You are on page 1of 15

Personality and Social Psychology

Bulletin http://psp.sagepub.com/

Closeness Discrepancies in Romantic Relationships : Implications for Relational Well-Being, Stability, and
Mental Health
David M. Frost and Cat Forrester
Pers Soc Psychol Bull published online 13 February 2013
DOI: 10.1177/0146167213476896

The online version of this article can be found at:


http://psp.sagepub.com/content/early/2013/02/13/0146167213476896

Published by:

http://www.sagepublications.com

On behalf of:

Society for Personality and Social Psychology

Additional services and information for Personality and Social Psychology Bulletin can be found at:

Email Alerts: http://psp.sagepub.com/cgi/alerts

Subscriptions: http://psp.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

>> OnlineFirst Version of Record - Feb 13, 2013

What is This?

Downloaded from psp.sagepub.com by guest on February 17, 2013


476896
ality and Social Psychology BulletinFrost and Forrester
PSPXXX10.1177/0146167213476896Person

Personality and Social

Closeness Discrepancies in
Psychology Bulletin
XX(X) 1­–14
© 2013 by the Society for Personality

Romantic Relationships: Implications and Social Psychology, Inc


Reprints and permission:
sagepub.com/journalsPermissions.nav
for Relational Well-Being, Stability, DOI: 10.1177/0146167213476896
pspb.sagepub.com

and Mental Health

David M. Frost1 and Cat Forrester2

Abstract
Experiences of closeness in romantic relationships are associated with heightened levels of relational well-being and mental
health. However, individuals differ in the degree of closeness they desire in their relationships. This study used the construct
of inclusion of other in self (IOS) to analyze discrepancies between individuals’ actual and ideal levels of closeness in their
relationships and the implications of these discrepancies for relational well-being and mental health. A longitudinal survey of
partnered individuals demonstrated that optimal levels of relational well-being and mental health existed when individuals had
minimal discrepancies between actual and ideal IOS, regardless of their actual levels of IOS. Individuals whose actual levels of IOS
moved closer to their ideal levels over a 2-year period reported improved relational well-being and mental health. Individuals
with little to no discrepancies between actual and ideal IOS were also less likely to break-up with their partners over time.

Keywords
Closeness, Health, Relationships, Discrepancies, Inclusion of Other in Self

Received June 5, 2012; revision accepted November 15, 2012

Closeness, Health, and the Quality of Research on IOS as an indicator of feelings of closeness in
Romantic Relationships romantic relationships has demonstrated that greater inclu-
sion of one’s partner in one’s self-concept is associated with a
Close relationships with others—particularly romantic rela- multitude of indicators of relational well-being, including
tionship partners—have a substantial impact on individuals’ greater relationship satisfaction and lower rates of dissolution
health and well-being (Holt-Lunstad, Smith, & Layton, over time (Le, Dove, Agnew, Korn, & Mutso, 2010; Tsapelas,
2010; Kiecolt-Glaser, Gouin, & Hantsoo, 2010; Reis & Aron, & Orbuch, 2009). In addition, findings across a variety
Aron, 2008; Reis, Collins, & Berscheid, 2000; Robles & of cognitive and linguistic studies indicate that a greater sense
Kiecolt-Glaser, 2003). Feelings of closeness within romantic of “we-ness” between partners within relationships is associ-
relationships play an important role in determining the qual- ated with increased relationship satisfaction and commitment
ity of romantic relationships and partners’ mental health (Agnew, Van Lange, Rusbult, & Langston, 1998; Fitzsimons
(Khaleque, 2004). Feeling close to a romantic partner is & Kay, 2004; Frost, in press; Seider, Hirschberger, Nelson, &
indicative of a more intimate, committed, and satisfying Levenson, 2009). Self-expansion through IOS may also result
relationship (Brunell, Pilkington, & Webster, 2007; in beneficial relational outcomes by promoting positive inter-
Hassebrauck & Fehr, 2002). Closeness in romantic relation- dependence between partners in the form of shared interests
ships is further representative of the degree to which indi- and mutuality of dependence (Rusbult & Van Lange, 2003).
viduals include aspects of their partners within their own
self-concept: inclusion of other in self (IOS; Agnew, Loving, 1
Columbia University, New York, NY, USA
Le, & Goodfriend, 2004; Aron, Aron, & Smollan, 1992; 2
San Francisco State University, CA, USA
Aron, Aron, Tudor, & Nelson, 1991). Seeking self-expansion
by including qualities of one’s partner in one’s self-concept Corresponding Author:
David M. Frost, Mailman School of Public Health, Heilbrunn Department of
(e.g., their identities, resources, experiences) is thought to Population and Family Health, Columbia University, 60 Haven Avenue B2,
produce beneficial outcomes at both the relational and indi- New York, NY 10032, USA
vidual levels (Aron, Norman, & Aron, 2001). Email: dmf2119@columbia.edu

Downloaded from psp.sagepub.com by guest on February 17, 2013


2 Personality and Social Psychology Bulletin XX(X)

The majority of psychological research therefore indicates attachment) may correspond to not feeling close enough to a
that closeness—or more overlap between one’s self and one’s partner (Bartholomew & Horowitz, 1991). Adult attachment
partner—is a key marker of quality, healthy relationships. theory positions attachment avoidance as detrimental to rela-
tionship quality and stability because attachment avoidance is
characterized by a fear of intimacy, as well as an increased
Closeness Discrepancies and Their desire to maintain distance and independence at the expense
Potential Negative Effects of intimacy within the context of close relationships (see
However, individuals vary in the amount of closeness they Hazan & Shaver, 1994, for a review). Empirical evidence
desire within their relationships (Aron et al., 2004; Fletcher, exists for this connection as well: Attachment avoidance is
Simpson, Thomas, & Giles, 1999; Goodboy & Booth- negatively associated with multiple indicators of well-being
Butterfield, 2009; Kashdan, Volkmann, Breen, & Han, 2007; in romantic and marital relationships (Birnbaum, Reis,
Mashek, Le, Israel, & Aron, 2011; Mashek & Sherman, Mikulincer, Gillath, & Orpaz, 2006; Brennan & Shaver,
2004). Recent advances in work on IOS have pointed to the 1995; Butzer & Campbell, 2008). The co-occurrence of
phenomena of desiring less closeness in romantic relation- closeness discrepancies with attachment avoidance therefore
ships. Specifically, too much IOS may pose a threat to per- suggests that actual–ideal IOS discrepancies will be indica-
sonal control and identity, and therefore cause someone to tive of negative relationship outcomes given the demon-
desire less closeness with an intimate partner (Aron et al., strated deleterious effects of avoidance on relationship
2004; Mashek & Sherman, 2004). Similar problems arise satisfaction and stability.
when considering the phenomena of enmeshment in marital
and romantic relationships. Increased experiences of close-
ness may be beneficial to relational well-being and mental Social Cognitive and Motivational
health when they take the form of intimacy, caregiving, and Perspectives on Closeness
nurturance, but detrimental when they take the form of intru- Discrepancies and Well-Being
siveness and loss of a sense of self and independence (Green
& Werner, 1996; Lewandowski, Nardone, & Raines, 2010; Self-discrepancy theory (e.g., Higgins, 1987)—and its exten-
Werner, Green, Greenberg, Browne, & McKenna, 2001). sion to discrepant self-concepts within dyads (e.g., Robins &
Thus, feeling “too close” to one’s partner and the corre- Boldero, 2003)—provides some insight into how closeness
sponding perceived loss of self may have a detrimental discrepancies may have a deleterious impact on relational
effect on relational well-being and mental health (Aron well-being, stability, and mental health. More specifically,
et al., 2004; Mashek & Sherman, 2004). one’s idealized level of closeness or IOS may serve as a cog-
Taken together, the research reviewed previously suggests nitive guide to which one compares his or her actual experi-
the utility of an understanding of feelings of closeness in ences of closeness or IOS. Although not focused on closeness,
romantic relationships that can account for individuals’ actual discrepancies between actual and ideal selves can result in
experiences of IOS in relation to their ideal experiences of “dejection-related emotions” such as disappointment and dis-
IOS. In other words, individuals in romantic relationships satisfaction (Higgins, 1987). When idealized standards for
who do not currently experience their ideal level of closeness characteristics of partners do not match perceived partner
with their partners may feel either “not close enough” or “too characteristics, people become dissatisfied with their rela-
close” to their partners. Although the theoretical importance tionships (Fletcher & Simpson, 2000).
and existence of these discrepancies between actual and ideal Extending this line of reasoning to understanding the
experiences of IOS have been documented, the implications influence of closeness on relational well-being, it follows
that IOS discrepancies actually have for the quality and sta- that discrepancies between actual and ideal levels of close-
bility of romantic relationships remain largely unexamined ness may translate to similar outcomes at the relational level:
(Aron et al., 2004; Mashek & Sherman, 2004). decreased relationship satisfaction and commitment.
Research on individual differences in experiences of IOS Furthermore, following findings relating self-discrepancies
discrepancies is further suggestive of the potential detrimen- to mental health (Heidrich & Powwattana, 2004; Scott &
tal effects of IOS discrepancies on relational well-being. O’Hara, 1993), actual–ideal closeness discrepancies may
Attachment avoidance is associated with experiencing a dis- result in negative mental health outcomes, such as depres-
crepancy between actual and ideal IOS in relationships, sion, given the centrality of dejection-related emotions to
whereas the lowest levels of attachment avoidance can be depressive symptomatology (e.g., Radloff, 1977). Finally,
found among individuals who experience no actual–ideal just as individuals are motivated to reduce actual–ideal self-
IOS discrepancies (Mashek & Sherman, 2004). Avoidance in discrepancies (Higgins, 1987) and discrepancies between
the form of fears of intimacy and perceived threats to inde- perceived and ideal partner characteristics (Fletcher &
pendence (i.e., dismissive-avoidant attachment) may mani- Simpson, 2000), the drive to bring one’s actual experience of
fest in feeling “too close” to a partner, while desiring closeness closeness in line with one’s ideal experience of closeness
yet anticipating rejection from a partner (i.e., fearful-avoidant may represent an energizing cognition that, when actualized

Downloaded from psp.sagepub.com by guest on February 17, 2013


Frost and Forrester 3

over time, should result in positive trajectories in relational and mental health above and beyond actual closeness alone
well-being and mental health. Conversely, failure to dimin- can potentially increase the explanatory power and applied
ish actual–ideal closeness discrepancies may result in nega- relevance of psychological models of health and well-being
tive trajectories in relational well-being and mental health in romantic relationships.
over time, and ultimately relationship dissolution.
Similar predictions about the effects of actual–ideal close-
ness discrepancies on relational well-being and mental health The Current Study: Aims and
can be gained from self-determination theory (SDT; Ryan & Hypotheses
Deci, 2000). Research using an SDT approach has demon- The current study uses an analysis of closeness as IOS and
strated that relational well-being and mental health result seeks to extend and replicate social cognitive theory and
from the satisfaction of fundamental psychological needs, evidence on actual–ideal discrepancies within the domain of
including relatedness and autonomy, in addition to compe- closeness. Utilizing the construct of actual–ideal IOS dis-
tency (e.g., Patrick, Knee, Canevello, & Lonsbary, 2007; crepancies as developed in research on experiences of desir-
Sheldon, Williams, & Joiner, 2003). Feeling less closeness ing less closeness (Aron et al., 2004; Mashek et al., 2011;
than idealized may be indicative of an unmet need for relat- Mashek & Sherman, 2004), the current study aims to exam-
edness in the context of a present partnership, whereas feel- ine the degree to which actual–ideal IOS discrepancies are
ing more closeness than idealized may indicate, in a sense, associated with relational well-being, stability, and mental
over-relatedness, which poses a threat to actualization of the health above and beyond the actual experience of closeness
need for autonomy. Alignment of actual and ideal experi- alone. The implications of IOS and actual–ideal discrepan-
ences of closeness may therefore represent an optimal bal- cies for mental health in particular have yet to be examined.
ance between the needs for relatedness and autonomy as Furthermore—with increased empirical attention being paid
subjectively experienced by the individual. As a result of to the ever-changing nature of romantic relationships (e.g.,
these threats to fundamental psychogenic needs for related- Berscheid, 2010)—it is imperative to understand how the
ness and autonomy, closeness discrepancies in romantic rela- attenuation or exacerbation of actual–ideal IOS discrepancies
tionships may put individuals at risk for diminished over time correspond to temporal changes in relational well-
well-being in both psychological and relational domains. being, stability, and mental health. Accounting for dynamic
experiences of actual–ideal IOS discrepancies and their
resulting implications is a necessary next step for a psycho-
Beyond Actual Closeness:The logical science that seeks to better its understanding of the
Potential Additive Influence experience of closeness, as well as improve its ability to
of Closeness Discrepancies on promote and facilitate healthy, lasting relationships.
Relational Well-Being and Mental The current study hypothesized that discrepancies between
Health actual and ideal levels of IOS would predict relational well-
being outcomes (i.e., relationship satisfaction, commitment,
Researchers have yet to understand the degree to which and break-up thoughts) and mental health (i.e., depression)
actual–ideal IOS discrepancies impact relational well-being above and beyond individuals’ actual experiences of IOS
above and beyond actual experiences of IOS alone. In other alone. Furthermore, it was hypothesized that the association
words, it is unclear what matters most for relational well- between IOS discrepancies and relational well-being and
being: how close people actually feel to their partners, or mental health outcomes would be nonlinear. Individuals with
whether people’s actual experiences of closeness in their little to no discrepancies between their actual and ideal levels
relationships match how close they would like to feel to their of IOS were expected to demonstrate optimal levels of rela-
partners. Closeness is a subjective psychological experience, tional well-being and mental health. Experiencing higher lev-
and individuals have been shown to vary in how much close- els of IOS than desired (i.e., feeling “too close” with one’s
ness they idealize within their romantic relationships (Aron partner) as well as lower levels of IOS than desired (i.e., feel-
et al., 2004; Fletcher et al., 1999; Goodboy & Booth- ing “not close enough” with one’s partner) were both hypoth-
Butterfield, 2009; Mashek & Sherman, 2004). Thus, a pre- esized to be associated with decreased levels of relational
scriptive model—one in which more closeness always results well-being and mental health. Furthermore, changes in
in better and healthier relationships—may not be the most actual–ideal IOS discrepancies over time were hypothesized
effective theoretical, empirical, and clinical approach to to be predictive of corresponding changes in relational well-
understanding the role of closeness in determining relational being and mental health. It was hypothesized that individuals
well-being, stability, and mental health. Although there is whose actual experiences of IOS with their partners moved
evidence that, on average, closer relationships tend to be bet- closer to their ideal IOS over time would show improvement
ter relationships (Brunell et al., 2007; Hassebrauck & Fehr, in relational well-being and mental health over time com-
2002), understanding whether and how actual–ideal close- pared with individuals whose actual experiences of IOS with
ness discrepancies can predict relational well-being, stability, their partners moved further away from their ideal levels of

Downloaded from psp.sagepub.com by guest on February 17, 2013


4 Personality and Social Psychology Bulletin XX(X)

IOS. Comparatively, little to no change in relational well- Forty-four percent (n = 732) of the originally coupled par-
being and mental health was hypothesized among individuals ticipants were retained for all waves of the study. Participants
with no change in actual–ideal IOS discrepancies over time. were invited to complete all waves of the survey regardless
Finally, it was hypothesized that individuals with no actual– of whether they were in the same relationship, had broken up
ideal IOS discrepancies would be less likely to break up with with their previous partner, and/or started a new relationship.
their partners over time than individuals who felt “too close” There were no significant differences in the demographic
or “not close enough” to their partners. and relationship characteristics reported above between
retained participants and those who were lost to follow-up,
with the exception of sexual orientation. More attrition was
Method observed among heterosexual participants (64%) than among
Sample gay men and lesbians (50%) and bisexuals (56%), χ2(2) =
16.08, p < .001. There were also no differences between
To test these hypotheses, data were obtained from the Lives retained participants and those lost to follow-up on any of
and Relationships Study: an Internet-based longitudinal survey the relational well-being outcomes in the study. However,
focused on understanding predictors of relational well-being those lost to follow-up were significantly more depressed as
and health among a diverse sample of individuals living in the measured by the Center for Epidemiological Studies—
United States and Canada. Four waves of data were collected, Depression scale (CESD; M = 9.06, SD = 6.09) than retained
with each wave occurring approximately 1 year (M = 370 participants (M = 7.91, SD = 5.76), t = 4.03, p < .001.
days, SD = 13) apart. Data from the first wave were not
included in the current analysis because IOS data were only
collected during the final three waves. Participants were ini- Procedures and Measures
tially recruited into the study using three recruitment strategies. After consenting to participate and entering their email
They were active strategies (e.g., emails and listserv postings addresses and passwords, participants were granted access to
that reached participants directly through their personal email the online questionnaire that contained the measures below.
accounts), passive strategies (e.g., posting an announcement
on discussion forums or classified websites), and snowball IOS (Times 1, 2, and 3). Aron and colleagues’ (1992) IOS scale
strategies (i.e., participants were provided with a link to the was used to measure how individuals conceptualized their
study that they were encouraged to share with their friends, own experiences of relational closeness. This pictorial scale
family, and coworkers). Announcements were not placed on depicted six sets of two circles in which one circle repre-
listservs or discussion forums specifically focused on dating, sented the participant’s “self” and the other represented the
relationships, sex, or relationship problems to avoid biasing the participant’s “partner.” The sets were presented with varying
sample. During the first wave of data collection, participants degrees of overlap ranging from completely separate to
provided their email addresses and unique passwords that were almost completely overlapping. The IOS has empirically
used to link participants’ data from wave to wave. At subse- demonstrated validity and reliability (see Aron et al., 1992),
quent waves, participants received an email directly from the most notably in its robust correlations with several other
first author announcing the new study wave, reminding them multi-item scales that measure closeness within relation-
of their username and password, and providing a link to the ships, such as the Subjective Closeness Index and the Rela-
study’s website. Incentive for participation at each wave was tionship Closeness Inventory. The current study used a
provided in the form of a lottery drawing for 1 of 20 US$100 2-item approach to assessing IOS, where one version of the
gift cards to a popular online retailer. scale assessed participants’ actual (i.e., “current”) levels of
Data for the current study were obtained from the sub- IOS and a second version assessed participants’ ideal levels
sample of individual participants who were in relationships of IOS (Aron, Melinat, Aron, Vallone, & Bator, 1997;
at the second wave of the study (N = 1,659) and their follow- Mashek & Sherman, 2004). IOS discrepancy scores were
up data from Waves 3 and 4. Waves 2, 3, and 4 are hereafter computed by subtracting each participant’s ideal IOS rating
referred to as Times 1, 2, and 3 to simplify the presentation from their actual IOS rating. Negative numbers on this IOS
of findings within the context of the current investigation. discrepancy score indicated feeling “not close enough” to
The sample was diverse in terms of age (M = 34.11, SD = one’s partner, while positive numbers indicated feeling “too
10), race/ethnicity (23% non-Caucasian ethnic/racial minor- close” to one’s partner, and scores of 0 indicated no discrep-
ity), sexual orientation (14% nonheterosexual), and educa- ancy between actual and ideal experiences of IOS.
tional attainment (54% having a 4-year college degree or
greater). The sample was primarily female (78% female). Relationship Satisfaction (Times 1, 2, and 3). The 4-item version
Participants reported being in a diverse array of relationships of the Couples Satisfaction Index (CSI; Funk & Rogge, 2007)
with regard to relationship length (M = 8.74 years, SD = was included to assess individuals’ satisfaction with their cur-
8.27), marital status (58% married or domestic partners), and rent primary romantic relationships. The CSI was developed
cohabitation (77% lived with their partners). using item response theory and is the result of a factor analysis

Downloaded from psp.sagepub.com by guest on February 17, 2013


Frost and Forrester 5

of items pooled from eight previously validated measures of original 20-item version (Radloff, 1977). Participants were
relationship satisfaction. Example items include “How asked how often during the past week they “could not get
rewarding is your relationship with your partner?” and “In going,” “felt depressed,” “felt hopeful about the future,” and
general, how satisfied are you with your relationship?” Par- “felt people dislike you.” Participants responded on a 4-point
ticipants responded to such items on a scale of “not at all” to scale ranging from 0 = rarely or none of the time (<1 day) to
“completely.” The CSI not only demonstrates strong validity 3 = most or all of the time (5-7 days). Numerous studies have
correlations with existing measures of the construct, but it also demonstrated the convergent validity of the CESD among
demonstrates less noise and more power in detecting individ- both clinical and nonclinical samples in the form of large
ual differences in satisfaction than existing measures (Funk & correlations with clinical reports of depression, Diagnostic
Rogge, 2007). The measure is scored on a scale of 0 to 21, and Statistical Manual of Mental Disorders depression diag-
with scores of 13.5 or below indicating relationship distress. noses, the Hamilton Rating Scale for Depression, and the
In its originating study (Funk & Rogge, 2007), CSI (4-item) Symptom Checklist–90 (for a review of validity evidence,
scale scores were internally consistent at .94 and were highly see McDowell & Newell, 1996; Roberts & Vernon, 1983).
correlated with the eight previously validated measures of Although the scale has been shown to correlate moderately
relationship satisfaction (rs ranged from .84 to .94). In the cur- to highly with other measures of anxiety and psychological
rent study, participants’ scores on the CSI were internally con- distress, it has been successful in identifying depression in
sistent with Cronbach’s alpha coefficients of .91 at each of the several clinical and community samples (McDowell & New-
three time points of the current study. ell, 1996) and as a result is one of the most widely used mea-
sures of depressive symptoms. Internal consistency for
Commitment (Time 1). Commitment was measured using the scores on the CESD in the current study ranged from .85 to
9-item Relationship Commitment Scale (Lund, 1985). Com- .87 over the three time points.
mitment was included as a complement to relationship satis-
faction given its ability to assess participants’ expectations
about the future of their relationships in addition to their cur- Analysis Plan
rent levels of satisfaction. Example items read, “How likely Cross-Sectional Segmented Regression Models. To estimate the
is it that your relationship will be permanent?” and “How effect of actual–ideal IOS discrepancies on relational well-
likely are you to pursue another relationship or single life in being and mental health, it was necessary to fit a spline to the
the future?” Participants responded to items on a scale of data. A spline, in its simplest form, can be tested in the form
1 = not at all to 7 = extremely. Final scale average scores of a segmented regression model in which linear change is
were created that ranged from 1 to 7. In its originating study, expected to be continuous up to a certain value (i.e., a knot),
scores on the measure demonstrated internal consistency at after which the size and direction of change will be altered. In
.82, and the scale demonstrated an ability to distinguish the current study, it was hypothesized that an actual–ideal
between other relational constructs focused on positive affect IOS discrepancy of 0—where actual closeness and ideal
(e.g., love; Lund, 1985). Scores on the Relationship Com- closeness are equal—would represent the point of expected
mitment Scale in the current study were internally consistent linear change. Specifically, it was hypothesized that a posi-
at .76 at Time 1. tive relationship would exist between closeness discrepancies
and positively valenced study outcomes (i.e., relationship sat-
Break-Up Thoughts (Times 1, 2, and 3). A single item was isfaction and commitment) as negative values of closeness
included in the survey to measure how often participants discrepancies (i.e., feeling not close enough) approached 0.
thought about ending their relationships with their current Alternatively, a negative relationship between closeness dis-
primary partners. The item read, “How often in the past crepancies and positively valenced outcomes was expected as
month have you considered ending your relationship with positive closeness discrepancies (i.e., feeling too close)
your current partner?” Participants responded on a scale of increased beyond 0. This pattern was expected to reverse for
0 = never to 4 = a lot. Higher scores were considered indica- negatively valenced outcomes (i.e., break-up thoughts and
tive of more seriousness of participants’ considerations to depression). Put more simply, having no actual–ideal IOS
dissolve their relationships, which have been demonstrated discrepancy was hypothesized to correspond to optimal lev-
to be key determinants of actual relationship dissolution els of relational well-being and mental health, with relational
(e.g., Gottman, 1993). well-being and mental health worsening as closeness discrep-
ancies increased beyond 0 in either direction.
Mental Health (Times 1, 2, and 3). Mental health was assessed Segmented linear regression models were fit to test these
in the form of symptoms of depression using the short form hypotheses (following Keele, 2008). These models included a
CESD (Andresen, Malmgren, Carter, & Patrick, 1994). The parameter to estimate the size and direction of the relation-
CESD short form is a 10-item measure of depressive symp- ship between closeness discrepancies and study outcomes for
toms experienced over a 1-week period prior to the study values of actual–ideal IOS discrepancies ranging from −6 to 0,
view. The 10-item version uses a subset of the items from the and a separate parameter to estimate the size and direction of

Downloaded from psp.sagepub.com by guest on February 17, 2013


6 Personality and Social Psychology Bulletin XX(X)

the relationship between actual–ideal IOS discrepancies and assessments among retained participants were minimal, with
study outcomes for values of closeness discrepancies ranging no more than 17 participants missing data on a given vari-
beyond 0 to 6. This segmented linear regression model is rep- able. Missing values were imputed using Expectation Maxi-
resented in the equation below where x = an individual’s mization, an imputation technique based in Maximum
closeness discrepancy score and k = the knot discrepancy Likelihood Estimation (Roth, 1994; Schafer & Olsen, 1998).
value of 0 at which the direction of linear change is expected
to be altered: Binary Logistic Regression Models. Finally, logistic regression
models were computed to examine whether actual–ideal IOS
Y = b0 + b1 x + b2 ( x) + + e,
discrepancies were predictive of relationship dissolution
over a 1-year period. Binary indicator variables were com-
 x if x > k  puted at Time 1 for those who reported higher actual IOS
where x+ =  .
 0 if x ≤ k  than ideal IOS with their partner (i.e., “too close”: 1 = yes;
0 = no) and lower actual IOS than ideal IOS with their part-
In this equation, b0 is the initial intercept for an outcome ner (i.e., “not close enough”: 1 = yes; 0 = no). Those who
Y. The sign and size of b1 represent the direction and slope of reported no actual–ideal IOS discrepancies with their part-
the segment of the equation linking negative actual–ideal ners were the reference group. These binary predictor vari-
IOS discrepancies (i.e., feeling not close enough) to study ables were entered into logistic regression models to obtain
outcomes. The sign and size of b2 represent the direction and the odds ratios (ORs) regarding relationship dissolution
slope of the segment of the equation linking positive actual– observed over a period of 1 year associated with each type of
ideal IOS discrepancies (i.e., feeling too close) to study out- actual–ideal IOS discrepancy. These analyses controlled for
comes. To examine the extent to which the association Time 1 actual IOS, gender, marital status, and length of
between actual–ideal IOS discrepancies and study outcomes relationship.
remained robust and statistically significant above and
beyond actual levels of closeness, actual IOS was controlled
for in all segmented regression models along with gender, Results
marital status, and length of relationship. It was not possible Preliminary and Descriptive Analyses
to include ideal IOS in these analyses because its inclusion
would produce a linear dependency within the models (i.e., Means, standard deviations, and bivariate Pearson correla-
participants’ ideal IOS is reflected in the difference between tions are presented in Table 1 for IOS-derived variables and
their actual IOS and actual–ideal IOS discrepancies). relational well-being and mental health outcomes across the
three time points in the study. As indicated by the mean
Repeated-Measures Mixed General Linear Models. To test the levels of actual and ideal IOS, participants’ actual experi-
study’s longitudinal hypotheses, change scores were com- ences of closeness were on average less than their ideal
puted reflecting the change in participants’ actual–ideal IOS levels of closeness, resulting in negative mean actual–ideal
discrepancies over Times 1, 2, and 3 of the current study. IOS discrepancies at each time point. Types of actual–ideal
Participants were classified into three groups for comparison IOS discrepancies at initial participation were not evenly
based on whether the absolute value of their actual–ideal distributed in the sample: 57.1% of the sample reported
IOS discrepancies (a) diminished over time, (b) increased negative discrepancies (i.e., actual IOS < ideal IOS); 37.6%
over time, or (c) remained the same over time. This between- reported no discrepancies (i.e., actual IOS = ideal IOS); and
subjects factor was examined within repeated-measures 5.3% reported positive discrepancies (i.e., actual IOS >
mixed general linear models, which examined whether ideal IOS).
change in the relational well-being and mental health out- Bivariate correlations demonstrated robust and statisti-
comes over time was dependent on the type of change in cally significant associations between actual IOS and all
participants’ actual–ideal IOS discrepancies. To examine the study outcomes, such that greater closeness was associated
extent to which changes in actual–ideal IOS discrepancies with higher levels of relationship satisfaction and commit-
corresponded to changes in study outcomes above and ment, as well as fewer break-up thoughts and depressive
beyond actual levels of IOS, average levels of actual IOS symptoms. Associations between ideal IOS and relational
across the three time points were included as a covariate well-being and mental health outcomes—although statisti-
given covariates in repeated-measured mixed general linear cally significant at times—were not robust in any instance
models are required to be time invariant. These models were given their corresponding effect sizes were uniformly small
also controlled for gender, and Time 1 marital status and (i.e., Pearson rs ~ .10; Cohen, 1992). Actual–ideal IOS dis-
length of relationship. Only participants who completed all crepancies demonstrated robust and statistically significant
waves of the study and reported being with the same rela- bivariate associations with relational well-being and mental
tionship partner at each time point were included in longitu- health. However, these bivariate relationships are not directly
dinal analyses. Missing data on variables within follow-up interpretable given both the scaling of the discrepancy score

Downloaded from psp.sagepub.com by guest on February 17, 2013


Frost and Forrester 7

Table 1. Descriptive Statistics and Correlations Between IOS Predictors and Relational Well-Being and Mental Health Outcomes

Bivariate correlations
Actual IOS Ideal IOS Actual–ideal IOS discrepancy
Time 1 Time 2 Time 3 Time 1 Time 2 Time 3 Time 1 Time 2 Time 3 M SD
Relationship satisfaction
  Time 1 .68*** .49*** .45*** .15*** .13*** .05 .61*** .41*** .45*** 17.89 4.89
  Time 2 .46*** .71*** .49*** .06 .19*** .04 .44*** .60*** .50*** 18.11 4.71
  Time 3 .43*** .51*** .69*** .08 .10* .10* .39*** .45*** .66*** 18.31 4.57
Commitment
  Time 1 .44*** .27*** .28*** .17*** .12*** .06 .36*** .19*** .26*** 6.00 0.92
Break-up thoughts
  Time 1 −.57*** −.37*** −.33*** −.12*** −.10** −.07 −.52*** −.31*** −.31*** 1.75 1.01
  Time 2 −.34*** −.59*** −.37*** −.08* −.15*** −.03 −.30*** −.51*** −.37*** 1.70 0.97
  Time 3 −.30*** −.43*** −.56*** −.05 −.11*** −.10* −.27*** −.35*** −.53*** 1.61 0.94
Depression
  Time 1 −.32*** −.24*** −.24*** −.08** −.08* −.05 −.28*** −.20*** −.22*** 8.62 5.99
  Time 2 −.22*** −.37*** −.29*** −.02 −.04 −.05 −.22*** −.35*** −.27*** 8.48 6.13
  Time 3 −.20*** −.30*** −.34** −.05 −.11** −.10* −.17*** −.23*** −.30*** 8.03 6.19
M 4.37 4.41 4.43 5.32 5.24 5.27 −.96 −.84 −.83  
SD 1.34 1.30 1.28 0.87 0.88 0.83 1.29 1.26 1.21  
Note: IOS = inclusion of other in self.
*p < .05. **p < .01. ***p < .001.

and the unequal representation of the types of actual–ideal feeling increasingly “too close for comfort” (IOS discrep-
IOS discrepancies in the dataset, and are thus better under- ancy > 0), their reported relational well-being and mental
stood via the results of the segmented regression analyses health decreased in kind. Similarly, feeling increasingly “not
reported below. close enough” (IOS discrepancy < 0) was associated with
corresponding decreases in relational well-being and mental
health. These associations remained robust and statistically
Associations Between IOS Discrepancies and significant controlling for gender, relationship length, and
Relational Well-Being and Mental Health marital status. Regression equations constructed from the
Outcomes values presented in Table 2 were used to compute adjusted
predicted scores for each outcome and visually plot the
Complete results of the segmented linear regression models resulting segmented linear relationships, as presented in
examining associations between actual–ideal IOS discrepan- Figure 1.
cies and relational well-being and mental health are pre-
sented in Table 2. In summary, these analyses demonstrated
that participants’ actual ratings of IOS were robust and sta- Type of Change in IOS Discrepancies as a
tistically significant predictors of relationship satisfaction Predictor of Change in Relational Well-Being
and commitment in a positive direction and break-up and Mental Health Over Time
thoughts and depression in a negative direction. However,
coefficients corresponding to the hypothesized segmented Participant representation across the three levels of the
association between IOS discrepancy scores and relational between-subjects factor of type of IOS discrepancy change
well-being and mental health outcomes were robust and were as follows: 125 participants’ actual–ideal IOS discrep-
statistically significant above and beyond the effects of ancies increased over time (i.e., demonstrated a net increase
actual IOS. These segmented equations indicated that opti- in a positive or negative direction, moving away from 0);
mal levels of relational well-being (i.e., highest levels of 126 participants’ actual–ideal IOS discrepancies diminished
relationship satisfaction, commitment, and lowest levels of over time (i.e., demonstrated a net decrease in a positive or
break-up thoughts) and mental health (i.e., lowest levels of negative direction, moving toward 0); and 260 participants’
depression) occurred among those who experienced no dis- actual–ideal IOS discrepancies remained unchanged over
crepancy between their actual and ideal levels of IOS with the 3 yearly assessments (i.e., were the same at Times 1 and
their partners (IOS discrepancy = 0). As participants reported 3). Means for each of these three groups defined by type of

Downloaded from psp.sagepub.com by guest on February 17, 2013


8 Personality and Social Psychology Bulletin XX(X)

Table 2. Segmented Associations Between IOS Discrepancies and Relational Well-Being and Mental Health Outcomes

Relationship satisfaction Commitment Break-up thoughts Depression


  B SE β p rs B SE β p rs B SE β p rs B SE β p rs
Intercept 9.44 0.60 .00 4.96 0.15 .00 2.45 0.14 .00 9.85 1.01 .00  
Female 0.61 0.20 .05 .00 .05 0.28 0.05 .13 .00 .12 −0.07 0.05 −.03 .17 −.03 0.24 0.34 .02 .47 .02
Relationship length −0.04 0.01 −.06 .00 −.05 0.01 0.00 .10 .00 .09 0.00 0.00 −.04 .10 −.03 0.01 0.02 .02 .55 .01
Married −0.02 0.20 .00 .92 .00 0.47 0.05 .25 .00 .21 −0.13 0.05 −.06 .01 −.05 −0.17 0.34 −.01 .62 −.01
Actual IOS 1.42 0.11 .39 .00 .22 0.16 0.03 .23 .00 .13 −0.21 0.03 −.27 .00 −.16 −0.65 0.18 −.15 .00 −.08
Negative IOS 1.65 0.12 .43 .00 .23 0.21 0.03 .30 .00 .16 −0.33 0.03 −.43 .00 −.22 −1.21 0.21 −.26 .00 −.14
discrepancy (≤0)
Positive IOS −3.94 0.28 −.29 .00 −.24 −0.61 0.07 −.23 .00 −.20 0.75 0.07 .27 .00 .22 3.61 0.47 .22 .00 .18
discrepancy (>0)
F 321.27 114.54 178.83 42.45
R2 .54 .35 .40 .14
Note: IOS = inclusion of other in self.

IOS discrepancy change on relational well-being and mental relationship satisfaction over time evidenced a dependency
health outcomes are displayed in Figure 2 for each of the 3 on average levels of actual IOS. The effects of gender, rela-
yearly assessments. tionship length, and marital status were also accounted for,
Results of the repeated-measures mixed general linear although they evidenced no influence on the rate of change
models examining the degree to which the type of change in in relational well-being or mental health over time.
participants’ actual–ideal IOS discrepancies was associated
with changes in relational well-being and mental health over
time are presented in Table 3. Only relationship satisfaction Actual–Ideal IOS Discrepancies as Predictors
demonstrated statistically significant change over time. of Relationship Dissolution
However, despite the lack of change in average levels of Results of the binary logistic regression models testing the
break-up thoughts and depression over time, all analyses extent to which IOS discrepancies at Time 1 were predictive
indicated uniform and statistically significant interactions of relationship dissolution at follow-up assessment are pre-
between the within-subjects factor of time and the between- sented in Table 4. Of the original Time 1 sample retained at
subjects factor of type of change in actual–ideal discrepan- the first yearly follow-up, 7.2% (n = 139) of participants
cies. In other words, the direction of change over time in reported ending their relationships with their previous part-
relational well-being and mental health depended on the type ners. Individuals who felt “not close enough” to their partners
of change participants experienced within their actual–ideal (i.e., had negative actual–ideal IOS discrepancies) at Time 1
IOS discrepancies. These interactions (see means plots in were approximately twice as likely to break up with their
Figure 2) indicated that, as hypothesized, those individuals partners at follow-up than individuals who experienced no
whose actual experiences of IOS became closer to their ideal discrepancy in actual and ideal IOS at Time 1. Similarly, indi-
IOS over time demonstrated increases in relational well- viduals who felt “too close” to their partners (i.e., had positive
being (i.e., increased satisfaction and decreased break-up actual–ideal IOS discrepancies) at Time 1 were approximately
thoughts) and mental health (i.e., fewer depressive symp- twice as likely to break up with their partners over 1 year than
toms) across the three time points. Those whose actual expe- individuals who experienced no discrepancy in actual and
riences of IOS moved further away from their ideal over ideal IOS at Time 1; however, this association was not statisti-
time demonstrated decreases in relational well-being (i.e., cally significant. These findings were obtained from models
decreased satisfaction and increased break-up thoughts) and controlling for actual IOS at Time 1, which demonstrated a
mental health (i.e., increased depressive symptoms) across protective effect on relationship dissolution along with rela-
the three time points. Finally, those whose actual–ideal IOS tionship length and marital status.
discrepancies evidenced no change over time demonstrated
relatively unchanging levels of relational well-being and
mental health, remaining virtually constant over the three Discussion
yearly assessments. The findings of the current study demonstrate the robust util-
The time by type of change in actual–ideal IOS discrep- ity of the construct of IOS in predicting relational well-
ancy interactions remained robust and statistically signifi- being, stability, and mental health among a diverse sample
cant even when the effects of average actual IOS were of individuals in romantic relationships. Consistent with
accounted for in the models. However, only changes in previous theory and research (Aron et al., 2004; Tsapelas

Downloaded from psp.sagepub.com by guest on February 17, 2013


Downloaded from psp.sagepub.com by guest on February 17, 2013
9
Figure 1. Actual–ideal IOS discrepancies as a predictor of relational well-being and mental health
Note: IOS = inclusion of other in self.
10 Personality and Social Psychology Bulletin XX(X)

Figure 2. Change in relational well-being and mental health as a function of type of change in actual–ideal IOS discrepancies
Note: IOS = inclusion of other in self.

et al., 2009), experiences of closeness with one’s relationship ings indicate that individuals’ actual experiences of IOS in
partner—as evidenced via greater inclusion of one’s partner their relationships are only part of a broader experience of
within one’s own self-concept—is indicative of heightened closeness, which must be accounted for in attempts to fully
relational well-being across a variety of indicators, including understand the association between closeness and well-
increased satisfaction, commitment, and fewer thoughts being in romantic relationships. Specifically, these findings
about ending one’s relationship. However, this study’s find- indicate that optimal levels of relational well-being and

Downloaded from psp.sagepub.com by guest on February 17, 2013


Frost and Forrester 11

Table 3. Change in Relational Well-Being and Mental Health Over Time As a Function of Type of Change in IOS Discrepancies

Relationship satisfaction Break-up thoughtsa Depressiona


Within-subjects effects F df η2p F df η2p F df η2p
Time 5.16** 2 0.01 1.34 1.97 0.00 0.19 2.00 0.00
Time × Change in actual– 19.55*** 4 0.07 13.86*** 3.93 0.05 3.41** 4.00 0.01
ideal IOS discrepancy
Time × Actual IOS 4.52** 2 0.01 1.10 1.97 0.00 0.80 2.00 0.00
Time × Female 1.61 2 0.00 0.55 1.97 0.00 1.83 2.00 0.00
Time × Relationship length 0.27 2 0.00 1.34 1.97 0.00 0.86 2.00 0.00
Time × Married 0.68 2 0.00 2.06 1.97 0.00 1.65 2.00 0.00
Note: IOS = inclusion of other in self.
a
Sphericity not assumed based on results of Mauchly’s Test of Sphericity, df adjusted using Huynh–Feldt correction.
*p < .05. **p < .01. ***p < .001.

Table 4. Actual–Ideal IOS Discrepancies as Predictors of relationships match how close they would like to feel to their
Relationship Dissolution partners.
Furthermore, this study provides empirical evidence that
B p OR 95% CI for OR
the abatement of discrepancies between experienced and
Intercept 0.24 .71 1.28   idealized levels of closeness in relationships can lead to
Negative IOS discrepancy 0.73 .04 2.07 [1.04, 4.14] improved relational well-being, stability, and mental health
Positive IOS discrepancy 0.77 .15 2.16 [0.75, 6.18] over time. Specifically, the findings from the current study
Actual IOS −0.33 .00 0.72 [0.58, 0.88] show that if experiences of closeness with a partner move in
Female −0.12 .66 0.89 [0.52, 1.52] line with ideal experiences of closeness, individuals will
Relationship length −0.21 .00 0.81 [0.74, 0.88] experience significant and substantial increases in relational
Married −0.99 .00 0.37 [0.19, 0.72] well-being and mental health over time. However, the oppo-
Note: IOS = inclusion of other in self; OR = odds ratio; CI = confidence site is true under conditions where experiences of closeness
interval. move further away from ideal levels of closeness over time,
resulting in decreased relational well-being and mental
health. Those whose discrepancies remained constant over
mental health are achieved when individuals’ actual experi- time did not experience much change in their relational well-
ences of closeness in their relationships match their ideal- being or mental health over time. It is important to note that
ized levels of closeness with their relationship partners. This the effects of participants’ closeness discrepancies on
finding provides further evidence for existing theory and changes in relational well-being and mental health persisted
research on the phenomena of feeling “too close” in roman- regardless of how close they actually felt to their partner.
tic relationships (Mashek & Sherman, 2004) by providing an This novel finding demonstrates the impact and importance
overarching explanatory model of closeness discrepancies of closeness discrepancy reduction—above and beyond the
and their impact on individual health and romantic relation- promotion of closeness itself—in efforts to improve rela-
ship quality. tional well-being and mental health over time.
The current study extends this line of research by demon- Even further, experiencing a discrepancy in actual–ideal
strating that individuals who feel “too close” and “not close levels of IOS was associated with an increased likelihood of
enough” to their partner—as operationalized via actual–ideal relationship dissolution, above and beyond actual experi-
IOS discrepancies—fare increasingly less well as the size of ences of closeness. Even though feeling “not close enough”
their IOS discrepancies increase, regardless of what their to a partner was significantly associated with dissolution rates
actual experiences of IOS may be. This finding was consis- and feeling “too close” to a partner was not significantly asso-
tent not only across a variety of indicators of relational well- ciated with dissolution rates, it is important to note that the
being, but extended to the domain of mental health as well. effect sizes for both were nearly equal (ORs = 2.07 and 2.16,
Thus, accounting for individuals’ actual experiences of respectively). Thus, it is not possible to conclude that there
closeness with their partners in relation to their desired levels were differential effects of the type of closeness discrepancy
of closeness is necessary to maximize the success of attempts on relationship dissolution. The magnitude and direction of
to understand the effect of closeness on relational well-being the effects associated with both types of closeness discrepan-
and mental health in relationships. What seems to matter cies indicate that individuals with any type of discrepancy
most for both relational well-being and mental health is not between their actual and idealized experiences of closeness
only how close people actually feel to their partners, but are at substantial risk for relationship dissolution compared
whether people’s actual experiences of closeness in their with individuals who experience no discrepancy.

Downloaded from psp.sagepub.com by guest on February 17, 2013


12 Personality and Social Psychology Bulletin XX(X)

Accounting for the existence and size of the discrepancy nature and causality cannot be inferred. Second, the sample
between individuals’ actual and ideal levels of closeness in is biased by self-selection and is not random or representa-
their relationships is paramount in attempts to explain its tive of the U.S. and Canadian populations from which it was
effects on relational well-being and mental health. Less impor- drawn. Although research indicates men and women do not
tant in this regard is the nature or direction of the discrepancy differ in their experiences of IOS discrepancies (Frost &
(Aron et al., 2004; Mashek & Sherman, 2004). In other words, Eliason, in press) and the effects of gender were controlled
optimal levels of relational well-being and mental health exist for in all analyses, the underrepresentation of men in the
when actual–ideal closeness discrepancies do not. Feeling study did not permit an examination of potential gender dif-
both “too close” and feeling “not close enough” to a relation- ferences in the associations between IOS discrepancies and
ship partner have detrimental effects on relational well-being study outcomes. Despite these limitations, it is important to
and mental health. The deleterious implications posed by the note that the size and diversity of the present sample pro-
encroachment on personal identity and control likely associ- vides a complement to the existing research on closeness
ated with feeling “too close” to one’s partner seem to be equal and IOS, which has been conducted primarily with under-
to those posed by the lack of closeness experienced in the form graduate student samples or geographically restricted sam-
of less-than-desirable levels of closeness. ples of mostly married individuals.
Understanding individuals’ experiences of closeness in Significant amounts of attrition were observed that were
their romantic relationships via their reports of actual–ideal likely due to the challenges of conducting longitudinal
IOS discrepancies has tremendous potential not only for future research on the Internet and the minimal incentive structure
research on closeness, IOS, and health and well-being in close used. However, there were very few differences between
relationships, but for clinical and counseling work with cou- retained participants and those lost to follow-up. As a result
ples and individuals as well. The present findings highlight of higher attrition among those who were more depressed,
necessary next steps in research on IOS and relational well- the effect sizes of coefficients linking IOS discrepancies
being and mental health, which have thus far overly relied on with mental health are likely attenuated and thus underesti-
single indicators of IOS in relational contexts. Measuring mated in the current study. Also, associations between IOS
actual–ideal IOS discrepancies takes seriously individuals’ discrepancies and dissolution are likely attenuated, given
own desires for closeness in relationships and avoids forcing individuals who break up with their partners may be less
couples’ lived experiences into a researcher-imposed expecta- motivated to continue participation in a study about relation-
tion of what their relationships “should” look like (e.g., health- ships than those who stay together. In addition, the lack of
ier relationships = more closeness). It is this kind of statistical significance associated with the robust OR linking
constructivist approach to understanding closeness in relation- positive actual–ideal IOS discrepancies to dissolution was
ships that, as the current study demonstrates, provides an likely a result of the low frequency of positive IOS discrep-
improved explanation of relational functioning over what ancies in the sample, as well as the difficulty in observing
static measures of closeness can explain by themselves. relationship dissolution in survey studies with yearly assess-
Accounting for actual–ideal IOS discrepancies in the clin- ments (Gottman, 1993).
ical intake process and assessment of couples counseling The study used a limited set of indicators of relational
interventions may thus prove to be a useful tool for clinicians well-being and mental health; thus, future work is needed to
seeking to improve relational well-being and mental health examine the extent to which the current findings generalize to
in relational contexts. Having couples reflect on their current other indicators of relationship quality and health. Given
and ideal levels of closeness on multiple occasions can assist depressive symptoms are just one indicator of a broader con-
clinicians and counselors in their abilities to target interven- struct of mental health, future studies should include addi-
tions aimed at eliminating closeness discrepancies and prob- tional measures of negative (e.g., anxiety) and positive
lematic experiences of enmeshment and intrusiveness, functionality (e.g., psychological well-being). Also, the cur-
thereby improving the health and well-being of their patients rent study examined only individuals in romantic relation-
and clients (e.g., Green & Werner, 1996). Furthermore, the ships. Future studies should explore how processes related to
present findings suggest that efforts focused on improving closeness and IOS discrepancies in relationships play out
relational well-being and mental health in couples should not within couples. For example, the use of dyadic analyses will
limit themselves to the promotion of closeness but instead allow for the examination of research questions surrounding
focus on the alignment of individuals’ experiences of close- partner effects of IOS discrepancies on relational well-being
ness and their ideal levels of closeness with their partners. and mental health (e.g., How do husbands’ IOS discrepancies
impact their wives’ depression?). Additional research is also
needed to explore the implications of actual–ideal IOS dis-
Limitations and Suggestions for Future crepancies in other types of close relationships given the cen-
Research trality of closeness to positive relations with others. Important
The findings of the current study should be interpreted in research questions can be posed regarding the impact
light of some limitations. Although the study reports on lon- of actual–ideal IOS discrepancies on a multitude of out-
gitudinal findings, the data are nonetheless correlational in comes across a variety of relationship contexts such as job

Downloaded from psp.sagepub.com by guest on February 17, 2013


Frost and Forrester 13

performance among coworkers, familial conflict in parent– Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close
child relationships, and satisfaction with clinical care in relationships as including other in the self. Journal of Person-
patient–provider interactions. ality and Social Psychology, 60, 241-253. doi:10.1037//0022-
3514.60.2.241
Aron, A., McLaughlin-Volpe, T., Mashek, D., Lewandowski, G.,
Conclusion Wright, S. C., & Aron, E. N. (2004). Including others in the
The present study’s focus on actual–ideal closeness discrep- self. European Review of Social Psychology, 15, 101-132.
ancies deepens psychological understandings of the effects doi:10.1080/10463280440000008
of closeness on relationship quality and individual health in Aron, A., Melinat, E., Aron, E., Vallone, R. E., & Bator, R. J.
the context of romantic relationships. Accounting for indi- (1997). The experimental generation of interpersonal closeness:
viduals’ actual experiences of closeness in romantic relation- A procedure and some preliminary findings. Personality and
ships is necessary but not sufficient in effectively explaining Social Psychology Bulletin, 23, 365-378.
their relational well-being, stability, and mental health. Aron, A., Norman, C. C., & Aron, E. N. (2001). Shared self-
Psychological inquiries into the phenomenon of closeness expanding activities as a means of maintaining and enhancing
must additionally account for individuals’ idealized notions close romantic relationships. In J. Harvey & A. Wenzel (Eds.),
of closeness in their relationships with their partners, given Close romantic relationships: Maintenance and enhancement
discrepancies between actual and ideal experiences of close- (pp. 47-66). Mahwah, NJ: Lawrence Erlbaum.
ness can be detrimental to relational well-being and mental Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles
health, both in the present and over time. Future work must among young adults: A test of a four-category model. Journal
take seriously where individuals’ experiences of closeness of Personality and Social Psychology, 61, 226-244.
with their partners are in relation to where they would ide- Berscheid, E. (2010). Love in the fourth dimension. Annual Review
ally like them to be to more fully understand and promote of Psychology, 61, 1-25.
their experiences of closeness, well-being, health, and rela- Birnbaum, G., Reis, H., Mikulincer, M., Gillath, O., & Orpaz, A.
tional stability over time. (2006). When sex is more than just sex: Attachment orienta-
tions, sexual experience, and relationship quality. Journal of
Acknowledgment Personality and Social Psychology, 91, 929-943.
The authors thank members of the Psychology of Sexuality and Brennan, K. A., & Shaver, P. R. (1995). Dimensions of adult attach-
Relationships Lab and the Columbia and NYU Couples Labs for ment, affect regulation, and romantic relationship functioning.
their comments on presentations of the data reported in this article. Personality and Social Psychology Bulletin, 21, 267-283.
Brunell, A. B., Pilkington, C. J., & Webster, G. D. (2007). Percep-
Declaration of Conflicting Interests tions of risk in intimacy in dating couples: Conversation and
The author(s) declared no potential conflicts of interest with relationship quality. Journal of Social & Clinical Psychology,
respect to the research, authorship, and/or publication of this 26, 92-119.
article. Butzer, B., & Campbell, L. (2008). Adult attachment, sexual satis-
faction, and relationship satisfaction: A study of married cou-
Funding ples. Personal Relationships, 15, 141-154.
The author(s) received no financial support for the research, Cohen, J. (1992). A power primer. Psychological Bulletin, 112,
authorship, and/or publication of this article. 155-159.
Fitzsimons, G. M., & Kay, A. C. (2004). Language and interper-
References sonal cognition: Causal effects of variations in pronoun usage
Agnew, C. R., Loving, T. J., Le, B., & Goodfriend, W. (2004). Think- on perceptions of closeness. Personality and Social Psychology
ing close: Measuring closeness as perceived self-other inclusion. Bulletin, 30, 547-557. doi:10.1177/0146167203262852
In D. J. Mashek & A. Aron (Eds.), Handbook of closeness and Fletcher, G. J. O., & Simpson, J. A. (2000). Ideal standards in close
intimacy (pp. 103-116). Mahwah, NJ: Lawrence Erlbaum. relationships: Their structure and functions. Current Directions
Agnew, C. R., Van Lange, P. A. M., Rusbult, C. E., & Langston, in Psychological Science, 9, 102-105.
C. A. (1998). Cognitive interdependence: Commitment and the Fletcher, G. J. O., Simpson, J. A., Thomas, G., & Giles, L.
mental representation of close relationships. Journal of Person- (1999). Ideals in intimate relationships. Journal of Person-
ality and Social Psychology, 74, 939-954.doi:10.1037/0022- ality and Social Psychology, 76, 72-89. doi:10.1037/0022-
3514.74.4.939 3514.76.1.72
Andresen, E. M., Malmgren, J. A., Carter, W. B., & Patrick, D. L. Frost, D. M., & Eliason, M. (in press). Challenging the assump-
(1994). Screening for depression in well older adults: Evalua- tion of fusion in female same-sex relationships. Psychology of
tion of a short form of the CES-D. American Journal of Preven- Women Quarterly.
tive Medicine, 10, 77-84. Frost, D. M. (in press). The narrative construction of intimacy and
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in affect in relationship stories: Implications for relationship qual-
the Self scale and the structure of interpersonal closeness. Jour- ity and mental health. Journal of Social and Personal Relation-
nal of Personality and Social Psychology, 63, 596-612. ships. DOI: 10.1177/0265407512454463

Downloaded from psp.sagepub.com by guest on February 17, 2013


14 Personality and Social Psychology Bulletin XX(X)

Funk, J. L., & Rogge, R. D. (2007). Testing the ruler with item McDowell, I., & Newell, C. (1996) Measuring health: A guide to
response theory: Increasing precision of measurement for rating scales and questionnaires (2nd ed.). New York, NY:
relationship satisfaction with the Couples Satisfaction Index. Oxford University Press.
Journal of Family Psychology, 21, 572-583. Patrick, H., Knee, C. R., Canevello, A., & Lonsbary, C. (2007). The
Goodboy, A. K., & Booth-Butterfield, M. (2009). Love styles and role of need fulfillment in relationship functioning and well-
desire for less closeness in romantic relationships. Psychologi- being: A self-determination theory perspective. Journal of Per-
cal Reports, 105, 191-197. sonality and Social Psychology, 92, 434-457.
Gottman, J. M. (1993). A theory of marital dissolution and stability. Radloff, L. S. (1977). The CES-D Scale: A self-report depression
Journal of Family Psychology, 7, 57-75. scale for research in the general population. Applied Psycho-
Green, R., & Werner, P. D. (1996). Intrusiveness and closeness-care- logical Measurement, 1, 385-401.
giving: Rethinking the concept of family “Enmeshment.” Family Reis, H. T., & Aron, A. (2008). Love: What is it, why does it matter,
Process, 35, 115-136. doi:10.1111/j.1545-5300.1996.00115.x and how does it operate? Perspectives on Psychological Sci-
Hassebrauck, M., & Fehr, B. (2002). Dimensions of relationship ence, 3, 80-86.
quality. Personal Relationships, 9, 253-270. doi:10.1111/1475- Reis, H. T., Collins, W. A., & Berscheid, E. (2000). The rela-
6811.00017 tionship context of human behavior and development.
Hazan, C., & Shaver, P. R. (1994). Attachment as an organizational Psychological Bulletin, 126, 844-872. doi:10.1037/0033-
framework for research on close relationships. Psychological 2909.126.6.844
Inquiry, 5, 1-22. Roberts, R. E., & Vernon, S. W. (1983). The Center for Epidemio-
Heidrich, S. M., & Powwattana, A. (2004). Self-discrepancy and logic Studies Depression Scale: Its use in a community sample.
mental health in older women with chronic illnesses. Journal of American Journal of Psychiatry, 140, 41-46.
Adult Development, 11, 251-259. Robins, G., & Boldero, J. (2003). Relational discrepancy theory:
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and The implications of self-discrepancy theory for dyadic relation-
affect. Psychological Review, 94, 319-340. ships and for the emergence of social structure. Personality and
Holt-Lunstad, J., Smith, T. B., & Layton, J. B. (2010). Social rela- Social Psychology Review, 7, 56-74.
tionships and mortality risk: A meta-analytic review. PLoS Robles, T. F., & Kiecolt-Glaser, J. K. (2003). The physiology of
Medicine, 7, e1000316. doi:10.1371/journal.pmed.1000316 marriage: Pathways to health. Physiology & Behavior, 79,
Kashdan, T. B., Volkmann, J. R., Breen, W. E., & Han, S. (2007). 409-416. doi:10.1016/S0031-9384(03)00160-4
Social anxiety and romantic relationships: The costs and ben- Roth, P. L. (1994). Missing data: A conceptual review for applied
efits of negative emotion expression are context dependent. psychologists. Personnel Psychology, 47, 537-560.
Journal of Anxiety Disorders, 21, 475-492. Rusbult, C. E., & Van Lange, P. A. M. (2003). Interdependence,
Keele, L. J. (2008). Semiparametric regression for the social sci- interaction, and relationships. Annual Review of Psychology,
ences. New York, NY: John Wiley. 54, 351-375.
Khaleque, A. (2004). Intimate adult relationships, quality of life Ryan, R., & Deci, E. (2000). Self-determination theory and the
and psychological adjustment. Social Indicators Research, 69, facilitation of intrinsic motivation, social development, and
351-360. doi:10.1007/s11205-004-1543-x well-being. American Psychologist, 55, 68-78.
Kiecolt-Glaser, J. K., Gouin, J.-P., & Hantsoo, L. (2010). Close rela- Schafer, J. L., & Olsen, M. K. (1998). Multiple imputation for mul-
tionships, inflammation, and health. Neuroscience & Biobehav- tivariate missing-data problems: A data analyst’s perspective.
ioral Reviews, 35, 33-38.doi:10.1016/j.neubiorev.2009.09.003 Multivariate Behavioral Research, 33, 545-571.
Le, B., Dove, N. L., Agnew, C. R., Korn, M. S., & Mutso, A. A. Scott, L., & O’Hara, M. W. (1993). Self-discrepancies in clinically
(2010). Predicting nonmarital romantic relationship dissolu- anxious and depressed university students. Journal of Abnor-
tion: A meta-analytic synthesis. Personal Relationships, 17, mal Psychology, 102, 282-287.
377-390. doi:10.1111/j.1475-6811.2010.01285.x Seider, B. H., Hirschberger, G., Nelson, K. L., & Levenson, R. W.
Lewandowski, G. W., Nardone, N., & Raines, A. J. (2010). The role (2009). We can work it out: Age differences in relational pro-
of self-concept clarity in relationship quality. Self and Identity, nouns, physiology, and behavior in marital conflict. Psychology
9, 416-433. doi:10.1080/15298860903332191 and Aging, 24, 604-613. doi:10.1037/a0016950
Lund, M. (1985). The development of investment and commit- Sheldon, K. M., Williams, G., & Joiner, T. (2003). Self-determination
ment scales for predicting continuity of personal relation- theory in the clinic: Motivating physical and mental health.
ships. Journal of Social and Personal Relationships, 2, 3-23. New Haven, CT: Yale University Press.
doi:10.1177/0265407585021001 Tsapelas, I., Aron, A., & Orbuch, T. (2009). Marital boredom now
Mashek, D. J., Le, B., Israel, K., & Aron, A. (2011). Wanting less predicts less satisfaction 9 years later. Psychological Science,
closeness in romantic relationships. Basic and Applied Social 20, 543-545. doi:10.1111/j.1467-9280.2009.02332.x
Psychology, 33, 333-345. Werner, P. D., Green, R. J., Greenberg, J., Browne, T. L., & McK-
Mashek, D. J., & Sherman, M. D. (2004). Desiring less closeness with enna, T. E. (2001). Beyond enmeshment: Evidence for the
intimate others. In D. J. Mashek & A. Aron (Eds.), Handbook of independence of intrusiveness and closeness-caregiving in
closeness and intimacy (pp. 343-356). Mahwah, NJ: Lawrence married couples. Journal of Marital & Family Therapy, 27,
Erlbaum 459-472.

Downloaded from psp.sagepub.com by guest on February 17, 2013

You might also like