Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Arch Sex Behav (2015) 44:1609–1620

DOI 10.1007/s10508-014-0426-8

ORIGINAL PAPER

Individual and Partner Correlates of Sexual Satisfaction and


Relationship Happiness in Midlife Couples: Dyadic Analysis of the
International Survey of Relationships
William A. Fisher • Kelly L. Donahue • J. Scott Long • Julia R. Heiman •

Raymond C. Rosen • Michael S. Sand

Received: 5 November 2012 / Revised: 18 June 2014 / Accepted: 10 August 2014 / Published online: 5 November 2014
 Springer Science+Business Media New York 2014

Abstract The current research reports a dyadic analysis of happiness included individuals’ reports of good health; frequent
sexual satisfaction, relationship happiness, and correlates of these kissing, cuddling, and caressing; frequent recent sexual activity;
couple outcomes in a large multinational dataset consisting of attaching importance to one’s own and one’s partner’s orgasm;
1,009midlifeheterosexualcouples(2,018individuals)recruitedin better sexual functioning; and greater sexual satisfaction, and once
Japan, Brazil, Germany, Spain, and the United States (Heiman again, even after controlling for individual-level effects, partners’
et al., 2011). Actor-Partner Interdependence Models (Kenny, reports of each of these correlates contributed significantly to
Kashy, & Cook, 2006) identified correlates of sexual satisfaction predicting and understanding individuals’ relationship happiness.
that included individuals’ reports of good health; frequent kissing, Interactions of individual and partner effects with participant
cuddling, and caressing; frequent recent sexual activity; attaching gender are also reported. Current results demonstrate empirically
importance to one’s own and one’s partner’s orgasm; better sexual that the partner‘‘matters’’to an individual’s sexual satisfaction and
functioning; and greater relationship happiness. Even after con- relationship happiness and indicate that a comprehensive under-
trolling for individual-level effects, partners’ reports of good standing of factors contributing to these couple outcomes requires
health; frequent kissing, cuddling, and caressing; frequent recent a couple-level research strategy. Partner effects, even when con-
sexual activity; attaching importance to one’s own and one’s trolling for individual effects, were consistently observed, and
partner’s orgasm; better sexual functioning; and greater relation- explanation of sexual satisfaction and relationship happiness
ship happiness contributed significantly to predicting and under- always depended on identifying and understanding mutual and
standingindividuals’sexualsatisfaction.Correlatesofrelationship concurrent individual and partner influences.

Keywords Sexual satisfaction  Relationship happiness 


W. A. Fisher (&) Dyadic analysis
Department of Psychology and Department of Obstetrics and
Gynaecology, University of Western Ontario, London,
ON N6A 5C2, Canada
e-mail: fisher@uwo.ca Introduction

K. L. Donahue
Sexual satisfaction and relationship happiness are assumed to be
Department of Pediatrics, Section of Adolescent Medicine, Indiana
University School of Medicine, Indianapolis, IN, USA valued goals for most human beings in most cultures, and these
outcomes are thought to be the joint result of individual and
J. S. Long partner characteristics and couple-level processes over the course
Department of Sociology, Indiana University, Bloomington, IN,
of relationships (Fincham & Beach, 2006; Laumann et al., 2006;
USA
Sprecher, Christopher, & Cate, 2006). Although sexual and
J. R. Heiman relationship satisfaction have been an enduring focus of research,
Kinsey Institute for Research in Sex, Gender, and Reproduction, work in this area has paradoxically concentrated on the search for
Indiana University, Bloomington, IN, USA
individual, not couple, characteristics that may prove to be cor-
R. C. Rosen  M. S. Sand relates of sexual satisfaction and relationship happiness, and has
New England Research Institutes, Watertown, MA, USA often examined such individual level correlates of sexual and

123
1610 Arch Sex Behav (2015) 44:1609–1620

relationship satisfaction among those in relatively new relation- and relationship happiness. For example, Rubin and Campbell
ships, rather than among individuals at midlife (Berscheid, 1999; (2011) found that individuals reporting increases in couple inti-
Rubin & Campbell, 2011; Kashy, Campbell, & Harris, 2006). macy from the previous day experienced increased relationship
Commenting on the emergence of what has been termed ‘‘rela- passion and positive affect the following day and that partners’
tionship science,’’ Berscheid (1999) noted that, ‘‘Reflecting psy- reports of increased intimacy the preceding day also contributed
chology’s individualistic orientation, virtually all of our significantly and independently to predicting individuals’ reports
methodologies and statistics are predicated on the individual as of relationship passion and positive affect. In related research,
the unit of analysis’’ (p. 261). Confirming the somewhat ironic Butzer and Campbell (2008) found that anxiously and avoidantly
individual focus of most relationship research, Kashy et al. (2006) attached married individuals reported lower levels of sexual sat-
surveyed published studies of relationships in five major journals isfaction and that avoidant attachment among partners contributed
andfoundthat70 %ofpeerreviewedstudiesofrelationshipswere significantly and independently to individuals’ lessened sexual
based on individual level—as opposed to couple-level—data. satisfaction. In additional relevant research, Birnbaum, Reis,
The same individual, not couple, level focus has characterized Mikulincer, Gillath, and Orpaz (2006) investigated the influence
research in the area of sexual dysfunction (Heiman, 2002). Thus, of individual and partner attachment orientation on the impact of
too little is known about the potentially mutual and concurrent positive and negative sexual experiences on relationship satis-
influence of the individual and the partner on sexual and rela- faction. Findings indicated that partners’ attachment anxiety and
tionship satisfaction, whether in couples in enduring relationships attachment avoidance contributed to individuals’ positive and
or in relationships of any duration. negative relationship-related behaviors on the days following
Efforts to understand sexual satisfaction and relationship hap- sexual interactions. For example, men with more anxiously
piness at the individual level of analysis have proven informative, attachedfemalepartnersreportedgreaterincreasesinrelationship
albeit limited in scope to half of the dyadic equation. For example, quality and lesser enactment of relationship damaging behav-
in a sample of middle-aged and older individuals from 29 coun- iors following days of reported sexual events, compared to men
tries, Laumann et al. (2006) found a correlation of 0.80 between with less anxiously attached partners. Women whose male part-
particular measures of individuals’ sexual satisfaction (e.g., ners were less avoidantly attached indicated a greater reduction
‘‘During the past 12 months, how physically pleasurable did you in relationship damaging behaviors following reported sexual
find your relationship with your partner to be?’’) and relationship events, compared to women whose male partner was more avoi-
satisfaction (e.g., ‘‘During the past 12 months, how emotionally dantly attached. Each of these studies—conducted at the couple
pleasurable did you find your relationship with your partner to level—serves to illustrate the mutual and concurrent influences
be?’’) (see also Sprecher et al., 2006). Gender differences in cor- of both the individual and the partner on sexual satisfaction and
relatesofsexualandrelationshipsatisfactionhavealsobeenfound. relationship happiness related outcomes, when a dyadic research
Heiman et al. (2011) reported that men’s sexual satisfaction was methodology is employed to explore these outcomes.
associated with men’s reports of good health, importance of The current study aimed to contribute to understanding the
partners’ orgasm, physical sexual intimacy (kissing and cuddling mutual influence of individual and partner characteristics on
often, being touched/caressed by partner often), good sexual sexual satisfaction and relationship happiness. To do so, we
functioning, and recent frequent sexual activity. Among women, investigated individual and partner correlates of these outcomes
sexual satisfaction was associated with relationship duration, in a large sample of midlife couples drawn from five nations. We
physical sexual intimacy (kissing and cuddling often, being tou- adopted a dyadic approach to analyzing a cluster of variables—
ched/caressed by partner often), good sexual functioning, and including good health, physical sexual intimacy, recent sexual
recent frequent sexual activity. In the same study, men’s rela- activity, and sexual functioning—that were found to influence
tionship happiness was found to be a function of reported good individuals’ reports of sexual and relationship happiness in Hei-
health, physical sexual intimacy, and good sexual functioning, man et al.’s (2011) individual-level analysis of the same dataset.
while women’s relationship happiness was related to relation- These variables were selected for study on the basis of existing
ship duration and good sexual functioning. Sexual satisfaction literature as promising conceptual and empirical correlates of
and relationship happiness have also been reported to change, sexual satisfaction and relationship happiness. We anticipated
concurrently, over time within individuals (Byers, 2005). Each finding that individual and partner reports of good health, good
of these findings concerning sexual satisfaction and relationship sexual functioning, recent physical intimacy, and recent sexual
happiness, however, is based upon individuals’ reports of their activitycontributeindependentlyandconcurrentlytoindividuals’
own characteristics and perceptions, and none of these studies sexual satisfaction and relationship happiness. We also explored
directly explored possible partner influences on individuals’ sex- gender-specific patterns of individual and partner effects on these
ual and relationship satisfaction. outcomes. The current study was thus designed to investigate the
Research that has adopted couple-level conceptual, methodo- potential mutual and concurrent influence of the individual and
logical, and statistical approaches has refined our understanding of the partner on individuals’ sexual and relationship satisfaction at
individual and partner factors that contribute to sexual satisfaction midlife,andmorebroadly,toaddtoexistingknowledgeregarding

123
Arch Sex Behav (2015) 44:1609–1620 1611

correlates of sexual and relationship satisfaction that has been thelanguagesofthecountriesinvolvedinthisresearch.Men’sand


drawn from individual-level analyses of smaller, younger, and women’s versions of the questionnaire were provided to couple
less heterogeneous samples. members, who were instructed not to discuss their answers with
their partner until their questionnaires were completed and
returned to the investigators. The sampling methodology of the
Method
International Survey of Relationships, initial refusal rates, and
refusal rates after potential participants had been informed of the
Participants
sexual content ofthesurvey, arereportedindetailbyHeiman etal.
(2011). This study was approved by the Institutional Review
The International Survey of Relationships was a cross-sectional,
Board of Indiana University.
self-administered, questionnaire-based survey of sexual satis-
faction and relationship happiness and potential correlates of
Measures
these outcomes, conducted with both members of intact midlife
heterosexual couples in five nations. Sample inclusion criteria
The International Survey of Relationships assessed self-reports
were men aged 40–70 years and their female partners, in a
of demographic characteristics, health status, sexual history,
committed relationship, married or cohabiting for at least 1 year.
sexual behavior, the importance of several aspects of sexual
Same-sex couples were not recruited due to constraints on
behavior, and sexual satisfaction and relationship happiness.
recruitment of sufficient sample size within the resources avail-
The instrument comprised 125 items, many of which were
able to this research.
drawn from standardized questionnaires and some of which
The current research was carried out in Brazil, Germany,
were developed specifically for this study.
Japan, Spain, and the U.S., sampling approximately 200 men in
Single item assessments of sexual satisfaction and relationship
the target age range and their female partners in each country. The
happiness were based on items from the International Index of
final sample included 1,009 couples (2,018 individuals), includ-
Erectile Function (Rosen et al., 1997), the Female Sexual Func-
ing 198 couples in Brazil, 198 in Germany, 207 in Japan, 199 in
tion Index (Rosen et al., 2000), and the Dyadic Adjustment Scale
Spain, and 207 in the U.S. The sample included men aged 39–70
(Spanier, 1976), respectively. Given the distribution of responses
(median age 55) and their female partners, who were aged 25–76
to items assessing sexual satisfaction and relationship happiness,
(median age 52). Couple relationship duration averaged 25 years
we dichotomized responses for sexual satisfaction into ‘‘not sat-
(median and mean) and ranged from 1 to 51 years (SD 11.34
isfied’’(comprising‘‘very dissatisfied,’’‘‘moderately dissatisfied,’’
years).1 Ninety percent of couples had children, 73.7 % of par-
‘‘dissatisfied,’’ and ‘‘equally satisfied and dissatisfied)’’ and ‘‘sat-
ticipants reported being in good health, and sexual activity had
isfied’’ (comprising ‘‘moderately satisfied’’ and ‘‘very satisfied’’).
occurred on average five to six times in the preceding 4 weeks.
Similarly, we dichotomized responses for relationship happiness
Procedure into‘‘not happy’’(comprising‘‘very unhappy,’’‘‘fairly unhappy,’’
and‘‘a little unhappy’’) and‘‘happy’’(comprising‘‘happy,’’‘‘very
Data collection was carried out by Synovate, an international happy,’’‘‘extremely happy,’’and‘‘perfect’’).2
healthcare market research firm, and employed sampling strat-
2
egies that were standard for each of the countries in this study. Dichotomization of the sexual satisfaction and relationship happiness
Participants in the U.S., Germany, and Spain were recruited by responses was based upon the skewed nature of the distributions of these
outcomes which is not uncommon in research on couple relationships. The
telephone using random digit dialing and market research panel
response distributions obtained were deemed to pose more challenges to
methods and were mailed self-administered questionnaires. In meaningful statistical analysis than the dichotomization of the sexual
Brazil and Japan, recruitment took place door-to-door, within satisfaction and relationship happiness outcomes and potential loss of
large cities in Brazil, and within randomly sampled locations for information that dichotomization entails. The dichotomous outcomes created
retain clearly anchored conceptual meaning that is faithful to the wording of
Japan. Self-administered questionnaires were provided to par-
the items, to our analytic aims, and to the interpretation of our results.
ticipants in Brazil and Japan. The research was described to ‘‘Sexuallysatisfied’’individuals(comprising‘‘verysatisfied’’and‘‘moderately
participants as‘‘…a study about people’s relationships and their satisfied’’individuals) are indeed reporting that they are sexually satisfied and
happiness with them. A number of questions deal with aspects of they differ from those who report that they are not sexually satisfied
(comprising‘‘very dissatisfied,’’‘‘moderately dissatisfied,’’‘‘dissatisfied,’’and
your personal relationship, including sexuality and sexual
‘‘equally satisfied and dissatisfied’’ individuals). Those who are happy with
experiences.’’ The survey was translated and back-translated in their relationships (comprising ‘‘happy,’’‘‘very happy,’’‘‘extremely happy,’’
and ‘‘perfect’’ ratings of relationship happiness) indeed report being happy
1
The interclass correlation for reported relationship duration within a with their relationships and differ from those who indeed report that they are
couple was 0.99. A total of 109 couples differed in their reported not happy with their relationships (comprising ‘‘very unhappy,’’‘‘fairly
relationship duration. For 81 of these couples, the reported difference unhappy,’’and‘‘a little unhappy’’). Balancing the challenges of the distri-
was 1 year, and the average difference among all discrepant couples was butions of sexual satisfaction and relationship happiness responses and
0.43 years. For discrepant couples, we used the average of couple’s the clearly anchored and meaningful dichotomization of these variables,
reports. we judged dichotomization to be the best solution.

123
1612 Arch Sex Behav (2015) 44:1609–1620

Table 1 Descriptions of measured variables


Variable label Survey question Response categories

Outcomes
Sexual satisfaction Over the past 4 weeks, how satisfied have you been 0 = Not satisfied (very dissatisfied; moderately
with your sexual relationship with your partner? dissatisfied, equally satisfied/dissatisfied)
1 = Satisfied (moderately satisfied, and very
satisfied)
Relationship happiness Which number best describes the degree of happiness, 0 = Unhappy (very unhappy, fairly unhappy, a
all things considered, in your relationship? little unhappy)
1 = Happy (happy, very happy, extremely happy,
and perfect)
Predictors
Good health Would you say your own health is excellent, good, fair, 0 = Poor/Fair
or poor? 1 = Excellent/Good
Relationship duration Constructed from years of marriage and years No. of years
cohabiting
Age at start of relationship Equal to age minus relationship duration No. of years
Importance of orgasm How important is it to you to reach orgasm yourself Scale from 1 to 10
during sex with your partner? (1 = Not at all; 10 = Very)
Importance of partner’s orgasm How important is it that your partner reaches orgasm Scale from 1 to 10
when you have sex together? (1 = Not at all; 10 = Very)
Kiss/cuddle often My partner and I kiss and cuddle each other: 0 = Very seldom/Seldom
1 = Often/Very often
Touched/caressed by partner often Over the past 4 weeks, how often have you been 0 = Not at all/Once or twice/Weekly
sexually touched and caressed by your partner 1 = 2–3 times a week/Daily or almost daily
(whether or not sexual activity occurred)?
Frequency of sex in last 4 weeks How many times during the past 4 weeks have you No. of times
engaged in sexual activity with your partner?
Lifetime sexual partners Over your lifetime, with how many partners have you No. of partners, both men and women
had sex (by sex we mean any genital contact and/or
stimulation)?
Four-week sexual functioning score Scale constructed separately by gender Average score on included items. Scale with higher
Men: Level of desire; frequency of sexual arousal; values indicating higher sexual functioning
frequency of erection; frequency of maintaining
erection after penetration; frequency of ejaculation;
frequency of premature ejaculation
Women: Level of desire; frequency of sexual arousal;
frequency of lubrication; frequency of orgasm

Potential correlates of sexual satisfaction and relationship Statistical Analysis


happiness were selected on the basis of the literatures in these
areas (Fincham & Beach, 2006; Laumann et al., 2006; Sprecher To exploretheinfluence ofboth individual (i.e.,actor)and partner
et al., 2006; Rubin & Campbell, 2011) and our own previous characteristics on actor sexual satisfaction and relationship hap-
research concerning individual-level correlates of sexual and piness, we ran a series of logistic regression actor-partner inter-
relationship happiness (Heiman et al., 2011). These potential dependence models (APIMs) (Kenny, Kashy, & Cook, 2006).
correlates included self-reported health, sexual history, impor- We used a random effects logit model, sometimes referred to as a
tance of several aspects of sexual intimacy, and sexual behavior ‘‘cluster-specific’’model, to account for the non-independence of
and sexual function. Item wording and scoring for these partners within a couple. This model is equivalent to a multi-level
potential correlates of sexual satisfaction and relationship hap- model with individuals at level 1 clustered within couples at level
piness appear in Table 1. Further details concerning these 2 (see Hosmer, Lemeshow & Sturdivant, 2013). We conducted
measures and research methodology are described in Heiman these analyses using Stata’s xtlogit command (StataCorp, 2013)
et al. (2011). to account for non-independence due to clustering.

123
Arch Sex Behav (2015) 44:1609–1620 1613

 
Each couple contributed two separate cases to the analyses— pij
one case in which the man was treated as the actor and the log ¼b0 þ aj þ baFemale aFemaleij þ baX aXij
1  pij
woman as the partner, and a second case in which the woman
þ bduration durationj þ bcountry countryj
was treated as the actor and the man as the partner. All actor- and
partner-level variables are within-couple variables that could ð1Þ
differ between individuals in a couple (e.g., an actor could report In Model 1, aj is assumed to be distributed normally with mean 0
that he or she was not in good health, while the partner might and variance r2.
report being in good health). The intercept was treated as a Model 2 estimates the effect of the partners’ response for the
random effect varying across couples (i.e., individuals were predictor on the outcome for the actor. Model 2 is equivalent to
clustered within couples who varied in their expected likelihood the model shown in Eq. (1) except that the partners’ response for
of reporting each outcome). The couple’s relationship duration the predictor was used in place of the actor’s own response.
and country of residence were included as between-couple Model 3 estimates both actor and partner effects for each
variables since values were the same for both members of the predictor on each outcome. Defining pXij as the value of a specific
couple. Dummy variables were created for country, with the predictor for the partner in couple j, Model 3 is:
United States excluded as the reference category in all analyses.
The slope for each predictor was treated as a fixed effect, as  
pij
couples do not have enough cluster members to allow for esti- log ¼ b0 þ aj þ baFemale aFemaleij
1  pij
mation of the variation in slopes across clusters (Kenny et al.,
2006). þ baX aXij þ bpX pXij
Listwise deletion of missing data was used in all models. Only þ bdur durationj þ bcountry countryj
cases with available data on all actor-level variables, partner-
level variables, and couple variables were used to estimate the Partners’ gender is not included in the model since it is perfectly
models. To measure within-couple similarity, we estimated collinear with the actor’s gender (i.e., if the actor is female, the
Pearson correlation coefficients for each of the predictors and partner must be male, in the current sample).
outcomes included in the current study (McMahon, Pouget, & Because each couple was comprised of a man and a woman
Tortu,2006).ThePearsoncorrelationcoefficient hasbeenshown potentially distinguishable by their role in the couple (Kenny
to be a reliable estimate of interdependence for binary variables et al., 2006), we also allowed the effects of the predictor to vary
in dyadic datasets with a sufficient number of clusters (Zou & by gender. Model 4 allows the effect of a regressor for an actor to
Donner, 2004). differ if the actor is male than if the actor is female; similarly, the
effects of the actor’s partner can differ by gender. For example,
the coefficient for aXij is bafX if the actor is female and bamX if the
Actor- and Partner-Level Influences on Sexual Satisfaction actor is male, and similarly for the actor’s partner. Defining
and Relationship Happiness aMale as 1 if the actor is male and else 0, with similar variables
for the partner, Model 4 is:
We first estimated the actor effect of each predictor on each out-
 
come, referred to as Model 1. For each predictor, the log odds of a pij
‘‘yes’’response to the outcome (e.g., the log odds of the actor being log ¼b0 þ aj þ baFemale aFemaleij
1  pij
sexually satisfied) were predicted by an intercept, the actor’s 
þ bafX aXij  aFemaleij
gender,the actor’s responsetothe predictor in question (e.g.,being 
in good health), and five couple-level variables (relationship þ bamX aXij  aMaleij

duration and four dummy codes representing country of residence þ bpfX pXij  pFemaleij
with the United States as the excluded category). A couple-level 
þ bpmX pXij  pMaleij
random coefficient aj was included to account for similarities
þ bdur durationj þ bcountry countryj
within a couple not accounted for by other measured variables. In
Model1below,letjindicatecouple,iindicateindividual,andpij be Since the coefficients in a fully interactive model for group
the probability of a positive response to the outcome. To simplify differences are not identified (Allison, 1999), we interpret
notation, we abbreviate the effects of country as: this model using predicted probabilities assuming the random

bcountry countryj ¼ bJapan Japanj þ bSpain Spain effect is zero (Long, 2014).
 The four models provide us with estimates of (1) the effect of
þ bGermany Germanyj þ bBrazil Brazilj
an actor’s characteristic on the actor’s odds of experiencing the
Defining aXij as the value of a specific predictor for actor i in outcome, (2) the effect of the actor’s partners’ characteristics on
couple j, and aFemaleij equal to 1 if the actor is female, else 0, the actor’s odds of experiencing the outcome; (3) the effects of
Model 1 is: both the actor’s and the actor’s partners’ characteristics when

123
1614 Arch Sex Behav (2015) 44:1609–1620

Table 2 Descriptive statistics for couple-level measured variables Table 3 Within-pair similarity on measured variables
Measures Min Max M SD % Measures r 95 % CI n
missing
Outcomes
Outcomes Sexual satisfaction 0.36 (0.30, 0.41) 981
Sexual satisfaction 0 1 66.7 % – 1.6 % Relationship happiness 0.35 (0.30, 0.41) 1,000
Relationship happiness 0 1 83.7 % – 0.4 % Predictors
Predictors Age at start of relationship 0.81 (0.79, 0.83) 996
Female 0 1 50.0 % – 0.0 % Good health 0.33 (0.28, 0.39) 1,006
Age at start of relationship 10 67 28.15 8.30 1.2 % Lifetime sexual partners 0.37 (0.31, 0.42) 929
Good health 0 1 73.7 % – 0.1 % Kiss/cuddle often 0.56 (0.52, 0.61) 993
Lifetime sexual partnersa 1 83 8.36 13.66 4.5 % Touched/caressed by partner often 0.53 (0.48, 0.57) 992
Kiss/cuddle often 0 1 62.6 % – 0.8 % Importance of own orgasm 0.37 (0.31, 0.42) 988
Touched/caressed by partner 0 1 41.1 % – 0.9 % Importance of partner’s orgasm 0.41 (0.35, 0.46) 992
often
Frequency of sex in last 4 weeks 0.82 (0.80, 0.84) 967
Importance of own orgasm 1 10 7.90 2.39 1.0 %
Sexual functioning score 0.60 (0.55, 0.64) 914
Importance of partner’s orgasm 1 10 8.25 2.27 0.8 %
Frequency of sex in last 4 weeksa 0 81 5.63 7.19 2.3 % n = number of couples with complete data
Sexual functioning score 1 5 3.51 1.16 5.6 %
Relationship duration 1 50.5 25.11 11.33 1.1 % functioning scores, and strong for reports of recent frequency of
From United Statesb 0 1 20.5 % – 0.0 % sexual activity.
From Japan 0 1 20.5 % – 0.0 %
From Spain 0 1 19.7 % – 0.0 % Actor and Partner Effects of Predictor Variables
From Germany 0 1 19.6 % – 0.0 %
From Brazil 0 1 19.6 % – 0.0 % We first estimated actor and partner effects for each within-
Variables with no reported SD are binary, with the provided percentage couple predictor, run separately for each predictor, on odds of
corresponding to the proportion of individuals responding‘‘yes.’’ sexual satisfaction (Table 4)and relationship happiness (Table 5).
a
Square-root transformation of variable used in analyses. Five indi- When each predictor variable was examined separately, actor
viduals reported frequency of sex in the last 4 weeks above 60. Supple- effects (Model 1) and partner effects (Model 2) were each sig-
mental analyses recoding these extreme values to 60 or dropping them nificantly associated with odds of sexual satisfaction and rela-
from the analysis showed that these cases did not affect the conclusions
b tionship happiness for almost all measured predictors. Impor-
Used as reference category and excluded in analyses
tantly, the results from Model 3 (Tables 4, 5) show that most
of these partner-level effects remain significant predictors of
considered simultaneously, and (4) the effects of the actor and sexual satisfaction and relationship happiness, even after con-
partner when the effects are allowed to vary by gender. trolling for actor-level effects.
With respect to sexual satisfaction, in Model 3 (Table 4),
individuals who reported more lifetime sexual partners were
Results significantly less likely to report being sexually satisfied, while
individuals who reported being in good health, often kissing and
Descriptive Statistics cuddling, often being touched and caressed by their partner,
attaching greater importance to their own, and their partners’,
Means, ranges, and SDs of measures of sexual satisfaction, rela- orgasm, having more frequent recent sexual activity, having
tionship happiness, and potential correlates of these outcomes higher sexual functioning scores, and being happy in their rela-
are shown in Table 2. tionship, were significantly more likely to report being sexually
satisfied.
Within-Couple Similarity Even after controlling for individual-level effects in Model 3
(Table 4), partners’ reports of good health, kissing and cuddling
Within-couple correlations, shown in Table 3, indicate signifi- often, being touched/caressed often, importance of their own
cant actor and partner interdependence for the measured out- orgasm and their partner’s (i.e., the actor’s) orgasm, frequency
comes. Within-couple correlations were modest for ratings of of recent sexual activity, better sexual functioning, and being
sexual satisfaction and relationship happiness, moderate for happy in their relationship were significantly associated with the
reports of frequency of kissing and cuddling and sexual odds that the individual was sexually satisfied. The odds ratios

123
Arch Sex Behav (2015) 44:1609–1620 1615

Table 4 Estimates of actor and partner influence on likelihood of actor’s sexual satisfaction (Models 1–3), examined separately for each predictor, OR
(95 % CI)
Predictor Model 1 Model 2 Model 3
Actor effect only Partner effect only Actor effect Partner effect n

Age at start of relationship 0.99 (0.97, 1.02) 0.98 (0.95, 1.00)* 1.00 (0.98, 1.03) 0.97 (0.95, 1.00) 1,960
Good health 1.68 (1.21, 2.33)** 2.38 (1.71, 3.32)** 1.78 (1.28, 2.49)** 2.44 (1.74, 3.42)** 1,958
Lifetime sexual partners 0.85 (0.77, 0.94)** 1.09 (0.98, 1.21) 0.85 (0.77, 0.94)** 1.09 (0.98, 1.21) 1,815
Kiss/cuddle often 6.83 (4.99, 9.34)** 3.72 (2.82, 4.89)** 5.39 (3.84, 7.57)** 2.38 (1.73, 3.27)** 1,936
Touched/caressed often 8.13 (5.76, 11.48)** 4.03 (2.98, 5.45)** 6.85 (4.72, 9.93)** 3.05 (2.16, 4.29)** 1,947
Importance of own orgasm 1.16 (1.09, 1.23)** 1.27 (1.19, 1.35)** 1.17 (1.10, 1.25)** 1.28 (1.20, 1.37)** 1,926
Importance of partner’s orgasm 1.23 (1.16, 1.32)** 1.26 (1.18, 1.35)** 1.23 (1.15, 1.32)** 1.26 (1.18, 1.35)** 1,934
Frequency of sex in last 4 weeks 3.62 (3.03, 4.33)** 2.75 (2.38, 3.17)** 2.71 (2.19, 3.35)** 1.51 (1.24, 1.82)** 1,895
Sexual functioning score 2.92 (2.50, 3.41)** 2.31 (2.03, 2.64)** 2.56 (2.15, 3.06)** 1.87 (1.59, 2.19)** 1,799
Happy with relationship 5.60 (3.78, 8.29)** 2.23 (1.55, 3.22)** 5.49 (3.65, 8.26)** 2.11 (1.43, 3.09)** 1,947
Model 1 parameters: Actor effect, actor gender, couple’s relationship duration, country of residence. Model 2 parameters: Partner effect, actor gender,
couple’s relationship duration, country of residence. Model 3 parameters: Actor effect, partner effect, actor gender, couple’s relationship duration,
country of residence
OR odds ratio, 95 % CI 95 % confidence interval, n total number of cases included in Model 3
p B .10; * p B .05; ** p B .01 for two-tailed test

Table 5 Estimates of actor and partner influence on likelihood of actor’s relationship happiness (Models 1–3), examined separately for each predictor,
OR (95 % CI)
Predictor Model 1 Model 2 Model 3
Actor effect only Partner effect only Actor effect Partner effect n

Age at start of relationship 1.03 (1.00, 1.06) 0.99 (0.96, 1.02) 1.04 (1.01, 1.08)* 0.97 (0.94, 1.01) 1,983
Good health 2.42 (1.63, 3.61)** 2.02 (1.36, 3.01)** 2.62 (1.73, 3.96)** 2.22 (1.47, 3.35)** 1,981
Lifetime sexual partners 1.05 (0.92, 1.19) 0.90 (0.80, 1.02) 1.02 (0.89, 1.16) 0.90 (0.79, 1.02) 1,829
Kiss/cuddle often 5.89 (3.88, 8.94)** 3.38 (2.31, 4.94)** 4.68 (3.00, 7.31)** 2.30 (1.51, 3.51)** 1,959
Touched/caressed often 5.12 (3.23, 8.12)** 2.44 (1.61, 3.71)** 4.65 (2.88, 7.51)** 2.02 (1.30, 3.16)** 1,954
Importance of own orgasm 1.14 (1.05, 1.23)** 1.07 (0.99, 1.16) 1.15 (1.06, 1.25)** 1.09 (1.01, 1.18)* 1,947
Importance of partner’s orgasm 1.19 (1.10, 1.29)** 1.09 (1.00, 1.18)* 1.20 (1.10, 1.31)** 1.11 (1.02, 1.20)* 1,955
Frequency of sex in last 4 weeks 1.96 (1.60, 2.38)** 1.53 (1.29, 1.82)** 1.89 (1.48, 2.41)** 1.07 (0.85, 1.35) 1,909
Sexual functioning score 2.13 (1.75, 2.60)** 1.59 (1.32, 1.91)** 2.12 (1.71, 2.63)** 1.48 (1.21, 1.82)** 1,803
Sexually satisfied 6.08 (4.01, 9.20)** 2.34 (1.62, 3.37)** 5.88 (3.79, 9.12)** 2.16 (1.45, 3.22)** 1,932
Model 1 parameters: Actor effect, actor gender, couple’s relationship duration, country of residence. Model 2 parameters: Partner effect, actor gender,
couple’s relationship duration, country of residence. Model 3 parameters: Actor effect, partner effect, actor gender, couple’s relationship duration,
country of residence
OR odds ratio, 95 % CI 95 % confidence interval, n total number of cases included in Model 3
p B .10; * p B .05; ** p B .01 for two-tailed test

for the partner effects of these variables on individuals’ sexual place greater importance on their orgasm and their partners’
satisfaction are somewhat smaller in magnitude than the cor- orgasm, report more frequent recent sexual activity, have a
responding actor effects when both effects were considered higher sexual functioning score, and report feeling sexually
simultaneously but they remain both statistically significant and satisfied were significantly more likely to report being happy in
conceptually quite meaningful. their relationship. Even after controlling for the individual effects
With respect to relationship happiness, in Model 3 (Table 5), of these variables, those with partners who reported being in
individuals who were older at the start of their relationship, who good health, kissing and cuddling often, being touched and
werein goodhealth,kiss andcuddleoften,touch andcaressoften, caressed often, placed greater importance on their orgasm and

123
1616 Arch Sex Behav (2015) 44:1609–1620

Table 6 Estimates of actor and partner influence on likelihood of actor’s sexual satisfaction by gender of actor (Model 4), examined separately for each
predictor, OR (95 % CI)
Predictor Female actor Male actor
Actor effect Partner effect Actor effect Partner effect

Age at start of relationship 0.98 (0.93, 1.02) 0.99 (0.95, 1.04) 1.03 (0.98, 1.07) 0.96 (0.91, 1.00)
Good health 1.17 (0.71, 1.93) 2.67 (1.63, 4.36)** 2.61 (1.61, 4.23)** 2.28 (1.39, 3.74)**
Lifetime sexual partners 0.78 (0.63, 0.96)* 1.12 (0.98, 1.28) 0.88 (0.77, 1.00)* 1.04 (0.84, 1.28)
Kiss/cuddle often 4.68 (2.88, 7.60)** 2.33 (1.46, 3.74)** 6.22 (3.83, 10.10)** 2.42 (1.52, 3.87)**
Touched/caressed often 4.82 (2.94, 7.90)** 3.16 (1.89, 5.26)** 10.39 (5.98, 18.06)** 2.92 (1.82, 4.70)**
Importance of own orgasm 1.21 (1.11, 1.32)** 1.17 (1.05, 1.29)** 1.13 (1.02, 1.26)* 1.37 (1.25, 1.50)**
Importance of partner’s orgasm 1.20 (1.10, 1.31)** 1.20 (1.08, 1.34)** 1.30 (1.16, 1.45)** 1.30 (1.18, 1.42)**
Frequency of sex in last 4 weeks 2.45 (1.82, 3.29)** 1.42 (1.08, 1.86)** 3.06 (2.23, 4.19)** 1.63 (1.22, 2.20)**
Sexual functioning score 3.24 (2.55, 4.11)** 1.19 (0.94, 1.51) 2.00 (1.56, 2.56)** 2.57 (2.06, 3.20)**
Happy with relationship 5.30 (3.03, 9.27)** 2.04 (1.12, 3.72)* 5.76 (3.12, 10.64)** 2.16 (1.26, 3.70)**
OR odds ratio, 95 % CI 95 % confidence interval
p B .10; * p B .05; ** p B .01 for two-tailed test

Table 7 Estimates of actor and partner influence on likelihood of actor’s relationship happiness by gender of actor (Model 4), examined separately for
each predictor, OR (95 % CI)
Predictor Female actor Male actor
Actor effect Partner effect Actor effect Partner effect

Age at start of relationship 1.02 (0.96, 1.08) 1.00 (0.94, 1.05) 1.07 (1.01, 1.14)* 0.94 (0.89, 1.00)
Good health 3.01 (1.67, 5.42)** 1.99 (1.11, 3.54)* 2.24 (1.22, 4.14)** 2.50 (1.34, 4.65)**
Lifetime sexual partners 1.02 (0.79, 1.33) 0.89 (0.76, 1.04) 1.01 (0.86, 1.20) 0.90 (0.69, 1.18)
Kiss/cuddle often 3.79 (2.03, 7.08)** 3.04 (1.65, 5.62)** 5.87 (2.98, 11.57)** 1.65 (0.86, 3.20)
Touched/caressed often 3.02 (1.60, 5.70)** 3.26 (1.67, 6.37)** 7.86 (3.64, 16.97)** 1.24 (0.64, 2.38)
Importance of own orgasm 1.20 (1.08, 1.33)** 1.07 (0.94, 1.22) 1.08 (0.94, 1.24) 1.11 (0.99, 1.24)
Importance of partner’s orgasm 1.23 (1.10, 1.38)** 1.11 (0.97, 1.27) 1.15 (1.00, 1.32)* 1.11 (0.99, 1.25)
Frequency of sex in last 4 weeks 2.65 (1.78, 3.93)** 0.89 (0.63, 1.27) 1.39 (0.96, 2.02) 1.28 (0.87, 1.90)
Sexual functioning score 2.21 (1.65, 2.95)** 1.50 (1.10, 2.05)** 1.98 (1.42, 2.77)** 1.49 (1.10, 2.00)**
Sexually satisfied 5.60 (3.10, 10.11)** 2.37 (1.34, 4.18)** 6.20 (3.29, 11.68)** 1.95 (1.06, 3.56)*
OR odds ratio, 95 % CI 95 % confidence interval
p B .10; * p B .05 ;** p B .01 for two-tailed test

their partners’ (i.e., the actor’s) orgasm, had higher sexual func- score, we plotted the probability of the outcome by gender.
tioning scores, and reported feeling sexually satisfied were Because of the focus on partner-level influences in the current
significantly more likely to be happy in their relationship. analyses, we will further highlight gender differences in part-
ner-level influences found in the current analyses (for a descrip-
Gender Differences tion of gender differences in individual-level influences on
sexual satisfaction and relationship happiness, analysed using
Estimates for Model 4 predicting sexual satisfaction are pre- an individual-level approach, see Heiman et al., 2011).
sented in Table 6, with estimates for relationship happiness in As shown in Table 6, significant partner effects were found
Table 7. In Model 4, the effect of each predictor was allowed to for both men and women for nearly all of the measured pre-
differ by gender and by actor/partner status. The effects of each dictors of actor’s sexual satisfaction. The magnitude of the
predictor variable and differences by gender and actor/partner estimated effect differed somewhat between men and women.
status were then assessed using predictions, sometimes referred Importantly, our results indicate a significant effect of partner
to as predictive margins (Graubard & Korn, 1999). For binary sexual functioning on men’s sexual satisfaction, while partners’
predictors, we computed tables of predicted probabilities of sexual functioning did not significantly predict women’s sexual
reporting the outcome by the level of the predictor and gen- satisfaction. Figure 1 illustrates the change in predicted proba-
der. For continuous predictors, such as the sexual functioning bility of sexual satisfaction for both genders across the range of

123
Arch Sex Behav (2015) 44:1609–1620 1617

partners’ response level for each of these two predictors. It is also


worth noting that there is an actor effect of women’s reported
importance of their own orgasm on their relationship happiness,
and a trend for the female partner effect of importance of own
orgasm on male actor’s relationship happiness.

Discussion

The current research illustrates the importance of adopting a


dyadic approach to understanding individual- and partner-level
factors that contribute to sexual satisfaction and relationship
happiness among individuals in enduring relationships. Current
results emphasize that the partner ‘‘matters’’ to an individuals’
Fig. 1 Predicted probability of actor’s sexual satisfaction by actor gender
sexual satisfaction andrelationshiphappiness, and that a compre-
and partners’ reported sexual functioning score. Predicted probabilities hensive understanding of factors that contribute to these couple
were calculated based on parameter estimates obtained in Model 4. All outcomes requires a couple-level research strategy. The number
other variables were held at their means, with random effect held at 0 and magnitude of actor and partner effects on sexual satisfaction
and relationship happiness was surprisingly similar. Partner
effects,even when controlling for actor effects, were very con-
Table 8 Predicted probability of actor’s relationship happiness by actor sistently observed, and explanation of sexual satisfaction and
gender and partner report of kissing/cudding often and being touched/ relationship outcomes always depended on identifying and
caressed often
understanding both individual and partner influences. Several
Predictor Female actor Male actor points in this connection are of special interest.
Kiss/cuddle often 0.95 0.96
First, our findings indicate that partners in ongoing relation-
ships, in this case of significant duration of an average of 25 years,
Do not kiss/cuddle often 0.86 0.94
show significant but modest agreement in reports of sexual sat-
Difference 0.09** 0.02 ns
isfaction and relationship happiness, significant and moderate
Touched/caressed often 0.96 0.96
agreement regarding levels of sexual functioning, and significant
Not touched/caressed often 0.88 0.95
and strong agreement in reports of recent sexual activity. Findings
Difference 0.08** 0.01 ns
for only modest agreement in partners’ reports of sexual satis-
Predicted probabilities were calculated based on parameter estimates faction and relationship happiness—outcomes that are presum-
obtained in Model 4. All other variables were held at their means, with
ably important to couples in long term relationships—are intrigu-
random effect held at 0
ing and potentially consequential, given that sexual satisfac-
** p B .01 for two-tailed test
tion significantly predicts relationship happiness, and rela-
tionship happiness is a significant predictor of sexual satisfac-
partners’ sexual functioning. For female actors, the predicted tion, for both men and women in the relationships under study.
probability of sexual satisfaction appears to remain relatively Within our large, multi-national sample, it appears that couples
high and stable across the range of partners’ sexual functioning that have stayed together over the long run may have accommo-
scores. In contrast, the predicted probability of male actors’ dated to less than perfect similarity in feelings of sexual and rela-
sexual satisfaction is relatively low when female partners’ tionship satisfaction. Can social exchange theory account for
sexual functioning is low but appears to increase significantly as relationship persistence in the face of partner discrepancy in sex-
the partners’ sexual functioning increases. ual and relationship satisfaction, if partners perceive that they
As shown in Table 7, partners’ good health, partners’ sexual lack favorable alternatives to their current relationship (Harvey
functioning, and partners’ sexual satisfaction were significantly & Wenzel, 2006)?Do mature couplessimply experience low lev-
associated with actors’ relationship happiness for both men and els of distress in relation to couple discordance in sexual and
women. Again, the magnitude of the estimated effect differed relationship satisfaction, as is the case with lessened distress
somewhat between men and women. Notably, our results indicate concerning sexual function problems as individuals age (Shifren,
significant effects of partner reports of frequent kissing and cud- Monz, Russo, Segreti, & Johannes, 2008)? Does the fact that
dlingaswellasbeingfrequentlytouchedandcaressedonwomen’s most couple members in the current study reported relatively
relationship happiness, while no significant partner effects were high levels of sexual satisfaction and relationship happiness
found for men. Table 8 shows the relative differences in actor’s diminish the importance of only modest partner agreement
predictedprobabilityofrelationshiphappinessbyactorgenderand concerning these outcomes? Over and above the issue of

123
1618 Arch Sex Behav (2015) 44:1609–1620

modest partner agreement in reports of sexual and relation- poor health, as are individuals who kiss and cuddle often, who
ship satisfaction, we note that the current finding for moderate were touched and caressed often, who rate the importance of
partner concordance in levels of sexual functioning is consis- their own and their partners’ orgasm to be greater, and had more
tent with the literature concerning the reciprocal effects of indi- frequent recent sexual activity, higher sexual functioning scores,
viduals’ and partners’ levels of sexual function and dysfunc- and who are sexually satisfied. Critically, it is also the case that
tion (Fisher et al., 2005; Sand, Fisher, Brock, & Goldstein, 2005). individuals whose partners are in good health, who report
Current findings for strong interpartner correlation in reports kissing and cuddling often and being touched and caressed
of recent sexual activity are also consistent with the published often, who rate the importance of their own and their partners’
literature (e.g., Kinsey, Pomeroy, Martin, & Gebhard, 1953; orgasm to be greater, who have higher sexual functioning scores,
Seal, 1997) and may be viewed as suggestive of the validity of and who are sexually satisfied, also reported higher levels of
participants’ responses to the survey measures. relationship happiness, even when controlling for individuals’
Second, our findings clearly and consistently establish that own reports of these characteristics. Furthermore, the findings
both individual and partner characteristics mutually and con- that individuals who are sexually satisfied report nearly six times
currently influence the sexual satisfaction of individuals in couple the likelihood of relationship happiness, and that individuals
relationships. Individuals in midlife who have good sexual whose partners are sexually satisfied report more than twice the
functioning are more likely to be sexually satisfied than individ- likelihood of relationship happiness, even after controlling for
uals with poorer sexual functioning, and individuals whose individuals’ reports of sexual satisfaction, suggest substantial
partners have good sexual functioning are also substantially more conceptual and empirical interdependence of relationship and
likely to be sexually satisfied than those whose partners have poor sexual satisfaction.
sexual functioning, even when controlling for the individuals’ As with sexual satisfaction, patterns of actor and partner
own level of sexual functioning. Our finding for the influence of a effects on relationship happiness were sometimes gender depen-
partners’ sexual functioning on the individuals’ sexual function- dent, and this was more often the case for correlates of rela-
ing underscores the clinical appropriateness of addressing sexual tionship happiness than for correlates of sexual satisfaction. For
function issues of both the individual and the partner in the sex example, women whose partners reported kissing and cuddling
therapy setting (for discussions of the interdependence of partner often were significantly more likely to indicate that they were
sexual functioning, see Dean et al., 2008; Fisher et al., 2005; Sand happy with their relationship compared to women whose part-
et al., 2005). Similarly, individuals at midlife who are happy with ners did not report kissing and cuddling often, while men’s
their relationship are more likely to be sexually satisfied than relationshiphappinesswassignificantlyaffectedbytheirownbut
individuals who are not happy with their relationship, and individ- not by their partners’ reports of kissing and cuddling. A similar
uals with partners who are happy with their relationship are also pattern of actor and partner effects of touching and caressing on
considerably more likely to be sexually satisfied than those with relationship happiness was also observed for women but not for
partnerswhoarenothappywiththeirrelationship,evenwhencon- men, who only reported actor effects of touching and caressing
trolling for the individuals’ own ratings of their relationship hap- on their relationship happiness. This pattern of female but not
piness. We found similar patterns of actor and partner influ-ences male relationship happiness in association with partner touching
for nearly all other measured correlates of sexual satisfaction. and kissing and cuddling is consistent with findings for women’s
The observed pattern of actor and partner effects on sexual preference for such forms of physical affection (Gulledge,
satisfaction occasionally varied by gender, with women show- Gulledge, & Stahmann 2003; Mansfield, Koch, & Voda, 1998;
ing an actor effect, but no partner effect, of sexual functioning on Peplau, 2003), with findings that cuddling is perceived as a
their sexual satisfaction while men showed both an actor and a nurturant nonsexual activity that is nonetheless associated with
partner effect of sexual functioning on their level of sexual sexual activity (van Anders, Edelstein, Wade, & Samples-Steele,
satisfaction. Our finding for an actor effect but no partner effect 2013), and with theorizing concerning the evolutionary function
of sexual functioning on women’s sexual satisfaction is sug- and hormonal foundations of cuddling (van Anders, Goldey, &
gestive of the independence of women’s sexual satisfaction in Kuo, 2011).
this respect. Our finding for actor and partner effects of sexual Fourth, our findings for couple-level correlates of sexual
functioning on men’s sexual satisfaction appears tobeconsistent satisfaction and relationship happiness demonstrate an over-
with recent research concerning men’s considerable sensitivity arching pattern of some conceptual significance. Specifically, in
to women’s sexual responsiveness in the context of sexual terms of both actor effects and partner effects, sexual satisfaction
intercourse (Salisbury & Fisher, 2014). contributes to relationship happiness, and relationship happiness
Third, our findings clearly and consistently establish that both contributes to sexual satisfaction. Moreover, a host of actor
individual and partner characteristics are related to relationship effects and partner effects for correlates of sexual satisfaction—
happiness. As was the case in models predicting sexual satis- including kissing and touching, cuddling and caressing, and
faction, individuals at midlife who are in good health are more importance of own and partner orgasm—show actor and partner
likely to be happy in their relationship than are individuals in effects for both male and female partners. In contrast, there is

123
Arch Sex Behav (2015) 44:1609–1620 1619

more heterogeneity in actor and partner effects for correlates of same-sex as well as mixed-sex couples) will be required to
relationship happiness for men and women. These observations confirm and extend the relationships we have reported.
for tight male–female correspondence in actor and partner cor-
relates of sexual satisfaction but somewhat lesser male–female Acknowledgments This study was supported by an independent
investigator initiated grant from Bayer-Schering Inc (J. R. Heiman, PI).
consistency in actor and partner correlates of relationship satis- The design, conceptualization, analysis, and interpretation of results are
faction require conceptualization within the area of gender, the sole product of the co-authors and have not been subject to editorial
sexuality, and relationships, as well as further research. influence of Bayer-Schering. The authors wish to acknowledge the
The current research had a number of strengths, including the helpful suggestions of Lorne Campbell regarding this article.
dyadic approach to analyzing both individual and partner influ-
ence on sexual satisfaction and relationship happiness, and it is
unique in its use of a large, heterogeneous, multinational sample
of couples in relationships of relatively long duration. At the References
same time, there are several limitations to this research, including
the fact that the couples recruited for this research were likely Allison, P. D. (1999). Comparing logit and probit coefficients across
groups. Sociological Methods & Research, 28, 186–208.
biased in the direction of sexual and relationship happiness and, Berscheid, E. (1999). The greening of relationship science. American
for purely pragmatic reasons, we sampled only mixed-sex cou- Psychologist, 54, 260–266.
ples. Not only were mean levels of sexual satisfaction and rela- Birnbaum, G. E., Reis, H. T., Mikulincer, M., Gillath, O., & Orpaz, A.
tionship happiness fairly high, the likelihood of self-selection of (2006). When sex is more than just sex: Attachment orientation,
sexual experience, and relationship quality. Journal of Personality
unhappy couples for participation in couples-focused research and Social Psychology, 91, 929–943.
would presumably be low. Moreover, bias may be introduced by Butzer, B., & Campbell, L. (2008). Adult attachment, sexual satisfac-
the requirement that both members of couples participate in the tion, and relationship satisfaction: A study of married couples.
current research and by the exclusion of unsuccessful couples Personal relationships, 15(1), 141–154.
Byers, E. S. (2005). Relationship satisfaction and sexual satisfaction: A
whose relationships no longer exist and who were not available longitudinal study of individuals in long term relationships.
for recruitment to the current research. (Catania, Gibson, Chit- Journal of Sex Research, 42, 113–118.
wood, & Coates, 1990; Saunders, Fisher, Hewitt, & Clayton, Catania, J. A., Gibson, D. R., Chitwood, D. D., & Coates, T. J. (1990).
1985). However, the current study was focused more on pro- Methodological problems in AIDS behavioral research: Influences
on measurement error and participation bias in studies of sexual
cess—in this case, the mutual and concurrent contribution of behavior. Psychological Bulletin, 108, 339–362.
individual and partner characteristics to relationship outcomes— Dean, J., Rubio-Aurioles, E., McCabe, M., Eardley, I., Speakman, M.,
than on absolute or normative levels of sexual satisfaction and Buvat, J., et al. (2008). Integrating couples in ED treatment:
relationship happiness, and we see no specific reason why such Improving the sexual experience of the couple. International
Journal of Clinic Practice, 62, 127–133.
processes would differ dramatically in a relatively self-selected Fincham, D., & Beach, S. R. H. (2006). Relationship satisfaction. In A.
sample, although this remains to be demonstrated. We note as L. Vangelisti & D. Perlman (Eds.), The Cambridge handbook of
well that the requirements of survey research in this large, mul- personal relationships (pp. 579–594). Cambridge: Cambridge
tinational sample required the use of single item, unidimensional University Press.
Fisher, W. A., Rosen, R. C., Mollen, M., Brock, G., Karlin, G., Pommerville,
measures of sexual satisfaction and relationship happiness. P., et al. (2005). Improving the sexual quality of life of couples affected
These constructs may of course be conceptualized and assessed by erectile dysfunction: A double-blind, randomized, placebo-con-
at the multidimensional level, which was not the case in the cur- trolled trial of vardenafil. Journal of Sexual Medicine, 5, 699–708.
rent research. Finally, we note that the current study is cross- Graubard, B., & Korn, E. (1999). Predictive margins with survey data.
Biometrics, 55, 652–659.
sectional and inference of causality is not possible. Cross-sec- Gulledge, A. K., Gulledge, M. H., & Stahmann, R. F. (2003). Romantic
tional research at the couple-level must be complemented by physical affection types and relationship satisfaction. American
longitudinal couple-level research as well as research testing Journal of Family Therapy, 31, 233–242.
theorized mechanisms underlying couple-level effects (e.g., see Harvey, J. H., & Wenzel, A. (2006). Theoretical perspectives in the study
of close relationships. In A. L. Vangelisti & D. Perlman (Eds.), The
Birnbaum et al., 2006; Butzer & Campbell, 2008; Rubin & Cambridge handbook of personal relationships (pp. 35–50).
Campbell, 2011). Cambridge: Cambridge University.
We conclude by noting that the current research clearly Heiman, J. R. (2002). Sexual dysfunction: Overview of prevalence,
demonstrates the significant contributions of both individuals etiological factors, and treatments. Journal of Sex Research, 39, 73–78.
Heiman, J. R., Long, J. S., Smith, S. N., Fisher, W. A., Sand, M. S., &
and their partners to sexual and relationship outcomes. Our Rosen, R. C. (2011). Sexual satisfaction and relationship happiness
findings underscore the value of adopting a dyadic approach to in midlife and older couples in five countries. Archives of Sexual
analyzing relationship data, while providing substantive infor- Behavior, 40, 741–753.
mation regarding important actor and partner correlates of sexual Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied
logistic regression (3rd ed.). New York: Wiley.
satisfaction and relationship happiness among couples at midlife. Kashy, D. A., Campbell, L., & Harris, D.W. (2006). Advances in data
Further dyadic research, adopting longitudinal approaches and analytic approaches for relationships research: The broad utility of
the sampling of additional heterogeneous populations (e.g., hierarchical linear modeling.

123
1620 Arch Sex Behav (2015) 44:1609–1620

Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. surrounding female coital orgasm occurrence in heterosexual partner-
New York: Guilford. ships. Journal of Sex Research, 51, 616–631.
Kinsey, A. C., Pomeroy, W. B., Martin, C. E., & Gebhard, P. H. (1953). Sand, M., Fisher, W. A., Rosen, R., Brock, G., & Goldstein, I. (2005). The
Sexual behavior in the human female. Philadelphia: Saunders. sexual function of women whose partners have sexual dysfunction. In
Laumann, E.O., Paik,A., Glasser,D.B., Kang, J.H.,Wang, T., Levinson,B., C. Meston, S. Davis, A. Traish, & I. Goldstein (Eds.), Women’s sexual
et al. (2006). A cross-national study of subjective sexual well-being function and dysfunction: Study, diagnosis, and treatment (pp. 314–
among older women and men: Findings from the Global Study of 322). London: Taylor and Francis, Blackwell.
Sexual Attitudes and Behaviors. Archives of Sexual Behavior, 35, 145– Saunders, D. M., Fisher, W. A., Hewitt, E. C., & Clayton, J. P. (1985). A
161. method of empirically assessing volunteer selection effects: Recruit-
Long, J. S. (2014). Group comparisons in logit and probit using predicted ment procedures and response to erotica. Journal of Personality and
probabilities. Manuscript in preparation, Indiana University. Social Psychology, 49, 1703–1712.
Mansfield, K. P., Koch, P. B., & Voda, A. M. (1998). Qualities midlife Seal, D. W. (1997). Interpartner concordance of self-reported sexual
women desire in their sexual relationships and their changing behavior among college dating couples. Journal of Sex Research,
sexual response. Psychology of Women Quarterly, 22, 285–303. 34, 39–55.
McMahon, J. M., Pouget, E. R., & Tortu, S. (2006). A guide for multi- Shifren, J., Monz, B., Russo, P. A., Segreti, A., & Johannes, C. B. (2008).
level modeling of dyadic data with binary outcomes using SAS PROC Sexual problems and distress in United States women: Prevalence
NLMIXED. Computational Statistics & Data Analysis, 50, 3663– and correlates. Obstetrics and Gynecology, 112, 970–978.
3680. Spanier, G. (1976). Measuring dyadic adjustment: New scales for
Peplau, L. A. (2003). Human sexuality: How do men and women differ? assessing the quality of marriage and similar dyads. Journal of
Current Directions in Psychological Science, 12, 37–40. Marriage Family, 38, 15–28.
Rosen, R., Brown, C., Heiman, J., Leiblum, S., Meston, C., Shabsigh, R., … Sprecher, S., Christopher, F. S., & Cate, R. (2006). Sexual satisfaction
D’Agostino, R. (2000). The Female Sexual Function Index (FSFI): A and sexual expression as predictors of relationship satisfaction and
multidimensional self-report instrument for the assessment of female stability. In J. Harvey, A. Wenzel, & S. Sprecher (Eds.), Handbook
sexual function. Journal of Sex and Marital Therapy, 26, 191–208. of sexuality in close relationships (pp. 235–256). Mahwah, NJ:
Rosen, R. C., Fisher, W. A., Eardley, I., Niederberger, C., Nadel, A., & Lawrence Erlbaum.
Sand, M. (2004). The The Multinational Men’s Attitudes Life StataCorp. (2013). Stata: Release 13. Statistical Software. College Sta-
Events Sexuality (MALES) Study: I. Prevalence of erectile tion, TX: StataCorp.
dysfunction and related health concerns in the general population. van Anders, S. M., Edelstein, R. S., Wade, R. M., & Samples-Steele, C.
Current Medical Research and Opinion, 20, 607–617. R. (2013). Descriptive experiences and sexual vs. nurturant aspects
Rosen, R. C., Riley, A., Wagner, G., Osterloh, I. H., Kirkpatrick, J., & of cuddling between adult romantic partners. Archives of Sexual
Mishra, A. (1997). The International Index of Erectile Function (IIEF): Behavior, 42, 553–560.
A multidimensional scale for assessment of erectile dysfunction. van Anders, S. M., Goldey, K. L., & Kuo, P. X. (2011). The steroid/
Urology, 49, 822–830. peptide theory of social bonds: Integrating testosterone and peptide
Rubin, H., & Campbell, L. (2011). Day-to-day changes in intimacy pre- responses for classifying social behavioral contexts. Psychoneu-
dict heightened relationship passion, sexual occurrence, and sexual roendocrinology, 36, 1265–1275.
satisfaction: A dyadic diary analysis. Social Psychological and Zou, G., & Donner, A. (2004). Confidence interval estimation of the
Personality Science, 3, 224–231. intraclass correlation coefficient for binary outcome data. Biomet-
Salisbury, C., & Fisher, W. A. (2014). Did you come? A qualitative explo- rics, 60, 807–811.
ration of gender differences in beliefs, experiences, and concerns

123

You might also like