trautwein, U. (2016). Self-esteem is mostly stable across young
adulthood: Evidence from latent STARTS models. Journal of Personality, 84(4),
523−535. https://doi.org/10.1111/jopy.12178
*Weidmann, R., Ledermann, T., & Grob, A. (2017). The interdependence of personality and
satisfaction in couples. European Psychologist, 21(4), 284295.
https://doi.org/10.1027/1016-9040/a000261
*Weidmann, R., Schönbrodt, F. D., Ledermann, T., & Grob, A. (2017). Concurrent and
longitudinal dyadic polynomial regression analyses of Big Five traits and relationship
satisfaction: Does similarity matter? Journal of Research in Personality, 70, 615.
https://doi.org/10.1016/j.jrp.2017.04.003
*Whiteman, S. D., McHale, S. M., & Crouter, A. C. (2007). Longitudinal changes in marital
relationships: The role of offspring's pubertal development. Journal of Marriage and
Family, 69(4), 10051020. https://doi.org/10.1111/j.1741-3737.2007.00427.x
*Whitton, S. W., Rhoades, G. K., & Whisman, M. A. (2014). Fluctuation in relationship
quality over time and individual well-being: Main, mediated, and moderated effects.
Personality and Social Psychology Bulletin, 40(7), 858871.
https://doi.org/10.1177/0146167214528988
Woods, S. A., & Hampson, S. E. (2005). Measuring the Big Five with single items using a
bipolar response scale. European Journal of Personality, 19(5), 373390.
https://doi.org/10.1002/per.542
RANK-ORDER STABILITY OF RELATIONSHIP SATISFACTION
92
*Zee, K. S., & Weiss, D. (2019). High-quality relationships strengthen the benefits of a
younger subjective age across adulthood. Psychology and Aging, 34(3), 374388.
https://doi.org/10.1037/pag0000349
93
Table S1
Estimates of Rank-Order Stability of Relationship Satisfaction as a Function of Relationship Duration (Sensitivity Analyses for Using 2 and
6 Years as Estimates of the Average Difference Between Relationship Duration and Duration of Living Together, and Relationship Duration
and Marriage Duration, Respectively)
Relationship-duration
group
k
N
Weighted mean
effect size
95% CI
Q
Variances
σ12
σ22
Without time lag as covariate
05 years
37
5,049
.652
[ .578, .715]
414.9
.042
.052
510 years
114
22,982
.736
[ .663, .796]
6094.5
.073
.167
1020 years
98
19,499
.812
[ .742, .864]
8914.5
.243
.092
2046 years
63
8,071
.885
[ .809, .932]
17297.8
.335
.164
With time lag as covariate
05 years
37
5,049
.583
[ .456, .687]
362.5
.043
.047
510 years
114
22,982
.695
[ .609, .764]
5615.8
.094
.130
1020 years
98
19,499
.821
[ .761, .867]
5604.9
.194
.089
2046 years
63
8,071
.904
[ .835, .945]
9661.1
.363
.143
Note. Computations without control of time lag were made with multilevel random-effects models and computations with control of time lag
were made with multilevel mixed-effects meta-regression models. In the analyses with time lag as covariate, time lag was grand-mean
centered at 2.30 years. k = number of effect sizes; N = number of participants in each group. Weighted mean effect size = disattenuated
correlation coefficient, indicating the rank-order stability of relationship satisfaction between two assessments. CI = confidence interval; Q =
test statistic of the test for (residual) heterogeneity; σ12 = variance proportion attributable to the level of the grouping variable (i.e., between
samples); σ22 = variance proportion attributable to the level nested within the grouping variable (i.e., within samples). Values in bold are
significant at p < .05.
94
Table S2
Overview of Measures in the Meta-Analytic Data Set and Their Categorization Into Global Satisfaction Measures and Adjustment
Measures
Global satisfaction measure
Couples Satisfaction Index (CSI; Funk & Rogge, 2007)
Global Measure of Relationship Satisfaction (GMRS; Lawrance & Byers, 1998)
Intimate Relations Questionnaire (RRQ Braiker & Kelley, 1979)
Kansas Marital Satisfaction Scale (KMSS; Schumm et al., 1986)
Kansas Marital Satisfaction ScaleRevised (KMSS-R; Chung, 2004)
Marital Opinion Questionnaire (MOQ; Huston & Vangelisti 1991)
Relationship Assessment Scale (RAS; Hendrick, 1988)
Relationship satisfaction subscale from the Investment Model Scale (RSI; Rusbult, Martz, & Agnew, 1998)
Self-Report Relationship Satisfaction (SRS; Schmitt et al., 1997)
Quality of Marriage Index (QMI; Norton, 1983)
Adjustment measure
Couple Relationship Domains Questionnaire (CRDQ; Huston, McHale, & Crouter, 1986)
Dyadic Adjustment Scale (DAS; Spanier, 1976)
ENRICH Marital Satisfaction (ENRICH; Fowers & Olson, 1993)
Marital Relationship Inventory (MRI; Burgess, Locke, & Thomes, 1971)
Marital Adjustment Test (MAT; Locke & Wallace, 1959)
Marital Satisfaction Questionnaire for Older Persons (MSQO; Haynes et al., 1992)
Partnership Questionnaire (PFB; Partnerschaftsfragebogen; Hahlweg, 1996)
Perceived Relationship Quality Components Inventory (PQRC; Fletcher, Simpson, & Thomas, 2000), mean
Relationship Evaluation Questionnaire (REQ; Busby et al., 2001)
Relationship Satisfaction Scale (RSS; Gilford & Bengtson, 197)
95
Table S3
Descriptive Information on Time Lag
Group
k
M
SD
Range
Age
2030 years
157
1.88
2.56
0.1715.00
3040 years
128
1.95
3.28
0.1720.00
4050 years
66
4.24
3.92
0.5015.00
5071 years
45
2.08
1.84
0.489.00
Relationship duration
05 years
79
1.36
0.97
0.175.00
510 years
99
1.37
1.75
0.1715.00
1020 years
79
3.30
4.65
0.1720.00
2046 years
55
3.13
3.26
0.4815.00
Note. Time lag is given in years. k = number of effect sizes
96
Table S4
Zero-Order Correlations Between Age, Relationship Duration, Sample Characteristics, Methodological Characteristics,
and Effect Size Measure
Variable
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1. Age
2. Relationship duration
.97
3. Year of birth
.72
.75
4. Sample type
.12
.25
.30
5. Country
.12
.22
.28
.15
6. Ethnicity
.37
.42
.21
.25
.18
7. Female
.09
.01
.07
.09
.05
.02
8. Shared household
.47
.43
.41
.17
.42
.31
.02
9. Married
.26
.30
.38
.01
.33
.08
.00
.86
10. Children
.70
.64
.03
.06
.12
.36
.03
.28
.08
11. Transition
.16
.18
.26
.05
.10
.21
.00
.06
.11
.09
12. Post transition
.27
.24
.05
.12
.11
.06
.02
.08
.13
.15
.14
13. Baseline mean
.42
.29
.28
.25
.20
.10
.00
.38
.05
.51
.12
.20
14. Time lag
.12
.25
.48
.43
.07
.17
.03
.14
.12
.03
.15
.00
.08
15. Dyadic data
.10
.07
.22
.43
.21
.19
.16
.09
.03
.15
.12
.08
.00
.35
16. Measure (established)
.05
.12
.19
.68
.09
.04
.09
.13
.12
.09
.06
.02
.32
.38
.38
17. Measure (global)
.28
.28
.39
.16
.20
.22
.03
.04
.21
.19
.26
.02
.47
.01
.14
n.a.
18. Effect size
.32
.29
.10
.22
.28
.02
.06
.20
.22
.16
.05
.14
.35
.36
.18
.12
.37
97
Note. Effect size = disattenuated correlation coefficient, indicating the rank-order stability of relationship satisfaction between Time 1 and
Time 2. The following variables were dichotomous: sample type (1 = community or student sample, 0 = nationally representative sample),
country, (1 = USA, 0 = other), ethnicity (1 = White, 0 = other), transition (1 = relationship transition between Time 1 and Time 2, 0 = no
relationship transition between Time 1 and Time 2), post transition (1 = relationship transition shortly before Time 1 , 0 = no relationship
transition shortly before Time 1), dyadic data (1 = dyadic data, 0 = no dyadic data), the two contrasts for measure of relationship satisfaction (1
= established scale, 0 = ad-hoc measure; 1 = global satisfaction, 0 = adjustment measure). Baseline mean refers to mean of relationship
satisfaction at Time 1 in the metric of POMP scores, and time lag refers to the interval between Time 1 and Time 2. n.a. indicates that this
correlation could not be calculated because of the nested structure of the variables. Values in bold are significant at p < .05.
98
Note. Time lag refers to the time interval between two assessments. The vertical blue
line represents the average time lag (2.30 years).
Count
Time lag (years)