Development of Self-Esteem and Relationship Satisfaction in Couples: Two Longitudinal Studies

Abstract

We examined the effects of self-esteem development on the development of relationship satisfaction in 2 samples of couples. Study 1 used data from both partners of 885 couples assessed 5 times over 12 years, and Study 2 used data from both partners of 6,116 couples assessed 3 times over 15 years. The pattern of results was similar across the 2 studies. First, development of relationship satisfaction could be modeled as a couple-level process. Second, initial level of self-esteem of each partner predicted the initial level of the partners’ common relationship satisfaction, and change in self-esteem of each partner predicted change in the partners’ common relationship satisfaction. Third, these effects did not differ by gender and held when controlling for participants’ age, length of relationship, health, and employment status. Fourth, self-esteem similarity among partners did not influence the development of their relationship satisfaction. The findings suggest that the development of self-esteem in both partners of a couple contributes in a meaningful way to the development of the partners’ common satisfaction with their relationship.

Full text

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 1
Development of Self-Esteem and Relationship Satisfaction in Couples:
Two Longitudinal Studies
Ruth Yasemin Erol
University of Basel
Ulrich Orth
University of Bern
This article has been accepted for publication but has not been through the
copyediting, typesetting, pagination, and proofreading process. This article may
not exactly replicate the final version published in the journal. It is not the copy
of record. Please cite this article as follows:
Erol, R. Y., & Orth, U. (2014). Development of self-esteem and relationship
satisfaction in couples: Two longitudinal studies. Developmental Psychology, 50,
2291-2303. http://dx.doi.org/10.1037/a0037370
Author Note
Ruth Yasemin Erol, Department of Psychology, University of Basel; Ulrich Orth,
Department of Psychology, University of Bern.
This research was supported by Swiss National Science Foundation Grants PP00P1-
123370 to Ulrich Orth and CRSII1-147614 to Alexander Grob, Mike Martin, Franciska Krings,
and Ulrich Orth.
Correspondence concerning this article should be addressed to Ruth Yasemin Erol,
Department of Psychology, University of Basel, Missionsstrasse 62, 4055 Basel, Switzerland. E-
mail: ruthyasemin.erol@unibas.ch.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 2
Abstract
We examined the effects of self-esteem development on the development of relationship
satisfaction in two samples of couples. Study 1 used data from both partners of 885 couples
assessed 5 times over 12 years, and Study 2 used data from both partners of 6,116 couples
assessed 3 times over 15 years. The pattern of results was similar across the two studies. First,
development of relationship satisfaction could be modeled as a couple-level process. Second,
initial level of self-esteem of each partner predicted the initial level of the partners’ common
relationship satisfaction, and change in self-esteem of each partner predicted change in the
partners’ common relationship satisfaction. Third, these effects did not differ by gender and held
when controlling for participants’ age, length of relationship, health, and employment status.
Fourth, self-esteem similarity among partners did not influence the development of their
relationship satisfaction. The findings suggest that the development of self-esteem in both
partners of a couple contributes in a meaningful way to the development of the partners’ common
satisfaction with their relationship.
Keywords: self-esteem, relationship satisfaction, common fate, growth curve modeling

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 3
Development of Self-Esteem and Relationship Satisfaction in Couples:
Two Longitudinal Studies
Why are some couples more satisfied than others? Even though this question has been the
focus of many studies for decades, no straightforward answer has been forthcoming. Relationship
satisfaction is a construct influenced by multiple factors (Bradbury, Fincham, & Beach, 2000),
and emerging evidence suggests that self-esteem is one of these factors (e.g., Erol & Orth, 2013;
Murray, Holmes, & Griffin, 2000; Orth, Robins, & Widaman, 2012). However, the precise nature
of the link between self-esteem and relationship satisfaction is still unclear. For example, does
growing self-esteem of one of the partners in a relationship have a positive effect on both
partners’ satisfaction with the relationship? If so, do men and women differ in how their self-
esteem influences relationship satisfaction? Does similarity in self-esteem in the two partners
lead to a more satisfactory relationship? The present research aims to fill this gap by examining
the link between development of self-esteem and development of relationship satisfaction, using
data from both partners of couples who participated in two large longitudinal studies.
Development of Relationship Satisfaction
The construct of relationship satisfaction has been examined in a large body of research
(for a review, see Bradbury et al., 2000). According to Bradbury et al. (2000), a major reason for
the strong scientific interest in relationship satisfaction is its importance for personal and family
well-being. The evidence suggests that being married or in a close relationship is related to higher
levels of personal well-being (Diener, Suh, Lucas, & Smith, 1999) and, moreover, that
relationship satisfaction is one of the most predictive factors for relationship continuation
(Hendrick, 1988; Hendrick, Hendrick, & Adler, 1988). Nevertheless, some aspects of the
construct are still controversially discussed. For example, some researchers define relationship
satisfaction as a global evaluation of the marriage or close relationship (e.g., Fincham &

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 4
Bradbury, 1987), whereas others suggest that it is multidimensional, including evaluations of
positive and negative aspects of the relationship (e.g., Hendrick, 1988; Spanier, 1976).
Researchers favoring the global definition argue that it avoids overlap between relationship
satisfaction and its correlates such as the frequency of quarrels (Fincham & Bradbury, 1987). In
contrast, opponents of the global definition claim that evaluation of both positive and negative
aspects is required for comprehensively covering the construct (Fincham & Linfield, 1997). In
the studies included in the present research, two different measures were employed, which
corresponded to the two definitions described above. Whereas the measure used in Study
1captures positive and negative aspects of the relationship, in Study 2 the global satisfaction with
the relationship was assessed. Therefore, if the results of the two studies converge, this will
provide evidence for the robustness of the findings.
The trajectory of relationship satisfaction has been examined in several studies. Some
studies reported a U-shaped curve with a decrease during the first years of the relationship and an
increase in later years (e.g., Orbuch, House, Mero, & Webster, 1996), whereas others found a
continuous decrease across time (e.g., VanLaningham, Johnson, & Amato, 2001) or found that it
remained stable (e.g., Vaillant & Vaillant, 1993). In addition, longitudinal studies are available
that have examined the trajectories of relationship satisfaction over the first few years after
marriage. Karney and Bradbury (1997) found a declining trajectory for both spouses, whereas
Lavner and Bradbury (2010) reported that, although on average relationship satisfaction declined,
a significant proportion of spouses experienced relatively high levels of relationship satisfaction
and only small, if any, decline over time. Similarly, in a recent study Lavner, Bradbury, and
Karney (2012) identified three distinct trajectories for husbands and wives: the first trajectory
started high and remained stable, the second started at a moderate level and showed a small
decline, and the third started low and showed a substantial decline. These findings suggest that

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 5
decline in relationship satisfaction is limited to couples with a low level at the outset. Given that
couples who experience declines in relationship satisfaction are more likely to dissolve the
relationship in the future (Karney & Bradbury, 1997), it is important to identify the factors that
contribute to change in satisfaction.
According to the vulnerability-stress-adaptation model (Karney & Bradbury, 1995),
factors influencing relationship satisfaction can be grouped into three domains: personality traits
and experiences prior to marriage (i.e., vulnerabilities), stressful events and circumstances during
marriage (i.e., stressors), and emotions and communication skills during marriage (i.e.,
adaptation). This framework suggests that the satisfaction of a couple depends on how the couple
deals with certain stressors during marriage, which is affected by the couple’s prior traits and the
quality of their social interaction (Bradbury & Lavner, 2012). Several lines of research provide
evidence for the propositions of the model. For the vulnerability domain, research indicates that
low neuroticism (Karney & Bradbury, 1997) and fewer premarital doubts (Lavner & Bradbury,
2012) are related to relationship satisfaction. With regard to the stressors domain, stressful events
(Lavner & Bradbury, 2010) and higher psychological distress (Kurdek, 2005) predict lower
relationship satisfaction. Influential factors in the adaptation domain include better
communication (Meeks, Hendrick, & Hendrick, 1998), less verbal and physical aggression
(Lavner & Bradbury, 2010), and greater love, trust, and cohesion (Kurdek, 2005). The findings
reviewed above help to design interventions that account for a wide range of possible causes of
low relationship satisfaction. However, with regard to high-risk couples, research suggests that it
is particularly important to address additional factors such as low self-esteem to optimize the
effectiveness of interventions (Lavner & Bradbury, 2010). In the present research, we therefore
focus on the role of self-esteem in the development of relationship satisfaction.
The Link between Self-Esteem and Relationship Satisfaction

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 6
Previous research suggests that self-esteem is positively related to relationship satisfaction
(Fincham & Bradbury, 1993; Murray, Holmes, & Griffin, 1996; Shackelford, 2001).
Furthermore, in a 12-year longitudinal study with several waves of data, cross-lagged regression
analyses indicated that self-esteem prospectively predicted relationship satisfaction; in contrast,
relationship satisfaction did not predict self-esteem (Orth et al., 2012). Moreover, another
longitudinal study found that self-esteem at Time 1 was correlated with relationship satisfaction
at Time 2 in both men and women, whereas relationship satisfaction at Time 1 was correlated
with self-esteem at Time 2 in men only (Fincham & Bradbury, 1993). Also, Lavner, Bradbury,
and Karney (2012) reported that initial differences in self-esteem distinguished between spouses
with positive vs. negative trajectories of marital satisfaction. In sum, the available findings
suggest that self-esteem has a positive effect on relationship satisfaction.
Thus, on the one hand the hypothesis that relationship satisfaction influences self-esteem
has intuitive appeal, given that relationships—in particular attachment relationships—are
assumed to be an important source of self-esteem (Swann & Bosson, 2010). On the other hand,
the available longitudinal evidence, as reviewed above, supports the reverse direction of effects,
i.e., that self-esteem is a predictor of relationship satisfaction rather than vice versa (Fincham &
Bradbury, 1993; Orth et al., 2012). A possible reason for this pattern of results is that self-esteem
is a relatively stable characteristic of individuals throughout the adult life span, comparable to
broad personality traits such as the Big Five (Kuster & Orth, 2013; Trzesniewski, Donnellan, &
Robins, 2003). Moreover, the pattern of results is consistent with research suggesting that self-
esteem prospectively predicts satisfaction and success in other important life domains (besides
relationships) such as work and health (Kuster, Orth, & Meier, 2013; Orth et al., 2012). For these
reasons, in the present research we focus on the possible influence of self-esteem on relationship
satisfaction.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 7
When examining the link between self-esteem of relationship partners and the
development of their relationship satisfaction, it is important to account for the fact that each
partner’s self-esteem itself changes over time. Although self-esteem is, as mentioned above, a
relatively stable individual-differences construct, a growing body of evidence indicates that, on
average, people experience systematic change in self-esteem across the life span. Longitudinal
studies suggest that self-esteem increases throughout young and middle adulthood, peaking at
about age 60 years, and then declines into old age (Erol & Orth, 2011; Orth et al., 2012; Orth,
Trzesniewski, & Robins, 2010). Moreover, evidence suggests that throughout young and middle
adulthood self-esteem not only increases but also becomes more stable and less contingent from
daily events, or, to put it differently, self-esteem becomes better adjusted (Meier, Orth, Denissen,
& Kühnel, 2011). Possible causes of the normative changes in self-esteem across the life span
are, for example, corresponding changes in the perceived control over one’s life (Erol & Orth,
2011) and changes in the individual’s health and socioeconomic status (Orth et al., 2010).
Given these findings on systematic change in self-esteem during adulthood and given that
in the present research we followed partners across 12 years (Study 1) and 15 years (Study 2),
respectively, in the analyses we examined developmental changes in both self-esteem and
relationship satisfaction. Thus, the analyses will answer the developmental question of how
intraindividual change in self-esteem predicts intraindividual change in relationship satisfaction.
Moreover, the present research capitalizes on the dyadic design of the data sets used. More
specifically, the development of self-esteem among both partners of couples will be modeled as
separate processes, which may independently predict the development of relationship satisfaction
in these couples.
An important question is whether developmental processes in men’s and women’s self-
esteem differentially affect the development of the couple’s relationship satisfaction. On the one

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 8
hand, meta-analytic findings indicate that men tend to have higher self-esteem than women, albeit
the effect size is small (Kling, Hyde, Showers, & Buswell, 1999). Moreover, due to the influence
of gender stereotypes, it is possible that relationship harmony benefits from increasing self-
esteem among men, but that relationship harmony is threatened by increasing self-esteem among
women (admittedly, this hypothesis is purely speculative). In this case, the analyses would reveal
that gender moderates the effect of self-esteem development on development of relationship
satisfaction. On the other hand, cross-sectional research suggests that gender does not moderate
the effect of an individual’s self-esteem on his or her partner’s satisfaction with the relationship
(Erol & Orth, 2013). Moreover, the methodological literature on dyadic analyses advises that
researchers should be cautious in accepting hypotheses about gender differences before robust
evidence is available (Ackerman, Donnellan, & Kashy, 2011). Therefore, in this research we will
test whether any gender differences that emerge replicate across the two studies.
In the analyses, it will be important to control for age and length of relationship, because
these two variables could confound the relations between development of self-esteem and
development of relationship satisfaction. For example, research suggests that both self-esteem
and relationship satisfaction change in systematic ways across the adult life span (Orth et al.,
2010; VanLaningham et al., 2001). Given that the samples examined in this research are
heterogeneous with regard to age (see below for the description of the samples), and because, in
contrast, many relationship partners are relatively homogeneous with regard to age, the relations
between growth processes in self-esteem and relationship satisfaction could be artificially inflated
or deflated. Similarly, it is possible that self-esteem and relationship satisfaction change
systematically during the course of a relationship; again, because the samples examined are
heterogeneous with regard to length of relationship, the relations between growth in self-esteem
and relationship satisfaction could be biased, if length of relationship is not controlled for.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 9
In addition, it is possible that not only each individual’s self-esteem but also the degree of
the partners’ similarity in self-esteem contributes to the couple’s relationship satisfaction.
Overall, theory suggests that similarity rather than complementarity contributes to satisfaction
with relationships, because similarity increases relationship functioning and reduces conflicts in
daily life, and because similar partners share more similar emotional responses (Dyrenforth,
Kashy, Donnellan, & Lucas, 2010). Therefore, self-esteem similarity could facilitate empathy
among relationship partners and lead to more positive social interaction. On the other hand,
theory suggests that similarity in self-esteem could be also problematic, given the link between
self-esteem and dominance (Zeigler-Hill, 2010). Research indicates that dominant people are
more satisfied when interacting with submissive rather than dominant people, and, vice versa,
that submissive people are more satisfied when interacting with dominant rather than submissive
people (Dryer & Horowitz, 1997). For these reasons, self-esteem similarity might be beneficial,
detrimental, or the advantages and disadvantages of self-esteem similarity might cancel each
other out. Empirical research indicates that the self-esteem of relationship partners is correlated at
small to medium effect size (Erol & Orth, 2013; Shackelford, 2001). However, previous research
failed to test for possible effects of self-esteem similarity on relationship satisfaction. Although
research has tested for the effects of similarity in other personality characteristics (such as the Big
Five), similarity effects are still a topic of debate. Whereas some studies suggest that similarity in
personality is related to relationship satisfaction (e.g., Gonzaga, Campos, & Bradbury, 2007),
other studies do not support this hypothesis (e.g., Donnellan, Assad, Robins, & Conger, 2007) or
report only very small similarity effects (Dyrenforth et al., 2010).
The Present Research
Our first goal was to examine whether the development of self-esteem in both partners
predicted the development of a couple’s relationship satisfaction. No previous study has tested

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 10
for relations between the development of self-esteem and relationship satisfaction. Based on the
findings of previous studies, we expected that for each partner a positive trajectory of self-esteem
would predict a more positive trajectory of the couple’s relationship satisfaction. Because cross-
couple differences in age and length of relationship could confound the findings, in the analyses
we controlled for the effects of these variables. In addition, we also tested whether the results
hold when controlling for the effects of health and employment status. Our second goal was to
test whether self-esteem similarity in a couple further contributes to the development of
relationship satisfaction. Moreover, before conducting the main analyses, we tested whether
development in relationship satisfaction can be modeled as a couple-level process. Specifically,
we tested whether a common fate growth curve model (which combines the satisfaction ratings of
both partners to form a single latent construct; Ledermann & Kenny, 2012; Ledermann & Macho,
2014) provides a good fit to the data. To answer these questions, we used data from two
independent studies with large samples of couples and repeated assessments over long periods,
which enables us to draw more accurate and generalizable conclusions.
Study 1
Method
Study 1 used data from the Longitudinal Study of Generations (LSG; Bengtson, 2009),
which includes members of families that were randomly drawn from a subscriber list of about
840,000 members of a health maintenance organization in Southern California. Participants were
assessed in 1971, 1985, 1988, 1991, 1994, 1997, and 2000. In 1971 and 1985, the LSG did not
include the full self-esteem measure. Because growth curve analyses require that the same
measure is available at each wave (Preacher, Wichman, MacCallum, & Briggs, 2008), in the
present research we used data of the five waves from 1988 to 2000 (denoted as Time 1 to Time
5).

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 11
Participants. The sample consisted of 885 married or cohabiting couples (at Time 1, 98%
of the couples were married and 2% were cohabiting). For couples to be included in the present
analyses, each partner had to provide data on at least one of the constructs examined in this
research in at least one wave. Data were available for 690 couples at Time 1; 688 couples at Time
2; 758 couples at Time 3; 718 couples at Time 4; and 710 couples at Time 5. At Time 1, men’s
age ranged from 21 to 100 years (M = 54.3, SD = 17.1) and women’s age from 19 to 96 years (M
= 51.8, SD = 16.7). Of the male participants, 91% were White, 3% were Hispanic, 1% were
Black, and 5% were from other ethnicities. Of the female participants, 94% were White, 3% were
Hispanic, 1% were Black, and 2% were from other ethnicities. Because of the low frequencies of
ethnicities other than White, we did not examine ethnic differences in Study 1. The length of
relationship at Time 1 ranged from 1 to 62 years (M = 23.3, SD = 15.9). To investigate the
potential impact of attrition, we compared individuals who did and did not participate in the most
recent wave of data collection (Time 5) on self-esteem and relationship satisfaction at the four
preceding assessments. The only significant difference was that women who dropped out (versus
those who did not) reported slightly lower relationship satisfaction at Time 3 (d = -0.04);
differences at Times 1, 2, and 4 (and for men at all assessments) were nonsignificant. Thus,
nonrepresentativeness because of attrition was not a serious concern in this study.
Measures. Self-esteem was assessed with the 10-item Rosenberg Self-Esteem Scale
(RSE; Rosenberg, 1965), a commonly used and well-validated measure of self-esteem (cf.
Robins, Hendin, & Trzesniewski, 2001). Responses were measured using a 4-point scale (ranging
from 1 = strongly disagree to 4 = strongly agree). The mean averaged across waves was 3.42 (SD
= 0.43) and the alpha reliability for men and women ranged from .81 to .87 across assessments.
Relationship satisfaction was assessed with the 10-item Gilford-Bengtson Marital
Satisfaction Scale (Gilford & Bengtson, 1979). Participants reported how frequently they

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 12
experienced situations such as “laughing together” or “disagreeing about something important”
(reverse-scored). Responses were measured on a 5-point scale (ranging from 1 = hardly ever to 5
= almost always), with M = 3.92 (SD = 0.68) averaged across waves. The alpha reliability for
men and women ranged from .86 to .89 across assessments. The favorable reliability estimates
have been replicated in another study using the scale (Markides, Roberts-Jolly, Ray, Hoppe, &
Rudkin, 1999). To examine the convergent validity of the scale, we computed its correlation with
a single-item measure of relationship satisfaction, which was included in the LSG in 1988, 1997,
and 2000 (“Taking everything into consideration, how would you describe your marriage [or
living arrangement] at this point in time?”). The correlation ranged from .70 to .80 across
assessments, supporting the validity of the Gilford-Bengtson scale. For the analyses, we used the
Gilford-Bengtson scale because it is a multi-item measure (typically, multi-item measures have
better psychometric properties and, moreover, a reliability estimate was available only for the
multi-item measure but not the single-item measure) and because the single item was available at
only three waves.
Health was assessed with one item (i.e., “Compared to people your own age, how would
you rate your overall physical health at the present time?”). Responses were measured on a 4-
point scale (1= poor; 4 = excellent), with M = 3.08 (SD = 0.75) averaged across waves. For
employment status, we used a dichotomous variable (i.e., employed vs. unemployed).
Statistical analyses. The analyses were conducted using the Mplus 6.1 program (Muthén
& Muthén, 2010). To deal with missing values, we employed full-information maximum
likelihood (FIML) estimation to fit models directly to the raw data, which produces less biased
and more reliable results than conventional methods of dealing with missing data, such as
listwise or pairwise deletion (Schafer & Graham, 2002). Although full-information maximum
likelihood methods are based on the assumption that data are at least “missing at random” (MAR)

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 13
or “missing completely at random” (MCAR), even if data were “missing not at random”
(MNAR) full-information maximum likelihood estimates are less biased compared to
conventional methods of dealing with missing data (Allison, 2003). In both Study 1 and Study 2,
we tested the effect of missing data on the results, by using multi-group models that compared
couples who had no missing data (and stayed together across all waves) with couples who had
missing data at some waves (e.g., due to divorce or because one of the partners was deceased).
We computed two multi-group models: one in which the structural coefficients were estimated
freely across groups and one in which the structural coefficients were constrained to be equal
across groups. In both studies, the constraints did not significantly decrease model fit, suggesting
that including couples with missing data did not confound the results. Model fit was assessed by
the Tucker-Lewis index (TLI), the comparative fit index (CFI), and the root-mean-square error of
approximation (RMSEA). Hu and Bentler (1999) suggest that good fit is indicated by values
greater than or equal to .95 for TLI and CFI, and less than or equal to .06 for RMSEA. To test for
differences in model fit, we used the test of small difference in fit recommended by MacCallum,
Browne, and Cai (2006, Program C).
With regard to the RSE, a methodological problem in the LSG is that the labels of the
middle response categories were slightly different in 1988 and 1991 (i.e., 2 = somewhat disagree
and 3 = somewhat agree) vs. 1994, 1997, and 2000 (i.e., 2 = disagree and 3 = agree), which
likely precludes measurement invariance of the raw scores across waves. As measurement
invariance is essential for growth curve modeling (Edwards & Wirth, 2009; Widaman, Ferrer, &
Conger, 2010), we used confirmatory factor analysis with categorical indicators (Wirth &
Edwards, 2007) and equating by common items (Edwards & Wirth, 2009) to establish
measurement invariance and to compute RSE factor scores. In these analyses, we followed the
procedure described for the LSG by Orth et al. (2012). In short, the procedure establishes

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 14
measurement invariance because all observed scores, by using common items as anchors, are
mapped on the same latent scale. As in Orth et al. (2012), we equated the RSE across waves
using two items from the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff,
1977) that were conceptually related to self-esteem and that were available at each wave in
identical response format (i.e., “I felt that I was just as good as other people” and “I thought my
life had been a failure”).
Results and Discussion
Table 1 shows means and standard deviations of the measures across waves, separately
for men and women.
Common fate model of relationship satisfaction. Before examining the link between
self-esteem and relationship satisfaction, we tested whether relationship satisfaction can be
modeled as a common fate variable—that is, as a truly dyadic variable that captures variance
common to both partners of a couple (Ledermann & Kenny, 2012). The model included five
latent relationship satisfaction factors (i.e., one for each of the five waves) that loaded onto the
observed relationship satisfaction scores of both partners of the couple. The latent factors were
allowed to correlate with each other. Moreover, the residuals of the indicators for each of the
partners were correlated across waves to control for person-specific variance that was not
captured by the dyadic latent factors. At each wave, the unstandardized loadings for both partners
were set to 1.1 The intercepts of one of the partners (i.e., the male partner) were set to 0 to
identify the model. The fit of the model was good (TLI = .99, CFI = 1.00, RMSEA = .025). The
factor loadings were all significant (ps < .05) and the standardized values ranged from .78 to .83
for men, and from .71 to .74 for women. Then, we tested for strong factorial invariance of the
relationship satisfaction factors; strong factorial invariance is required when conducting growth
curve analyses, because it ensures that the means of latent factors are measured on the same scale

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 15
over time (Ledermann & Kenny, 2012; Widaman et al., 2010). In this model, the intercepts of
both partners were constrained to be equal within-person across waves. The fit of the model was
good (TLI = .99, CFI = 1.00, RMSEA = .023) and was not significantly worse than the fit of the
model without constrained intercepts. Consequently, we used these constraints in subsequent
analyses.
Next, we estimated a common fate growth curve model (Ledermann & Macho, 2014)
based on the latent relationship satisfaction factors, as measured in the model described above
(Figure 1). The loadings of the intercept factor were set to 1 and the loadings of the slope factor
were set to 0, 0.25, 0.5, 0.75, and 1. The fit of the model was good (TLI = .98, CFI = .99,
RMSEA = .041).2 We also tested whether a quadratic growth curve model provided for a better
fit to the data compared to the linear model (by including an additional slope factor, for which
quadratic values of the loadings were used); however, the quadratic model did not converge
properly. When we fixed the variance of the quadratic term to zero, the quadratic model
converged but did not fit significantly better than the linear model (for the test of small difference
in fit, dflinear = 28, dfquadratic = 27, critical Δχ2 = 45.6, observed Δχ2 = 1.3, ns). Consequently, we
continued to use the linear model.
Main analyses. Then, we examined the effects of men’s and women’s self-esteem
development on the development of relationship satisfaction. To model the growth processes of
men’s and women’s self-esteem, we used a bivariate growth curve model (McArdle, 1988),
which allows examining two parallel growth processes simultaneously (for an illustration, see the
upper half of Figure 2). Specifically, the model included two sets of growth factors (i.e., one
intercept and one slope factor for each of the two partners), which explained the corresponding
self-esteem scores (i.e., men’s and women’s self-esteem). All growth factors were allowed to
correlate with each other. The loadings of the intercept factors were set to 1 and the loadings of

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 16
the slope factors were set to 0, 0.25, 0.5, 0.75, and 1. The residual variances of self-esteem
indicators were allowed to covary between men and women within waves to account for
occasion-specific systematic variance. The fit of the model was good (TLI = .97, CFI = .97,
RMSEA = .067). We also tested whether a quadratic growth curve model provided for a better fit
to the data than the linear model; however, the quadratic model did not fit significantly better
than the linear model (for the test of small difference in fit, dflinear = 36, dfquadratic = 34, critical Δχ2
= 63.9, observed Δχ2 = 0.9, ns).
Next, we examined the complete model, including self-esteem and relationship
satisfaction (Figure 2). To model development in relationship satisfaction, we used the common
fate growth curve model described above. The model included six growth factors: two intercept
factors for self-esteem (i.e., men’s and women’s intercepts), two slope factors for self-esteem
(i.e., men’s and women’s slope), and one intercept and one slope factor for relationship
satisfaction. The growth factors (i.e., intercept and slope) for relationship satisfaction were
regressed on the growth factors for self-esteem. In the analyses, we controlled for the effects of
age and length of relationship, by regressing the growth factors for self-esteem and relationship
satisfaction on the control variables (we used the mean age across both partners, given that the
age of female and male partners was correlated at .98 averaged across waves).In both Study 1 and
Study 2, the results were virtually unaltered when the control variables were not included in the
models.
First, we estimated the effects of self-esteem on relationship satisfaction freely for men
and women. Overall, the fit of the model was good (TLI = .94, CFI = .96, RMSEA = .056).
Second, we constrained the effects to be equal for men and women. The constraints did not
significantly diminish model fit, suggesting that men and women did not differ in the effects of
self-esteem development on relationship satisfaction. Consequently, in the remainder of the

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 17
analyses, we used the model with cross-gender equality constraints on structural effects. Figure
3A shows the standardized estimates of the effects. The self-esteem intercept factors positively
predicted the intercept factor of relationship satisfaction with about medium effect sizes (for both
men and women, β = .24, p < .05). Furthermore, the self-esteem slope factors positively predicted
the slope factor of relationship satisfaction with about medium effect sizes (for men, β = .24, p <
.05; for women, β = .23, p < .05).3 These findings suggest that higher initial levels of self-esteem
predict a higher initial level of relationship satisfaction, and, more importantly, that positive
changes in self-esteem predict positive change in relationship satisfaction. It is noteworthy that
the effects of self-esteem intercepts were controlled for the effects of self-esteem slopes and, vice
versa, the effects of self-esteem slopes were controlled for the effects of self-esteem intercepts,
which strengthens the conclusions that can be drawn from the findings.
In addition, we tested whether age moderated the effects of self-esteem development on
development of relationship satisfaction, using multi-group models (i.e., comparing participants
who were, at Time 1, 40 years or younger vs. participants who were older than 40 years).
Constraining the effects to be equal across groups did not significantly decrease model fit,
suggesting that age did not moderate the effects. Similarly, we tested whether length of
relationship satisfaction moderated the effects (comparing couples who were, at Time 1, less than
10 years together vs. couples who were more than 10 years together). The results suggested that
length of relationship did not moderate the effects. Moreover, we tested whether the results hold
when the participants’ health and employment status were included as time-varying covariates in
the model (i.e., using the wave-specific scores on these covariates); however, the results were
essentially unaltered.
Finally, we tested whether similarity in self-esteem between partners has an effect on the
development of relationship satisfaction. Similarity was operationalized as the difference between

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 18
men’s and women’s self-esteem (thus, a difference of 0 indicates perfect similarity). On average,
self-esteem did not differ between men and women (M = 0.01, SD = 1.42, averaged across waves;
the positive mean indicates that men scored higher than women, although the difference was
nonsignificant, d = 0.01). The wave-specific difference scores were used as indicators for a
growth curve model, allowing us to examine whether the intercept and slope in self-esteem
similarity predicted the growth factors of relationship satisfaction. The fit of this model was good
(TLI = .96, CFI = .97, RMSEA = .044). However, neither the intercept nor the slope for self-
esteem similarity had a significant effect on the growth factors for relationship satisfaction. Thus,
the findings suggest that similarity in self-esteem does not influence the development of
relationship satisfaction. To further test for effects of self-esteem similarity, we also used another
indicator of similarity, by computing the absolute values of the difference between men’s and
women’s self-esteem (i.e., discrepancy scores). When we examined a growth curve model using
the wave-specific discrepancy scores, the fit of this model was good (TLI =.97, CFI = .97,
RMSEA = .039). However, neither the intercept nor the slope factor of discrepancy scores
significantly predicted the growth factors for relationship satisfaction.
Study 2
Method
Study 2 used data from the National Survey of Families and Households (NSFH; Sweet &
Bumpass, 2002), a nationally representative study providing a broad range of information on
American family life. Participants were first assessed in 1987 or 1988 (Time 1), and then approx.
6 years (Time 2) and approx. 15 years (Time 3) later.
Participants. The sample consisted of 6,116 married or cohabiting couples (no
information on the percentage of cohabiting couples is available). For couples to be included in
the present analyses, each partner had to provide data on at least one of the constructs examined

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 19
in this research in at least one wave. Data were available for 6,115 couples at Time 1; 3,956
couples at Time 2; and 1,836 couples at Time 3 (the assessment at Time 3 included only a small
number of couples due to budget constraints; Sweet & Bumpass, 2002). At Time 1, men’s age
ranged from 17 and 90 years (M = 43.0, SD = 15.7) and women’s age from 16 to 90 years (M =
40.3, SD = 15.2). Of both male and female participants, 81% were White, 11% were Black, 7%
were Hispanic, and 1% were from other ethnicities. The length of relationship at Time 1 ranged
from 0.8 to 68 years (M = 15.9, SD = 15.1). To investigate the potential impact of attrition, we
compared individuals who did and did not participate in the most recent wave of data collection
(Time 3), on self-esteem and relationship satisfaction at the preceding assessments. At Time 1,
men who dropped out (versus those who did not) reported slightly lower self-esteem (Ms = 4.09
vs. 4.12; d = -0.04), and women who dropped out reported slightly higher relationship
satisfaction (Ms = 6.06 vs. 6.05; d = 0.01). Given that these two differences were very small and,
moreover, all other differences were nonsignificant, we concluded that nonrepresentativeness
because of attrition was not a concern in this study.
Measures. Self-esteem was assessed with three items from the RSE. The items were “I
feel that I am a person of worth, at least on an equal plane with others,” “On the whole, I am
satisfied with myself,” “I am able to do things as well as other people.” Responses were
measured using a 5-point scale (1= strongly agree; 5 = strongly disagree), with M = 4.08 (SD =
0.60) averaged across waves. The alpha reliability for men and women ranged from .60 to .68
across assessments.
Relationship satisfaction was assessed with one item (i.e., “Taking things all together,
how would you describe your relationship?”). Responses were measured on a 7-point scale (1 =
very unhappy; 7 = very happy), with M = 6.06 (SD = 1.22) averaged across waves.
Health was assessed with one item (i.e., “Compared with other people your own age, how

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 20
would you describe your health?”). Responses were measured on a 5-point scale (1= very poor; 5
= excellent), with M = 4.01 (SD = 0.84) averaged across waves. For employment status, we used
a dichotomous variable (i.e., employed vs. unemployed).
Statistical analyses. The analyses were conducted using the Mplus 6.1 program. Model
fit was assessed using the same procedures as in Study 1.
Results and Discussion
Table 1 shows means and standard deviations of the measures across waves, separately
for men and women. The models and statistical procedures were identical to Study 1, except that
the Study 2 models were based on only three instead of five waves of data.
Common fate model for relationship satisfaction. As in Study 1, we tested whether
relationship satisfaction can be modeled as a common fate variable. The model included three
latent relationship satisfaction factors (i.e., one for each wave) that loaded onto the observed
relationship satisfaction scores of both partners of the couple. The fit of the model was good (TLI
= 1.00, CFI = 1.00, RMSEA = .006). The factor loadings were all significant (ps < .05) and
standardized values ranged from .62 to .67 for men, and from .58 to .61 for women. Then, we
tested for strong factorial invariance of relationship satisfaction, constraining the intercepts of
both partners to be equal within-person across waves. The fit of the model was good (TLI = 1.00,
CFI = 1.00, RMSEA = .000) and not significantly worse than the fit of the model without
constrained intercepts. Consequently, we used these constraints in subsequent analyses.
Next, we estimated a common fate growth curve model based on the latent relationship
factors, as measured in the model described above. The loadings of the intercept factor were set
to 1 and the loadings of the slope factor were set to 0, 0.4, and 1 (corresponding to the timing of
assessments). The fit of the model was acceptable (TLI = .89, CFI = .94, RMSEA = .058).4
Main analyses. We then examined the effects of men’s and women’s self-esteem

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 21
development on the development of relationship satisfaction. As in Study 1, we used a bivariate
growth curve model to capture change in men’s and women’s self-esteem. The loadings of the
intercept factors were set to 1 and the loadings of the slope factors were set to 0, 0.4, and 1. The
fit of the model was acceptable (TLI = .95, CFI = .99, RMSEA = .034).
Next, we examined the complete model including the growth factors for relationship
satisfaction regressed on the growth factors for self-esteem. First, we estimated the effects of self-
esteem on relationship satisfaction freely for men and women. The fit of the model was good
(TLI = .90, CFI = .94, RMSEA = .032). Second, we constrained the effects to be equal for men
and women. The constraints did not significantly decrease model fit, suggesting that the effects
did not differ between men and women. Consequently, in the remainder of the analyses, we used
the model with cross-gender equality constraints. Figure 3B shows the standardized estimates of
the structural effects. The self-esteem intercept factors positively predicted the intercept factor of
relationship satisfaction with about medium effect sizes (for men, β = .33, p < .05; for women, β
= .34, p < .05). Furthermore, the self-esteem slope factors positively predicted the slope factor of
relationship satisfaction with about medium effect sizes (for men, β = .26, p < .05; for women, β
= .22, p < .05). As in Study 1, these findings suggest that higher initial levels of self-esteem
predict a higher initial level of relationship satisfaction, and, more importantly, that positive
changes in self-esteem predict positive change in relationship satisfaction. In Study 2, two
additional effects were significant given that the self-esteem intercepts negatively predicted the
slope factor of relationship satisfaction (for both men and women, β = -.17, p < .05). However,
the interpretation of effects of intercepts on slopes is often problematic. As in the present case,
the growth process is often captured as it unfolds and, consequently, the location of the intercept
is arbitrary and the intercept cannot be interpreted as the starting point of the process (Grimm,
2007). However, the size, and even the direction, of the relation between intercept and slope

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 22
directly depend on the location of the intercept (Grimm, 2007). For this reason, and because the
effect of the intercept on the slope emerged in only one of the two studies included in the present
research, we did not draw substantive conclusions from these effects.
As in Study 1, we tested whether age moderated the effects of self-esteem development
on development of relationship satisfaction, using multi-group models (i.e., comparing
participants who were, at Time 1, 40 years or younger vs. participants who were older than 40
years). The results suggested that age did not moderate the effects. Similarly, we tested whether
length of relationship satisfaction moderated the effects (comparing couples who were, at Time 1,
less than 10 years together vs. couples who were more than 10 years together). Again, the results
suggested that length of relationship did not moderate the effects.5 We also tested whether
ethnicity moderated the effects of self-esteem development on development of relationship
satisfaction, using multi-group models (i.e., we compared 4,710 couples of which both partners
were White, 649 couples of which both partners were Black, and 472 couples of which both
partners were Hispanic). Constraining the effects to be equal across groups did not significantly
decrease model fit, suggesting that ethnicity did not moderate the effects.6 Moreover, as in Study
1, we tested whether the results hold when health and employment status were included as time-
varying covariates; however, the results were essentially unaltered.
Finally, we tested whether similarity in self-esteem between partners—as indicated by
difference scores—had an effect on the development of relationship satisfaction (TLI = .87, CFI
= .92, RMSEA = .045). On average, men reported slightly higher self-esteem than women (M =
0.08, SD = 1.31, averaged across waves; p < .05, d = 0.06). However, neither the intercept nor the
slope factor for self-esteem similarity had a significant effect on the growth factors for
relationship satisfaction. Thus, as in Study 1, the findings suggest that similarity in self-esteem
does not influence the development of relationship satisfaction. Again, we further tested for

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 23
effects of self-esteem similarity by using the absolute values of the difference between men’s and
women’s self-esteem (i.e., discrepancy scores) as indicators for a growth curve model (TLI = .89,
CFI = .93, RMSEA = .038). However, the intercept and slope factor of discrepancy scores did not
significantly predict the growth factors for relationship satisfaction (except for the effect of the
intercept of self-esteem discrepancy on the intercept of relationship satisfaction, which was -.26,
p < .05).
General Discussion
In the present research, we examined the effects of self-esteem development on the
development of relationship satisfaction in two samples of couples. Study 1 used data from both
partners of 885 couples assessed 5 times over 12 years, and Study 2 used data from both partners
of 6,116 couples assessed 3 times over 15 years. The pattern of results was similar across the two
studies. First, development of relationship satisfaction could be modeled as a couple-level
process. Second, initial level of self-esteem of each partner predicted the initial level of the
partners’ common relationship satisfaction, and change in self-esteem of each partner predicted
change in the partners’ common relationship satisfaction. Third, these effects did not differ by
gender and held controlling for participants’ age, length of relationship, health, and employment
status. Fourth, self-esteem similarity among partners did not influence the development of their
relationship satisfaction. In the following, we discuss these findings in more detail.
The results suggest that relationship satisfaction can be modeled as a truly dyadic
variable. Specifically, in our analyses we combined the satisfaction ratings of both partners of a
couple to form a single—i.e., couple-level—latent construct, using the common fate model
(Ledermann & Kenny, 2012). Then, we assessed the developmental trajectory of relationship
satisfaction using a common fate growth curve model (Ledermann & Macho, 2014). For both
data sets examined in this research, the results supported the usefulness of the model: (a) the fit of

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 24
the model was good and (b) the loadings of the couple-level factors on the individual-level
indicators were large. The findings support that—at least in the data sets examined in this
research—men’s and women’s reports of relationship satisfaction are highly related and can be
used to model a dyadic variable.
In both studies, the analyses showed that the developmental trajectories of self-esteem of
each partner predicted the developmental trajectory of the couple’s common relationship
satisfaction. Although previous studies had suggested that self-esteem prospectively predicts the
individual’s relationship satisfaction and, moreover, although cross-sectional research indicates
that self-esteem is linked in a dyadic way to relationship satisfaction (Erol & Orth, 2013; Murray
et al., 2000), no previous study has tested whether change in self-esteem is related to change in
the couple’s satisfaction. By using growth curve modeling of longitudinal data, the present
research suggests that the developmental trajectories of self-esteem and relationship satisfaction
(i.e., in particular their slope factors) are meaningfully related within couples. When we regressed
the growth factors for relationship satisfaction on growth factors for self-esteem, the effects were
of about medium size. Importantly, given that we simultaneously examined the effects of both
partners’ self-esteem, the cumulative effects of the partners’ self-esteem on relationship
satisfaction are twice as large as the individual effects of self-esteem.
In both studies the effect sizes did not differ significantly between male and female
partners. Thus, although men and women might differ in their typical social roles (Eagly &
Wood, 1999) and although it might be plausible that change (e.g., improvement) in men’s and
women’s self-esteem differentially affects the couple’s harmony, the present findings suggest that
the structural relations between self-esteem and relationship satisfaction hold across gender,
corresponding to results from cross-sectional dyadic analyses (Erol & Orth, 2013). Moreover, the
findings are in line with longitudinal results reported by Kurdek (2005) who examined whether

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 25
there is a “his or her version of marital processes” (p. 82). Kurdek (2005) concluded that
although men and women differ in average levels on variables such as marriage-specific
appraisals and satisfaction with social support by the partner, the effects of these variables on
relationship satisfaction were similar for men and women (Kurdek, 2005).
The present research rules out an additional, or alternative, hypothesis on the relation
between development in self-esteem and relationship satisfaction, specifically the hypothesis that
the partners’ similarity in self-esteem contributes to the development of relationship satisfaction.
In both studies, level and change in self-esteem similarity was not significantly related to the
developmental trajectory of relationship satisfaction. Thus, the findings on self-esteem similarity
are largely consistent with studies on similarity in other personality characteristics, many of
which suggested that similarity does not influence, or has negligible effects on, the couple’s
relationship satisfaction (Donnellan, Trzesniewski, Conger, & Conger, 2007; Dyrenforth et al.,
2010). As outlined in the Introduction, there are competing theoretical perspectives suggesting
that self-esteem similarity could be either beneficial for relationships (because partners share
more similar emotional responses) or detrimental (because high self-esteem is linked to a
dominant interpersonal style and because dominant people are more satisfied when interacting
with submissive rather than dominant partners). Consequently, it is possible that we did not find
an effect of self-esteem similarity because its advantages and disadvantages cancel each other
out. Another perspective is provided by Murray, Holmes, Bellavia, Griffin, and Dolderman
(2002) who argue that couples perceive similarity even when in reality the similarity is small, but
that the mere perception of similarity is sufficient to feel better understood and, consequently, to
be more satisfied with the relationship. Therefore, perceiving to have similar self-esteem might
be more important than actual similarity for a satisfying relationship.
A limitation of this research is that the study designs, although longitudinal, do not allow

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 26
for causal conclusions about the effects of self-esteem on the development of relationship
satisfaction. As in all non-experimental studies, effects may be caused by third variables that
were not included in the analyses. However, given that we controlled for several relevant
variables that could have confounded the self-esteem effect (i.e., age, length of relationship,
health, and employment status), the present research strengthens the case for the unique effect of
self-esteem on relationship satisfaction.
Another limitation is that both data sets examined in this research included participants
from a single country, i.e., the United States. Although the findings are likely to generalize to
other Western countries, future research should replicate the analyses using samples from other
cultural contexts. For example, research suggests that culture influences communication
behaviors in relationships and, moreover, influences whether these behaviors are linked to
relationship satisfaction (Williamson et al., 2012). However, in Study 2 it was possible to test for
ethnic differences and the results suggested that the findings did not differ between White, Black,
and Hispanic participants. Thus, although the participants of Study 2 were American, the fact that
the pattern of results replicated across different ethnic groups is a first step toward establishing
the cross-cultural generalizability of the findings. Nevertheless, future research is needed to
examine differences that may exist within Hispanic (e.g., Puerto Rican, Mexican, etc.) and Black
(e.g., African American, Caribbean, etc.) populations to better understand how generalizable the
present findings are.
In Study 2, self-esteem was measured using only three items of the RSE. Therefore, as
expected, the reliability of this measure was lower compared to the 10-item RSE (Robins et al.,
2001). However, in the LSG data set used in Study 1, the three-item RSE (as used in Study 2)
correlated at .80 with the 10-item RSE, supporting the validity of the three-item RSE. In addition,
based on an item response analysis, Gray-Little, Williams, and Hancock (1997) suggested that

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 27
the 10-item RSE “could be shortened without compromising the measurement of global self-
esteem” (p. 450).
One strength of the present research is the convergence of findings across Studies 1 and 2,
which helps allay some methodological concerns. For example, a limitation of Study 2 was that
self-esteem was assessed with only three items (as mentioned above) and relationship satisfaction
with a single item; however, this limitation was addressed in Study 1, in which the complete 10-
item RSE and a 10-item scale for relationship satisfaction was used. A limitation of Study 1 was
that the sample was not representative of the U.S. population; however, this limitation was
addressed in Study 2, in which data from a national probability sample were used, increasing
confidence in the generalizability of our findings. Additional strengths of the present research are
that both studies included very large samples of couples and that the samples were heterogeneous
with regard to the participants’ age (ranging from 16 to 96 years across studies) and length of
relationship (ranging from 0.8 to 68 years across studies). Also, in both studies participants were
repeatedly assessed across long periods (12 years in Study 1 and 15 years in Study 2).
As yet there is scant evidence with regard to the mechanisms that account for the link
between self-esteem and relationship satisfaction. The study by Murray et al. (2000) suggested
that felt security about the partner’s love might mediate the effect of self-esteem on relationship
satisfaction. More specifically, individuals with low self-esteem tend to think that their partners
see them as negatively as they see themselves; therefore, in order to avoid disappointment, they
might defensively find faults in their partner or distance themselves, which in turn diminishes
both partners’ satisfaction with the relationship. Furthermore, a study by Erol and Orth (2013)
found that secure romantic attachment between the partners mediated large portions of actor and
partner effects of self-esteem on relationship satisfaction. Secure romantic attachment to the
current partner was assessed as having low levels of attachment-related anxiety and avoidance,

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 28
both of which contributed independently to the mediation effect. Possible theoretical explanations
for the mediating effects are that, for example, attachment-related anxiety is linked to problematic
interpersonal behavior such as excessive reassurance seeking (Shaver, Schachner, & Mikulincer,
2005), whereas attachment-related avoidance is linked to paying less attention to emotional
information provided by the partner (Fraley, Garner, & Shaver, 2000). Therefore, in future
research, it would be of interest to test whether growth in secure attachment between the partners
mediates the effects of growth in self-esteem on growth in relationship satisfaction. Moreover, it
is possible that the degree of attachment security is related to whether couples are married or
cohabiting; future research should therefore explore the role of marital status in the link between
self-esteem and relationship satisfaction.
In conclusion, the present research suggests that the development of self-esteem among
relationship partners influences the development of their common relationship satisfaction,
regardless of the partners’ age, the length of their relationship, and the similarity of their self-
esteem. The findings contribute to the emerging understanding of the link between self-esteem
and satisfaction in romantic relationships. Furthermore, the findings suggest that interventions
aimed at improving self-esteem may be an effective way to enhance a couple’s relationship
satisfaction and, consequently, increase the stability of their relationship. Given that the findings
indicate that each partner’s self-esteem has a unique effect on the couple’s satisfaction, not only
interventions that target both partners’ self-esteem (e.g., in the context of relationship
counseling), but also interventions that focus on one of the partners might contribute to the
quality of the relationship.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 29
References
Ackerman, R. A., Donnellan, M. B., & Kashy, D. A. (2011). Working with dyadic data in studies
of emerging adulthood: Specific recommendations, general advice, and practical tips. In
F. D. Fincham & M. Cui (Eds.), Romantic relationships in emerging adulthood. New
York, NY: Cambridge University Press.
Allison, P. D. (2003). Missing data techniques for structural equation modeling. Journal of
Abnormal Psychology, 112, 545-557.
Bengtson, V. L. (2009). Longitudinal Study of Generations, 1971, 1985, 1988, 1991, 1994, 1997,
2000 [Data file and codebook]. Ann Arbor, MI: Inter-University Consortium for Political
and Social Research [Distributor].
Bradbury, T. N., Fincham, F. D., & Beach, S. R. H. (2000). Research on the nature and
determinants of marital satisfaction: A decade in review. Journal of Marriage and
Family, 62, 964-980.
Bradbury, T. N., & Lavner, J. A. (2012). How can we improve preventive and educational
interventions for intimate relationships? Behavior Therapy, 43, 113-122.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three
decades of progress. Psychological Bulletin, 125, 276-302.
Donnellan, M. B., Assad, K. K., Robins, R. W., & Conger, R. D. (2007). Do negative interactions
mediate the effects of negative emotionality, communal positive emotionality, and
constraint on relationship satisfaction? Journal of Social and Personal Relationships, 24,
557-573.
Donnellan, M. B., Trzesniewski, K. H., Conger, K. J., & Conger, R. D. (2007). A three-wave
longitudinal study of self-evaluations during young adulthood. Journal of Research in
Personality, 41, 453-472.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 30
Dryer, D. C., & Horowitz, L. M. (1997). When do opposites attract? Interpersonal
complementarity versus similarity. Journal of Personality and Social Psychology, 72,
592-603.
Dyrenforth, P. S., Kashy, D. A., Donnellan, M. B., & Lucas, R. E. (2010). Predicting relationship
and life satisfaction from personality in nationally representative samples from three
countries: The relative importance of actor, partner, and similarity effects. Journal of
Personality and Social Psychology, 99, 690-702.
Eagly, A. H., & Wood, W. (1999). The origins of sex differences in human behavior: Evolved
dispositions versus social roles. American Psychologist, 54, 408-423.
Edwards, M. C., & Wirth, R. J. (2009). Measurement and the study of change. Research in
Human Development, 6, 74-96.
Erol, R. Y., & Orth, U. (2011). Self-esteem development from age 14 to 30: A longitudinal study.
Journal of Personality and Social Psychology, 101, 607-619.
Erol, R. Y., & Orth, U. (2013). Actor and partner effects of self-esteem on relationship
satisfaction and the mediating role of secure attachment between the partners. Journal of
Research in Personality, 47, 26-35.
Fincham, F. D., & Bradbury, T. N. (1987). The assessment of marital quality: A reevaluation.
Journal of Marriage and the Family, 49, 797-809.
Fincham, F. D., & Bradbury, T. N. (1993). Marital satisfaction, depression, and attributions: A
longitudinal analysis. Journal of Personality and Social Psychology, 64, 442-452.
Fincham, F. D., & Linfield, K. J. (1997). A new look at marital quality: Can spouses feel positive
and negative about their marriage? Journal of Family Psychology, 11, 489-502.
Fraley, R. C., Garner, J. P., & Shaver, P. R. (2000). Adult attachment and the defensive
regulation of attention and memory: Examining the role of preemtive and postemptive

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 31
defensive processes. Journal of Personality and Social Psychology, 79, 816-826.
Gilford, R., & Bengtson, V. (1979). Measuring marital satisfaction in three generations: Positive
and negative dimensions. Journal of Marriage and Family, 41, 387-398.
Gonzaga, G. C., Campos, B., & Bradbury, T. N. (2007). Similarity, convergence, and relationship
satisfaction in dating and married couples. Journal of Personality and Social Psychology,
93, 34-48.
Gray-Little, B., Williams, V. S. L., & Hancock, T. D. (1997). An item response theory analysis of
the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 23, 443-
451.
Grimm, K. J. (2007). Multivariate longitudinal methods for studying developmental relationships
between depression and academic achievement. International Journal of Behavioral
Development, 31, 328-339.
Hendrick, S. S. (1988). A generic measure of relationship satisfaction. Journal of Marriage and
the Family, 50, 93-98.
Hendrick, S. S., Hendrick, C., & Adler, N. L. (1988). Romantic relationships: Love, satisfaction,
and staying together. Journal of Personality and Social Psychology, 54, 980-988.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability:
A review of theory, method, and research. Psychological Bulletin, 118, 3-34.
Karney, B. R., & Bradbury, T. N. (1997). Neuroticism, marital interaction, and the trajectory of
marital satisfaction. Journal of Personality and Social Psychology, 72, 1075-1092.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York, NY:
Guilford Press.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 32
Kling, K. C., Hyde, J. S., Showers, C. J., & Buswell, B. N. (1999). Gender differences in self-
esteem: A meta-analysis. Psychological Bulletin, 125, 470-500.
Kurdek, L. A. (2005). Gender and marital satisfaction early in marriage: A growth curve
approach. Journal of Marriage and Family, 67, 68-84.
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, 677-690.
Kuster, F., Orth, U., & Meier, L. L. (2013). High self-esteem prospectively predicts better work
conditions and outcomes. Social Psychological and Personality Science, 4, 668-675.
Lavner, J. A., & Bradbury, T. N. (2010). Patterns of change in marital satisfaction over the
newlywed years. Journal of Marriage and Family, 72, 1171-1187.
Lavner, J. A., & Bradbury, T. N. (2012). Do cold feet warn of trouble ahead? Premarital
uncertainty and four-year marital outcomes. Journal of Family Psychology, 26, 1012-
1017.
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, 606-616.
Ledermann, T., & Kenny, D. A. (2012). The common fate model for dyadic data: Variations of a
theoretically important but underutilized model. Journal of Family Psychology, 26, 140-
148.
Ledermann, T., & Macho, S. (2014). Analyzing change at the dyadic level: The common fate
growth model. Journal of Family Psychology, 28, 204-213.
MacCallum, R. C., Browne, M. W., & Cai, L. (2006). Testing differences between nested
covariance structure models: Power analysis and null hypotheses. Psychological Methods,
11, 19-35.
Markides, K. S., Roberts-Jolly, J., Ray, L. A., Hoppe, S. K., & Rudkin, L. (1999). Changes in

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 33
marital satisfaction in three generations of Mexican Americans. Research on Aging, 21,
36-45.
McArdle, J. J. (1988). Dynamic but structural equation modeling of repeated measure data. In J.
R. Nesselroade & R. B. Cattell (Eds.), Handbook of multivariate experimental data (Vol.
2, pp. 561-614). New York, NY: Plenum Press.
Meeks, B. S., Hendrick, S. S., & Hendrick, C. (1998). Communication, love and relationship
satisfaction. Journal of Social and Personal Relationships, 15, 755-773.
Meier, L. L., Orth, U., Denissen, J. J. A., & Kühnel, A. (2011). Age differences in stability,
contingency, and level of self-esteem across the life span. Journal of Research in
Personality, 45, 604-612.
Murray, S. L., Holmes, J. G., Bellavia, G. M., Griffin, D. W., & Dolderman, D. (2002). Kindred
spirits? The benefit of egocentrism in close relationships. Journal of Personality and
Social Psychology, 82, 563-581.
Murray, S. L., Holmes, J. G., & Griffin, D. W. (1996). The benefits of positive illusions:
Idealization and the construction of satisfaction in close relationships. Journal of
Personality and Social Psychology, 70, 79-98.
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, 478-498.
Muthén, L. K., & Muthén, B. O. (2010). Mplus user's guide: Sixth edition. Los Angeles, CA:
Muthén and Muthén.
Orbuch, T. L., House, J. S., Mero, R. P., & Webster, P. S. (1996). Marital quality over the life
course. Social Psychology Quarterly, 59, 162-171.
Orth, U., Robins, R. W., & Widaman, K. F. (2012). Life-span development of self-esteem and its

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 34
effects on important life outcomes. Journal of Personality and Social Psychology, 102,
1271-1288.
Orth, U., Trzesniewski, K. H., & Robins, R. W. (2010). Self-esteem development from young
adulthood to old age: A cohort-sequential longitudinal study. Journal of Personality and
Social Psychology, 98, 645-658.
Preacher, K. J., Wichman, A. L., MacCallum, R. C., & Briggs, N. E. (2008). Latent growth curve
modeling. Los Angeles, CA: Sage.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement, 1, 385-401.
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, 151-161.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University
Press.
Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.
Psychological Methods, 7, 147-177.
Shackelford, T. K. (2001). Self-esteem in marriage. Personality and Individual Differences, 30,
371-390.
Shaver, P. R., Schachner, D. A., & Mikulincer, M. (2005). Attachment style, excessive
reassurance seeking, relationship processes, and depression. Personality and Social
Psychology Bulletin, 31, 343-359.
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of
marriage and similar dyads. Journal of Marriage and Family, 38, 15-28.
Swann, W. B., & Bosson, J. K. (2010). Self and identity. In S. T. Fiske, D. T. Gilbert & G.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 35
Lindzey (Eds.), Handbook of social psychology (Vol. 1, pp. 589-628). Hoboken, NJ:
Wiley.
Sweet, J. A., & Bumpass, L. L. (2002). The National Survey of Families and Households, 2001-
2003, Wave III [Data file and codebook]. Madison, WI: University of Wisconsin, Center
for Demography and Ecology [Distributor].
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, 205-220.
Vaillant, C. O., & Vaillant, G. E. (1993). Is the U-Curve of marital satisfaction an illusion? A 40-
year study of marriage. Journal of Marriage and Family, 55, 230-239.
VanLaningham, J., Johnson, D. R., & Amato, P. (2001). Marital happiness, marital duration, and
the U-shaped curve: Evidence from a five-wave panel study. Social Forces, 78, 1313-
1341.
Widaman, K. F., Ferrer, E., & Conger, R. D. (2010). Factorial invariance within longitudinal
structural equation models: Measuring the same construct across time. Child Development
Perspectives, 4, 10-18.
Williamson, H. C., Ju, X., Bradbury, T. N., Karney, B. R., Fang, X., & Liu, X. (2012).
Communication behavior and relationship satisfaction among American and Chinese
newlywed couples. Journal of Family Psychology, 26, 308-315.
Wirth, R. J., & Edwards, M. C. (2007). Item factor analysis: Current approaches and future
directions. Psychological Methods, 12, 58-79.
Zeigler-Hill, V. (2010). The interpersonal nature of self-esteem: Do different measures of self-
esteem possess similar interpersonal content? Journal of Research in Personality, 44, 22-
30.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 36
Footnotes
1Although for identification purposes it would be sufficient to fix the loadings for one of
the partners, fixing both loadings provides for more stable solutions when only two indicators are
available per latent factor (Kenny, Kashy, & Cook, 2006).
2 We also tested a bivariate growth curve model in which men and women had separate
growth factors for relationship satisfaction. The model fit was good (TLI = .98, CFI = .98,
RMSEA = .045). However, given that both the common fate growth model and the bivariate
growth model provided a good fit, we used the simpler model (i.e., the common fate growth
model) for the analyses.
3 Although the coefficients were constrained to be equal across gender, the constraints
were imposed on unstandardized coefficients (as typically recommended), which led to slight
variation in the resulting standardized coefficients.
4 As in Study 1, we also tested a bivariate growth curve model in which men and women
had separate growth factors for relationship satisfaction. The fit was acceptable (TLI = .80, CFI =
.95, RMSEA = .076). However, given that both the common fate growth model and the bivariate
growth model provided an acceptable fit, we used the simpler model (i.e., the common fate
growth model) for the analyses.
5 To allow for convergence of the model, we constrained the variances of the growth
factors to be equal across groups.
6 To allow for convergence of the model, we constrained the variances of the growth
factors to be equal across groups and in the Hispanic group we fixed the correlation between the
partners’ self-esteem slopes to the value of the model with cross-group constraints.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 37
Table 1
Means and Standard Deviations of Self-Esteem and Relationship Satisfaction Across Waves,
Separately for Men and Women
Self-esteem
Relationship satisfaction
Men
Women
Men
Women
Study 1 (LSG)
Time 1
3.52 (0.45)
3.52 (0.44)
3.90 (0.64)
3.91 (0.70)
Time 2
3.55 (0.44)
3.53 (0.43)
3.89 (0.70)
3.89 (0.76)
Time 3
3.35 (0.43)
3.31 (0.42)
3.89 (0.66)
3.91 (0.67)
Time 4
3.36 (0.41)
3.32 (0.42)
3.96 (0.64)
3.92 (0.74)
Time 5
3.36 (0.45)
3.33 (0.44)
3.98 (0.64)
3.98 (0.67)
Study 2 (NSFH)
Time 1
4.10 (0.57)
4.11 (0.60)
6.12 (1.19)
6.05 (1.28)
Time 2
4.12 (0.61)
4.03 (0.64)
5.95 (1.28)
5.91 (1.34)
Time 3
4.08 (0.57)
4.04 (0.58)
6.19 (1.06)
6.13 (1.16)
Note. Values shown in parentheses are standard deviations. In Study 1, the response scales
ranged from 1 to 4 for self-esteem and from 1 to 5 for relationship satisfaction. In Study 2, the
response scales ranged from 1 to 5 for self-esteem and from 1 to 7 for relationship satisfaction.
LSG = Longitudinal Study of Generations; NSFH = National Survey of Families and
Households.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 38
Figure 1. Generic model used for the common fate growth curve model for relationship
satisfaction. RS = relationship satisfaction; RS-M = men’s relationship satisfaction; RS-W =
women’s relationship satisfaction; T1 to T5 = Time 1 to Time 5.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 39
Figure 2. Generic model used for the analysis of the effects of men’s and women’s growth
factors for self-esteem on the growth factors for relationship satisfaction. The model controls for
the effects of age and length of relationship, by regressing the growth factors for self-esteem and
relationship satisfaction on the control variables (not shown in the figure). SE = self-esteem; SE-
M = men’s self-esteem; SE-W = women’s self-esteem; RS = relationship satisfaction; RS-M =
men’s relationship satisfaction; RS-W = women’s relationship satisfaction; T1 to T5 = Time 1 to
Time 5.

SELF-ESTEEM AND RELATIONSHIP SATISFACTION 40
Figure 3. Standardized estimates of self-esteem predicting relationship satisfaction in Study 1
(Panel A) and Study 2 (Panel B). The models control for the effects of age and length of
relationship, by regressing the growth factors for self-esteem and relationship satisfaction on the
control variables (not shown in the figure). SE = self-esteem; RS = relationship satisfaction. * p <
.05.