Arch Sex Behav (2011) 40:741–753

DOI 10.1007/s10508-010-9703-3

ORIGINAL PAPER

Sexual Satisfaction and Relationship Happiness in Midlife

and Older Couples in Five Countries
Julia R. Heiman • J. Scott Long • Shawna N. Smith •

William A. Fisher • Michael S. Sand •

Raymond C. Rosen

Received: 1 July 2009 / Revised: 19 October 2010 / Accepted: 19 October 2010 / Published online: 26 January 2011
 Springer Science+Business Media, LLC 2011

Abstract Sexuality research focuses almost exclusively on sexual satisfaction early in the relationship and greater sexual
individuals rather than couples, though ongoing relationships satisfaction later. Within the long-term committed relationship
are very important for most people and cultures. The present context, there were significant gender differences in correlates
study was the first to examine sexual and relationship parame- of sexual and relationship satisfaction, with sexual functioning a
ters of middle-aged and older couples in committed relation- common predictor of both types of satisfaction and physical
ships of 1–51 years duration. Survey research was conducted in intimacy a more consistent and salient predictor for men.
Brazil, Germany, Japan, Spain, and the U.S. targeting 200 men
aged 40–70 and their female partners in each country, with Keywords Couples
Midlife
Sexual satisfaction

1,009 couples in the final sample. Key demographic, health, Relationship happiness
Sexual function

physical intimacy, sexual behavior, sexual function, and sexual Physical intimacy
history variables were used to model relationship happiness and
sexual satisfaction. The median ages were 55 for men and 52 for
women; median relationship duration was 25 years. Relation- Introduction
ship satisfaction in men depended on health, physical intimacy,
and sexual functioning, while in women only sexual functioning Research efforts to understand the place of sexuality in human
predicted relationship satisfaction. Models predicting sexual lives rarely study intact couples in ongoing relationships. Scat-
satisfaction included significant physical intimacy and sexual tered exceptions can be found, primarily in studies of sexual
functioning for both genders and, for men, more frequent recent problems or medical conditions such as HIV risk prevention
sexual activity and fewer lifetime partners. Longer relationship (Remien et al., 2005; Turner et al., 2008) or HPV transmission
duration predicted greater relationship happiness and sexual (Benevolo et al., 2008), infertility (Peterson, Pirritano, Chris-
satisfaction for men. However, women in relationships of 20 to tensen, & Schmidt, 2008), chronic pelvic pain (Smith, Pukall,
40 years were significantly less likely than men to report rela- Tripp, & Nickel, 2007), and sexual dysfunction treatment
tionship happiness. Compared to men, women showed lower (Fisher, Rosen, Eardley, Sand, & Goldstein, 2005; Fisher,
Rosen, Mollen, et al., 2005; Heiman et al., 2007). Consequently,
there is limited research evidence concerning sexual patterns
J. R. Heiman (&)
J. S. Long
S. N. Smith and sexual relationships across the life span and concerning the
The Kinsey Institute for Research in Sex, Gender, and independent, additive, and interactive contributions of couple
Reproduction, Indiana University, Morrison Hall 313, partners to one another’s sex and relationship outcomes. Among
1165 E. Third Street, Bloomington, IN 47405, USA
the reasons for this research lacunae have been the tradition of an
e-mail: [email protected]
individual focus in psychological, behavioral, and medical
W. A. Fisher sciences and the cost and complexity of recruiting and studying
Department of Psychology, University of Western Ontario, both members of ongoing relationships.
London, ON, Canada
Despite the paucity of research evidence concerning cou-
M. S. Sand
R. C. Rosen ple’s sexuality, enduring relationships appear to be linked to life
New England Research Institutes, Watertown, MA, USA quality, health, and satisfaction for many individuals, and

123
742 Arch Sex Behav (2011) 40:741–753

sexuality appears to play an integral, albeit not fully predictable, predict relationship happiness and sexual satisfaction. To the
role in relationship durability and satisfaction (Laumann, Ga- extent allowed by the sample sizeand composition, we also report
gnon, Michael, & Michaels, 1994; Sprecher, 2002). In a compre- the impact of cultural context on the primary outcomes. There
hensive meta-analysis examining predictors of marital satisfac- were three areas of more intensive attention, since we expected
tion and stability (separation or divorce), Karney and Bradbury them to have unique contributions to the prediction ofrelationship
(1995) found that sexual satisfaction was among the strongest happiness and sexual satisfaction: (1) relationship duration; (2)
predictors of both outcomes. For men, sexual satisfaction (r = sexual functioning; and (3) physical intimacy. We hypothesized
.33) showed the highest effect size among predictors of marital that there would be few gender differences in these longer term
stability, comparable to the predictive value of marital satis- committed couples, but that women would be less sexually sat-
faction (r = .29). Results of cross-lagged models of 283 married isfied and that physical intimacy would play a larger role in their
midlife couples revealed causal sequences for both husbands sexual and relationship satisfaction compared to men.
and wives such that individuals satisfied with their sexual rela-
tions tended to be satisfied and happy with their marriages, and
better marital quality, in turn, helped reduce marital instabil- Method
ity (Yeh, Lorenz, Wickrama, Conger, & Elder, 2006). Yet, as
Sprecher and Cate (2004) acknowledge, there are exceptions Participants
to the general finding that sexual contact and satisfaction are
related to relationship satisfaction and durability, such that sex- Five countries were included in the study: Brazil, Germany,
ual and relationship satisfaction can operate somewhat inde- Japan, Spain, and the United States. A benchmark of 200 cou-
pendently. ples was set for each country with the final sample including
A recent examination of sexual behavior and satisfaction in 1,009 couples (2,018 individuals): 207 couples from each of
older individuals (40–80 years old) found that the predictors of Japan and the U.S.; 198 couples from Brazil and Germany; and
sexual well being, defined using four variables (physical plea- 199 couples from Spain. Men in the sample ranged in age from
sure with relationship in the last 12 months, emotional satis- 39 to 70 with a median age of 55. Female partners ranged in age
faction with relationship in the last 12 months, current sexual from 25 to 76 with a median age of 52. Ninety percent of the
functioning/sexual health satisfaction, and importance of sex to couples had children.
life overall), were largely consistent across 29 nations and that Sampling targeted men aged 40–70 in committed relation-
men usually reported higher levels of sexual satisfaction than ships with women, either married or living with a partner a
did women, regardless of sociocultural context (Laumann et al., minimum of 1 year. Gender-specific questionnaires were admin-
2006). The study focused on individuals, not persons and their istered for each partner, with couples instructed not to discuss
partners, and the percent of the sample that was married or cohab- their answers with their partner until all questionnaires were
iting was, in fact, not used in the analyses. There thus remain completed and returned. Data collection, directed and managed
unanswered questions concerning the role and relative impor- by Synovate Healthcare, an international healthcare market
tance of sexual behaviors and interactions, health status, rela- research company, varied by country, using sampling strategies
tionship variables, and historical sexual events that might help to standard for each country. In the U.S., Germany, and Spain,
conceptualize the patterns of sexual and relationship satisfaction participants were recruited by phone, using both random digit
in committed couples, particularly those in longer term relation- dialing (RDD) techniques and established market databases,
ships. and then sent questionnaires by mail for self-completion. In
The present study, the International Survey of Relationships Brazil and Japan, recruitment was done door-to-door, within
(ISR), is the first to focus on sexuality and relationship param- large cities for Brazil, and within randomly sampled locales for
eters among middle-aged and older individuals in committed Japan, and questionnaires then left for respondent self-com-
relationships lasting 1–51 years. The study was conducted in pletion. Quota samples based on age were used in all countries.
five countries—Brazil, Germany, Japan, Spain, and the U.S.— Except for Japan, quota sampling for geographic regions was
targeting men 40–70 and their female partners. Its general pur- also used. Initial response rates, before finding out about the
pose was to assess, among couples in committed relationships, sexual content of the survey, were calculated only for the U.S.
the importance of the relationship, sexual behavior, and the role Details on sampling follow.
sexuality plays in men’s and women’s health and life satisfaction. Brazil: Recruitment in Brazil was done in person, within the
As a first step, the present article describes the sample on key vari- major cities. Recruiters visited homes of potential participants to
ables and develops models predicting relationship happiness and administer surveys. The first part of the survey was administered
sexual satisfaction among this sample of middle and older-aged to one member of the couple; if the first member agreed to
couples in committed relationships. We identified demographic, participate and met the participation criteria, a screener survey
health, physical intimacy, sexual behavior, sexual function, and was then administered to the second member. If both part-
sexual history variables which, based on prior studies, might ners met participation criteria, the full questionnaire was left for

123
Arch Sex Behav (2011) 40:741–753 743

self-completion and collected several hours later. Recruiters did not opt-in). Males in married households between the ages of
reported an initial refusal rate of approximately 25%, with 412 40 and 70 were targeted, with quotas first set by Census region
participants (206 couples) agreeing to complete the survey. with sub-quotas for age within region. Participants were recruited
However, 12 participants (2.9%) refused to complete the entire first by telephone, and sent questionnaires by mail after agreeing
survey, after sexual content was revealed, resulting in 400 par- to participate. 2,274 individuals were contacted, of which 78%
ticipants (200 couples) completing the survey. immediately refused to complete the survey. 500 individuals
Germany: Two methods of sampling were used in recruiting agreed to the survey; however, a further 86 (17.2%) refused to
German participants. For parts of Germany (mostly Eastern), a complete the survey after learning of the sex-related content.
database of people previously contacted for marketing research Overall, 414 participants (207 couples) completed the survey in
was used. This database was compiled through various meth- the U.S. By Census region, 46 couples came from the Northeast;
ods, including phone book sampling, snowball sampling, and 57 from the South; 50 from the Midwest; and 56 from the West.
respondent opt-ins. Participants recruited for our study had not
been contacted for market research within the preceding six
Measures
months. Participants from this database made up 30% of the
total sample. The remaining participants, from those geographi-
International Survey of Relationships
cal areas not represented by the market research database (includ-
ing most of West Germany), were recruited via RDD with
The ISR is a multi-dimensional survey instrument assessing
questionnaires sent via post for respondent self-completion.
domains of demographics, health, mood, selected sexual his-
Information on initial refusal rates was not collected. 440 par-
tory, sexuality behaviors and experiences over the past 4 weeks
ticipants (220 couples) agreed to complete the survey; however,
and 12 months, and the importance of different life areas and
44 participants (10.0%) refused to complete the survey once
sexual activities. It includes 125 questions, many of which were
they discovered its sex-related content. In total, 396 participants
selected from other surveys or standardized questionnaires, with
(198 couples) completed the survey.
some questions developed specifically for this study.1 The
Japan: Recruitment was done door-to-door using a two-
survey was designed by the authors, using a number of selected
step stratified sample: first, sampling locales were selected ran-
questions from prior surveys and several unique to this survey,
domly; second, interviewers called on every nth household.
to provide potentially important information for increasing our
Questionnaires were left with interested households and picked
knowledge of enduring relationships and for designing future
up several hours later. Due to this method of recruitment, as well
clinical programs dealing with sexual and relationship quality in
as cultural constraints in Japan, participants were informed of
older adults. The survey was described to participants as ‘‘a
the sexual nature of the survey without first being administered a
study about people’s relationships and their happiness with
screener survey. Recruiters reported an initial refusal rate of
them. A number of questions deal with aspects of your personal
12% with 414 (207 couples) participants completing the survey.
relationship, including sexuality and sexual experiences.’’ Par-
Spain: Respondent recruitment relied on several established
ticipants were assured that their responses would be confiden-
marketing research databases containing demographic infor-
tial, not shared with their partner, and only analyzed in the
mation on individuals. Individuals from these databases were
aggregate with responses never connected to a specific indi-
identified as eligible on the basis of age and gender, then sam-
vidual. The survey was translated and back-translated for the
pled at-random by region. Respondent quotas were set by both
given language in the countries involved. The study received
regions and age group. Recruitment of participants took place
approval from the Indiana University Institutional Review
initially via phone, with participants then sent questionnaires by
Board for the Protection of Human Subjects.
post for self-completion. Information on initial refusal rates was
not collected. 468 individuals (234 couples) initially agreed to
complete the survey; however, 68 participants (14.5%) refused Dependent Variables
to complete the survey once they were informed of its sex-
related content. Overall, 400 participants (200 couples) com- We examined predictors of relationship happiness and sexual
pleted the survey in Spain, including 40 couples from each of satisfaction. Table 1 presents the wording of the two items
Madrid and Barcelona and 30 from each of Seville, Valencia, used—the first item asked about general relationship happiness
Bilbao, and Vigo. and the second asked about sexual satisfaction over the past 4
U.S.: Sample was pulled from the InfoUSA Listed House- weeks. The relationship happiness question was adapted from
hold database, which comprises all white pages listings from the the Dyadic Adjustment Scale (Spanier, 1976) and included the
U.S. Sampling was limited to listings which specified age original response categories: very unhappy, fairly unhappy, a
(between 50 and 60% of the total listings), and sample infor-
mation was purchased for 12,000 individuals. This sample was 1
For further details about the questionnaire or to request a copy of the
gathered without knowledge of participants (i.e., participants questionnaire, please contact the corresponding author.

123
744 Arch Sex Behav (2011) 40:741–753

little unhappy, happy, very happy, extremely happy, and per- Duration measures were constructed for each individual, rather
fect. Due to our sample size and small marginal distributions for than each couple. Although duration measures could differ within
some outcome categories, our analyses of relationship happi- couples, the intra-couple correlation for our duration measure was
ness used a dichotomized measure comparing happy to unhappy .99. Intra-couple correlations for the original marital years and
relationships by collapsing across the original seven categories. cohabitation years measures were both .96.
Similarly, our measure of sexual satisfaction was adapted from Duration of relationships ranged from 1 to 51 years with
the International Index of Erectile Function (IIEF) and Female a mean of 25.1 years. The median duration measure was
Sexual Function Index (FSFI) (Rosen et al., 1997, 2000) and 25 years with the 25th and 75th percentiles at 17 years and
originally included five categories: very dissatisfied, moder- 34 years. To allow for nonlinear effects of duration in the logit
ately dissatisfied, equally satisfied/dissatisfied, moderately sat- models, we included duration, squared-duration, and cubed-
isfied, and very satisfied. For our analyses, responses were com- duration. Since cubed-duration was not significant in any of
bined into two categories: Not satisfied comprised the first three the models examined, it was not included in the models we
responses and satisfied comprised the latter two. Table 2 pre- presented. To allow age to affect outcomes independently of
sents summary statistics by gender for each variable and Table 3 duration of relationship, the respondent’s age at the beginning
provides the distribution of dependent variables in converting of the current relationship was included in our models.
from multiple to binary categories. Men were significantly more Four categories of relational and sexual variables were
likely to report being happy in their relationship, LRv2(2) = included in our models: measures related to partner physical
7.23, p\.01, while women were more likely to report being intimacy, sexual history, sexual behavior, and sexual func-
satisfied with their sexual relationship, LRv2(2) = 4.59, p = .03. tioning. Five measures of physical intimacy were used. Two
were self-reports of the importance of one’s own orgasm and
Independent Variables partner’s orgasm measured using a 10-point scale with 1 = Not
at all important and 10 = Very important. Binary indicators
Independent variables used to predict relationship happiness were created for whether the respondent reported kissing and
and sexual satisfaction are also described in Table 1. The same cuddling with their partner often and whether they reported
variables were used to predict both dependent variables and the their partner sexually touching or caressing them often (irre-
effects of all variables were allowed to differ by gender. Robust spective of sexual activity). Fifth, a measure of the number of
standard errors that correct for country level clustering were times they had engaged in sexual activity with their partner
used. over the past 4 weeks was included. Sexual activity was
Education and self-reported health were used as demographic defined as any mutually voluntary activity that involves genital
and health controls. Education was measured with a five-category contact and/or stimulation, even if intercourse or orgasm did
scale: (1) did not finish elementary school/primary school; (2) not occur. Sexual history was measured by the total number of
finished elementary school/primary school; (3) finished high reported sexual partners, combining separate reports of male
school/secondary school; (4) finished university/college degree; and female partners over the respondent’s lifetime.3 To min-
(5) finished post-college/university graduate degree. Again, imize the effects of outliers on measures of sexual frequency
because of small marginal frequencies in some categories (espe- and total sexual partners, extreme values were truncated to the
cially by country), three categories were used in our analyses: 95th percentile and then the square root of these two variables
more than high school, high school, and less than high school. was used in our models.
High school was the omitted category in the models. A binary A measure of sexual functioning was used in all models. This
indicator of poor/fair health compared to good/excellent health measure scaled multiple items, where the specific items differed
was also included. for men and women. All items were adapted from questions in
Duration of the couple’s relationship was a key variable of the IIEF and/or FSFI except for a question on premature ejac-
interest. The original ISR instrument did not include a direct ulation. All items were measured with 5-point Likert scales
measure of duration, but participants were asked to report how
many years they had been (1) cohabitating and/or (2) married to
their partner. These measures were used to construct a measure of
Footnote 2 continued
relationship duration for each individual. For non-married indi-
10 years of marriage and 10 years of cohabitation, we were unable to
viduals, years of cohabitation was used; for married individuals, determine with certainty whether these years were contemporaneous
years married was used for the duration measure unless partici- making relationship duration 10 years total or additive making duration
pants reported a significantly longer length of cohabitation.2 20 years total, or a combination thereof.
3
A total of 21 participants reported no male or female lifetime sexual
partners. We confirmed that these participants were in relationships and
2
An ideal measure of duration would include both years cohabitating that they had been sexually active (via other variables, such as reports of
and later years married; however, the survey questions did not guarantee age of first sex or frequency of sex over the past four weeks) and these
mutual exclusivity. For example, for an individual who reported cases were assigned a value of 1.

123
Arch Sex Behav (2011) 40:741–753 745

Table 1 Variable descriptions

Variable label Question Response categories

Dependent variables
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)
Sexual satisfaction over past Over the past 4 weeks, how satisfied have you been 0 = Not satisfied (very dissatisfied; moderately
4 weeks with your sexual relationship with your partner? dissatisfied, equally satisfied/dissatisfied)
1 = Satisfied (moderately satisfied, and very
satisfied)
Independent variables
In good health? Would you say your own health is excellent, good, fair, 0 = Poor/Fair
or poor? 1 = Excellent/Good
Education What is the highest level of education you have \HS/secondary
completed? HS/secondary
[HS/secondary
Relationship duration Constructed from years of marriage and years cohabiting # of years
Age at start of relationship Equal to age minus relationship duration # of years
Importance of orgasm How important is it to you to reach orgasm yourself during Scale from 1 to 10 (1 = Not at all; 10 = Very)
sex with your partner?
Importance of partner orgasm How important is it that your partner reaches orgasm when Scale from 1 to 10 (1 = Not at all; 10 = Very)
you have sex together?
Kiss & cuddle often? My partner and I kiss and cuddle each other: 0 = Very seldom/Seldom
1 = Often/Very often
Touch & caress by partner often? Over the past 4 weeks, how often have you been sexually 0 = Not at all/Once or twice/Weekly
touched and caressed by your partner (whether or not 1 = 2–3 times a week/Daily or almost daily
sexual activity occurred)?
Freq of sex in last 4 weeks How many times during the past 4 weeks have you # of times
engaged in sexual activity with your partner?
Lifetime sexual partners Over your lifetime, with how many partners have you had # of partners, both men and women
sex (by sex we mean any genital contact and/or
stimulation)?
Four-week sexual functioning Scale constructed separately by gender: Men: Level of Average score on included items. Scale with
score desire; frequency of sexual arousal; freq of erection; higher values indicating higher sexual
frequency of maintaining erection after penetration; functioning
frequency of ejaculation; frequency of premature
ejaculation
Women: Level of desire; frequency of sexual arousal;
frequency of lubrication; frequency of orgasm

except for frequency of sexual desire which used a 7-point these men (N = 108).4 Similarly, if a man had not ejaculated
Likert scale. This item was rescaled to 5-points to match the over the past 4 weeks, the item for premature ejaculation was
other items. The final sexual functioning score was the average not included for that person. Four items were used for women:
response to the items that applied to a given individual. For men, (1) frequency of sexual desire over the past 4 weeks; (2) fre-
the sexual functioning was based on six items: (1) frequency of quency of sexual arousal over the past 4 weeks; (3) frequency of
sexual desire; (2) frequency of sexual arousal over the past 4 lubrication over the past 4 weeks; and (4) frequency of orgasm
weeks; (3) frequency of erections over the past 4 weeks; (4) over the past 4 weeks. Men who had ejaculated were allowed
frequency of maintaining an erection after penetration over the two missing items; men who had not ejaculated were allowed
past 4 weeks; (5) frequency of ejaculation over the past 4 weeks; one missing item; women were allowed one missing item.
and (6) frequency of premature ejaculation over the past 4 Among non-missing items, the average response was computed
weeks. Since men who reported ‘‘Never or almost never ejac-
ulating’’were likely to report‘‘Never or almost never ejaculating 4
Of the 108 men who reported‘‘Never or almost never ejaculating,’’96
prematurely,’’the premature ejaculation item was excluded for reported ‘‘Never or almost never ejaculating too quickly.’’

123
746 Arch Sex Behav (2011) 40:741–753

Table 2 Descriptive statistics by gender

Variables Men Women
Min Max M SD Min Max M SD

Dependent variables
Relationship happiness 0 1 0.86 – 0 1 0.81 –
Sexual satisfaction 0 1 0.64 – 0 1 0.69 –
Independent variables
Good health 0 1 0.72 – 0 1 0.75 –
Education
Less than high school 0 1 0.21 – 0 1 0.24 –
High school 0 1 0.43 – 0 1 0.49 –
More than high school 0 1 0.36 – 0 1 0.27 –
Relationship duration (in years) 1 50 25.13 11.35 1 51 25.08 11.33
Age at start of relationship (in years) 15 67 29.85 8.36 10 64 26.46 7.88
Importance of orgasm 1 10 8.45 2.02 1 10 7.35 2.60
Importance of partner orgasm 1 10 8.53 2.00 1 10 7.96 2.49
Kiss/cuddle often? 0 1 0.63 – 0 1 0.62 –
Touch/caress by partner often? 0 1 0.38 – 0 1 0.44 –
Freq of sex in last 4 weeks 0 81 5.74 7.55 0 80 5.52 6.82
Lifetime sexual partners 1 83 11.91 16.62 1 83 4.95 8.75
Sexual functioning score 1 5 3.81 1.02 1 5 3.22 1.23
Note: Variables where no SD is listed are binary; the mean can be read as the proportion answering‘‘yes’’or converted into the appropriate percentage
(e.g., 75% of women reported good health). Transformations of some variables are performed for analysis (see text for details)

for our scale. For men who had ejaculated over the past 4 weeks, Modeling Strategy
the scale reliability was .70. For men who had not ejaculated, the
scale reliability coefficient was .79. For women, the scale reli- Logit models were estimated for each dependent variable where
ability coefficient was .89. the effects of predictors were allowed to differ by gender. While
Finally, country-level controls were included in all models. odds ratios are reported in tables, we also considered how changes
Due to differences in sampling strategies (see discussion above), in a given predictor affected the probability of the outcome. The
country-level differences cannot be interpreted substantively, effect of a variable was illustrated by comparing the predicted
but reflect different mean-level responses in reference to the probability of the outcome (e.g., relationship happiness), at spe-
U.S. which was used as the excluded category. A reference cific values of the predictors (e.g., at the mean) as the variable of
country merely provides a basis of comparison and does not interest (e.g., duration) changes; for further details, see Long and
imply special status for that country. Freese (2005).

Missing Data Results

When one partner provided no information on relationship dura- Tetrachoric correlations between our binary measures of sex-
tion, their partner’s information was used. Allowances were also ual satisfaction and relationship happiness by gender were .40
made for missing variables in the construction of the sexual for men and .41 for women, indicating 16% shared variation
functioning scales as described above. For model estimations, a between the dependent variables.
listwise deletion of cases was used (i.e., only those cases that
included data for all variables were used to estimate coeffi- Relationship Happiness
cients). This left 860 males (85.2% of the total cases) for both the
relationship happiness and sexual satisfaction models, 899 Table 4 presents the odds ratios from binary logit models of
females (89.1%) for the relationship happiness model, and 898 relationship happiness for men and women. The odds ratios
for the sexual satisfaction model (88.9%). Table 2 presents show the factor change in the odds of being happy versus not
descriptive statistics by gender for the variables used in our happy for a one-unit change in a variable, holding other vari-
models. ables constant.

123
 Table 3 Distribution of dependent variables by gender

Original response categories

Binary categories
Happy with 1 Extremely 2 Fairly 3 Little 4 Happy 5 Very 6 Extremely 7 Perfectly Refused/ Total
relationship unhappy unhappy unhappy happy happy happy omitted

0 Unhappy
Arch Sex Behav (2011) 40:741–753

M 21 47 74 0 0 0 0 0 142
F 18 63 105 0 0 0 0 0 186
1 Happy
M 0 0 0 337 309 159 59 0 864
F 0 0 0 314 284 151 68 0 817
Ref
M 0 0 0 0 0 0 0 3 3
F 0 0 0 0 0 0 0 6 6
Total
M 21 47 74 337 309 159 59 3 1009
F 18 63 105 314 284 151 68 6 1009
Satisfied with sexual 1 Very 2 Moderately 3 Equally 4 Moderately 5 Very Refused/ Total
relationship with partner satisfied satisfied satisfied/ dissatisfied dissatisfied omitted
dissatisfied

0 Dissatisfied
M 0 0 168 118 67 0 353
F 0 0 196 70 42 0 308
1 Satisfied
M 292 348 0 0 0 0 640
F 314 371 0 0 0 0 685
Ref
M 0 0 0 0 0 16 16
F 0 0 0 0 0 16 16
Total
M 292 348 168 118 67 16 1009
F 314 371 196 70 42 16 1009
747

123
748 Arch Sex Behav (2011) 40:741–753

Table 4 Odds ratios from logit model of having a happy relationship by these variables on the predicted probabilities of reporting rela-
gender tionship happiness—both independently and jointly. For a man
Men Women with average characteristics, kissing and cuddling frequently or
frequent sexual caressing by the partner increased his proba-
In good health? 1.67* (2.29) 1.41 (1.37)
a
bility of reporting relationship happiness by .11 and .09. A man
Education (Ref = high school) who reported both frequent kissing/cuddling and sexual caress-
Less than high school 1.65 (0.87) 1.62 (1.48) ing by partner had a predicted probability of reporting rela-
More than high school 1.20 (0.99) 1.30 (0.84) tionship happiness that was .21 higher than his counterpart who
Relationship durationb 1.06 (1.45) 1.00 (1.65) reported neither. Sexual functioning also proved important for
Relationship duration squared 1.00 (0.08) 1.00** (3.23) male relationship happiness, with each point higher increasing
Age at start of relationship 1.03 (1.31) 1.02 (1.20) the odds of happiness by a factor of 1.44. Duration also has a
Importance of orgasm 0.90 (1.05) 0.96 (0.41) significant and positive effect on relationship happiness that is
Importance of partner orgasm 1.17*** (4.21) 1.08 (1.22) discussed further below.
Kiss/cuddle often? 3.00*** (4.92) 1.59 (1.25) Fewer significant effects on relationship happiness were
Touch/caress by partner often? 3.11*** (3.66) 1.35 (0.84) found for women (Table 4). None of the demographic variables
Freq of sex in last 4 wks (square root) 0.87 (1.50) 1.24 (1.24) were significant, although the same pattern of country-level
Lifetime sexual partners (square root) 0.97 (0.46) 0.93 (0.49) effects was found for women as for men. None of the physical
Sexual functioning score 1.44*** (3.72) 1.42** (2.84) intimacy measures significantly distinguished between women
Country (Ref = USA) who reported relationship happiness and those who did not, net
Brazil 0.52*** (3.68) 0.40*** (6.82) of other characteristics. Relationship duration, however, was a
Germany 1.12 (0.81) 0.91 (1.02) significant predictor of relationship happiness for women. As
Japan 2.36** (3.25) 1.81* (2.14) with men, women with higher levels of sexual functioning were
Spain 4.89** (3.20) 0.71* (2.22) more likely to report relationship satisfaction, with each one-
N 860 899 point increase increasing the odds of reporting happiness by a
Pseudo R2 0.152 0.124 factor of 1.42.
Figures 1 and 2 plot the predicted probabilities of relationship
Note: Odds ratios are the exponential of the slope coefficients and indi-
cate the factor change in the odds of a happy relationship compared to happiness for men and women as duration and sexual func-
an unhappy relationship. Z-statistics are given in parentheses. Models tioning change. In these and later figures, the predicted proba-
include robust standard errors that adjust for clustering by country. bilities for men are plotted with lines that include a square
Figures are rounded to 2 decimal points symbol while lines with circles show predictions for women.
a
Joint significance for education for men and for women was not sig- We tested whether the predicted probabilities (e.g., the proba-
nificant
b
bility of being in a happy relationship) were equal for men and
Joint test of significance for duration and duration squared for men
was significant at the .01 level; for women, at the .001 level women. When the gender differences were significant at the .05
*p\.05; **p\.01; ***p\.01 for two-tailed tests

Demographic factors were generally unimportant for men

in predicting relationship happiness. However, health was an
important predictor with men in good health having odds of a
happy relationship that were 1.67 times larger than those
reporting fair or poor health, net of other variables. Net of other
predictors, Brazilian and Spanish men reported significantly
lower odds of relationship happiness than American men,
while Japanese men reported significantly higher levels.
Three measures of physical intimacy were significant pre-
dictors of relationship happiness for men. Men who valued their
partner’s orgasm were more likely to report relationship hap-
piness, with each additional point of valuation increasing the
odds of happiness by a factor of 1.17. Reporting frequent kissing Fig. 1 Predicted probabilities of relationship happiness by duration and
and cuddling and frequent sexual caressing by partner both gender. Note: Solid lines indicate that the predicted probabilities for men
and women were significantly different at the .05 level; dashed lines
increased the odds of reporting relationship happiness by a indicate that the differences were not significant. All variables except
factor of approximately 3. Table 6 illustrates the importance of duration and duration-squared were held at their means

123
Arch Sex Behav (2011) 40:741–753 749

Table 5 Odds ratios from logit model for being sexually satisfied by
gender
Men Women

In good health? 1.60** (2.99) 0.74 (1.48)

a
Education (Ref = high schools)
Less than high school 1.00 (0.01) 0.94 (0.15)
More than high school 1.02 (0.11) 1.21 (0.93)
Relationship durationb 1.02 (1.01) 1.07* (2.51)
Relationship duration squared 1.00 (0.02) 1.00 (0.84)
Age at start of relationship 1.01 (0.73) 1.03 (1.30)
Importance of orgasm 0.91 (1.82) 0.93 (1.11)
Importance of partner orgasm 1.19* (2.49) 1.02 (0.25)
Kiss/cuddle often? 1.93* (2.28) 2.03* (2.19)
Touch/caress by partner often? 3.23*** (7.98) 1.51* (2.54)
Fig. 2 Predicted probabilities of relationship happiness by sexual func- Freq of sex in last 4 wks 2.05*** (5.88) 1.59** (2.91)
tioning score and gender. Note: Solid lines indicate that the predicted (square root)
probabilities for men and women were significantly different at the .05
Lifetime sexual partners 0.86*** (6.02) 0.90 (1.52)
level; dashed lines indicate that the differences were not significant. All
(square root)
variables except duration and duration-squared were held at their means
Sexual functioning score 1.37*** (3.76) 2.34*** (12.50)
Country (Ref = USA)
level, the lines are solid; if gender differences were not signifi- Brazil 1.19 (0.82) 1.68** (2.74)
cant, the lines are dashed.5 Germany 0.92 (0.48) 1.51 (1.83)
For men, duration had a significant and positive impact on Japan 2.62*** (10.82) 4.32** (3.23)
reporting relationship satisfaction, LRv2(2) = 11.24, p\.01. Spain 0.91 (0.38) 1.24 (0.75)
For example, a man who had been in a relationship for five N 860 898
years and was average on all other characteristics had a pre- Pseudo R2 0.290 0.258
dicted probability of reporting relationship happiness of about Note: Odds ratios are the exponential of the slope coefficients and indi-
.75 compared to a probability of about .97 in the 45th year of cate the factor change in the odds of a happy relationship compared to
the relationship. Although duration measures were also signif- an unhappy relationship. Z-statistics are given in parentheses. Models
icant predictors of relationship happiness for women, LRv2(2) = include robust standard errors that adjust for clustering by country.
Figures are rounded to 2 decimal points
13.13, p\.01, the effect was more complex. From years 1 a
Joint significance for education for men and women was not signifi-
through 15 of the relationship, duration decreased the proba- cant
bility of relationship happiness. Beginning in year 20, the b
Joint test of significance for duration and duration squared for men
effect turned positive through year 50. was not significant; for women, it was significant at the .001 level
Figure 2 shows that, for both men and women, increasing *p\.05; **p\.01; ***p\.01 for two-tailed tests
sexual functioning had a persistent, positive effect on the prob-
ability of relationship happiness. For women, moving from the
lowest sexual functioning score to the highest, holding other gender. The coefficients indicate how a unit increase in a
variables constant at the mean, increased the probability of predictor affects the odds of being satisfied as opposed to not
reporting relationship happiness from .68 to .90. There was a satisfied, holding all other variables constant.
similar magnitude of change in predicted probabilities for For men, we found that good health had a positive effect,
men, with a change from lowest level of functioning to highest increasing the odds of sexual satisfaction by a factor of 1.60. At
increasing the probability from .78 to .94. the country level, Japanese men had odds of reporting signifi-
cantly greater sexual satisfaction that were 2.62 times greater
Sexual Satisfaction than American men. As with relationship happiness, physical
intimacy measures were also important for men in predicting
Table 5 presents the odds ratios from a binary logit for sexual sexual satisfaction. Valuing one’s partner’s orgasm signifi-
satisfaction with the effects of each variable differing by cantly increased the odds of sexual satisfaction, with each addi-
tional point increasing the odds of satisfaction by 1.19. Fre-
5 quent kissing and cuddling, as well as frequent sexual caress-
The significance of the difference in predicted probabilities depends
both on the magnitude of the difference but also on factors such as the ing, were important predictors for men. Table 6 presents the
number of cases at a given level of the independent variable. magnitudes of these effects in terms of predicted probabilities

123
750 Arch Sex Behav (2011) 40:741–753

Table 6 Predicted probabilities of relationship happiness and sexual

satisfaction, by gender and intimacy measures
Men Women

Probability of reporting relationship happiness

Seldom kiss/cuddle 0.83 0.78
Often kiss/cuddle 0.93 0.85
Difference 0.11 0.07 ns
Seldom caress by partner 0.86 0.81
Often caress by partner 0.95 0.85
Difference 0.09 0.04 ns
Seldom kiss/cuddle & caress 0.75 0.76
Often kiss/cuddle & caress 0.97 0.87
Difference 0.21 0.11
Probability of reporting sexual satisfaction
Seldom kiss/cuddle 0.65 0.67 Fig. 3 Predicted probabilities of sexual satisfaction by relationship
duration and gender. Note: Solid lines indicate that the predicted prob-
Often kiss/cuddle 0.78 0.80
abilities for men and women were significantly different at the .05 level;
Difference 0.13 0.14 dashed lines indicate that the differences were not significant. All vari-
Seldom caress by partner 0.64 0.72 ables except duration and duration-squared were held at their means
Often caress by partner 0.85 0.80
Difference 0.21 0.07 the change was more gradual, increasing from .66 in year 1 to .78
Seldom kiss/cuddle & caress 0.53 0.63 in year 40.
Often kiss/cuddle & caress 0.88 0.84 The odds of reporting sexual satisfaction doubled for
Difference 0.34 0.21 women who reported frequent kissing and cuddling compared
to their counterparts who did not report frequent kissing and
Note: Predicted probabilities are calculated for at the within gender
mean level of predictors; the mean for duration squared is held at the
cuddling. Frequent sexual caressing by the partner similarly
square of the mean for duration. All differences were significant at the raised the odds of sexual satisfaction by a factor of 1.51.
.05 level unless marked ns Table 6 shows the magnitude of these effects on predicted
probabilities, for a woman average on other characteristics.
Similar to men, sexual frequency over the past 4 weeks was
of reporting relationship happiness and sexual satisfaction for also positively related to sexual satisfaction for women. A unit
specific values of the predictors in the model. increase in the square root of sexual frequency increased the
In contrast to the relationship happiness model, frequency of odds of sexual satisfaction by a factor of 1.59. Sexual func-
sex over the past 4 weeks had a positive effect on the odds of tioning again had a marked impact for women, with each
sexual satisfaction for men. An increase of one unit in the square additional point more than doubling the odds of sexual
root of frequency increased the odds of reporting sexual satis- satisfaction.
faction by a factor of 2.05. Sexual history also had a small Figures 3 and 4 show the predicted probabilities of sexual
negative effect, with a unit increase in the square root of number satisfaction by relationship duration and sexual functioning for
of lifetime sexual partners decreasing the odds of sexual satis- men and women. As noted above, duration of the relationship
faction by a factor of .86. As with relationship happiness, sexual has a smaller effect on sexual satisfaction for men than women.
functioning was positively related to sexual satisfaction for men. For women, however, the impact of duration on sexual satis-
Each additional point increase in sexual functioning score faction is both significant and substantial. During the first
increased the odds of satisfaction by a factor of 1.37. 15 years of the relationship, women had significantly lower
For women, there was no relationship between demographic probabilities of reporting satisfaction than men. From year 30
and health variables and the outcome variable. Country-level on in the relationship, however, women had a significantly
patterns indicated that Japanese and Brazilian women were sig- higher probability of reporting satisfaction.
nificantly more likely than U.S. women to report sexual satis- Figure 4 shows the magnitude of the effect of sexual func-
faction, net of all other characteristics. As with relationship hap- tioning on predicted probabilities of sexual satisfaction for
piness, duration was a significant predictor of sexual satisfac- individuals who were average on all other characteristics. For
tion, LRv2(2) = 20.50, p\.001. As Fig. 3 shows, for a woman both genders, we found substantial effects on the probability of
who was average on all other characteristics, the probability of reporting sexual satisfaction, with larger effects for women.
reporting sexual satisfaction increased from .40 at the beginning Moving from lowest to highest levels of sexual functioning for
of the relationship to .86 in year 40 of the relationship. For men, a women average on other characteristics, for example,

123
Arch Sex Behav (2011) 40:741–753 751

genders as a large majority reported being happy with their

relationship (see Tables 2, 3). As with prior studies, relation-
ship and sexual satisfaction, the two dependent variables used
here, were significantly correlated, demonstrating 16% shared
variance. Women reported significantly more sexual satisfac-
tion than men and men more relationship happiness than women,
contrary to our hypothesis. In the context of these specific char-
acteristics, our focus was on modeling predictors of relationship
happiness and sexual satisfaction.
The models tested relevant variables contributing to sexual
and relationship outcomes and found different patterns by
outcome and by gender. For example, the physical intimacy
variables (kissing/cuddling, partner touch/caressing and, to a
lesser extent, importance of partner’s orgasm) unexpectedly
predicted relationship happiness for men only, whereas fre-
Fig. 4 Predicted probabilities of sexual satisfaction by sexual function- quent kissing and cuddling and frequent partner touching and
ing score and gender. Note: Solid lines indicate that the predicted proba-
bilities for men and women were significantly different at the .05 level; caressing were important for both men’s and women’s sexual
dashed lines indicate that the differences were not significant. All vari- satisfaction. The degree to which physical intimacy (that was
ables except duration and duration-squared were held at their means not necessarily sexual) was rated as important to men’s but not
women’s relationship happiness was striking, suggesting a
increased the predicted probability of reporting sexual satis- need for reconsideration of the role of physical affection and its
faction from .32 to .93. Men had a somewhat attenuated, but meanings for each gender in longer term relationships. For
still large, effect, increasing the probability from .54 to .80. both men and women, sexual frequency was related to sexual
Paralleling our findings with duration, women had signifi- satisfaction but not relationship happiness, suggesting some
cantly lower probabilities of reporting satisfaction with func- independence of sexual activity and relationship satisfaction.
tion levels from 1 to 2, with significantly higher probabilities The degree of independence/dependence of sexuality and rela-
with function scores above 3.5. tionship satisfaction is a point of disagreement in the literature
with evidence in both directions but often supporting a more
independent factor structure. For example, Hassebrauk and Fehr
Discussion (2002) conducted a series of studies in which they explored the
dimensions of relationship quality. Using principal compo-
The present study examined couples in a committed rela- nents analysis, Hassebrauk and Fehr reported that sexuality
tionship where the men were between 39 and 70 years (med- formed one of four factors (without any item overlap), the
ian, 55) and their female partners between 25 and 76 years others being intimacy, agreement, and independence. The
(median, 52) from Brazil, Germany, Japan, Spain, and the U.S. correlation between the sexuality and other three factors was
On average, these were long term relationships with 50% of r = .13 to .35 across three samples of German and Canadian
them more than 25 years in duration (only 5% were together men and women.
7 years or less). Relationship duration is usually examined Sexual functioning predicted both relationship and sexu-
indirectly through divorce statistics, but they reveal that long ality outcomes, with increasing levels of functioning related to
duration marriages do exist. For example, according to 2001 higher levels of positive outcome in the sex and relationship
figures, a U.S. Census Report noted that approximately 50% of domains. However, the data suggest that level of sexual func-
first marriages (52% for men and 49% for women) last 25 years tioning had a more dramatic effect on sexual satisfaction for
or more (U.S. Census Bureau, 2005). Marriages expected to women than men; at the low end of sexual functioning for both
continue beyond 20 years were estimated at approximately genders, women were less likely than men to report sexual sat-
92% in Spain, 70% in Germany and Japan, and 56% in the U.S. isfaction, while at the high end of sexual functioning, women
(Andersson & Philipov, 2001; Raymo, Iwasawa, & Bumpass, were more likely than men to report satisfaction. This gender
2004). Comparable data could not be located for Brazil. Thus, difference was important and suggests a closer examination by
longer term relationships are a significant fact of social and sexual function domain, including sexual desire, erections/
sexual life for a large number of couples. arousal, and orgasm, as this result may differ by each sexual
Given the length of the relationships in the present sample, function category that in the present study were collapsed into
we expected that relationship happiness might be fairly con- a single scale.
stant but sexual satisfaction lower for those together the most Lifetime number of sexual partners was a significant pre-
years. The mean satisfaction levels were quite high for both dictor of sexual but not relationship satisfaction, and this only

123
752 Arch Sex Behav (2011) 40:741–753

for men. Men reported more lifetime sexual partners than somewhat less relationship happiness due to inter-partner adjust-
women (M = 11.9 vs. 4.9). More sexual partners predicted less ments in roles and possible health issues.
sexual satisfaction. Searching for a better partner or sexual There were several limitations to this study. It incorporated
experience may emerge from or be connected to a lack of multiple modes for administering the surveys, each with dif-
sexual satisfaction rather than just a desire for sexual recrea- ferent sampling methods. The generalizability of our findings
tion and variety. Alternatively, more partners might indicate is therefore constrained, since the protocols were not represen-
different standards based on greater experience. tative of national populations. Thus, although country differ-
The focus on relationship duration as a key variable allowed ences were of interest, our interpretation of them is necessarily
us to examine patterns which predicted sexual satisfaction and limited. Differing sampling strategies per country, relatively
relationship happiness at a given time since the couple had small sample sizes, and the fact that the response items may
been together. Laumann et al.’s (1994) data suggested, using have been interpreted differently in each language/culture means
single item indicators, gender differences associated with age the country differences are only suggestive. We can make no
in a random probability sample of U.S. individuals (relation- claims about the relative relationship happiness or sexual satis-
ship duration was not examined with regard to these variables). faction levels by country. While one might look at whether the
The percent of women who reported that they were ‘‘extre- patterns found here in predicting relationship and sexual out-
mely’’ emotionally satisfied or physically satisfied with their comes would be similar within each country, our sample sizes
partner decreased in the 55–59 year olds (to 26.3 and 28.8%, were too small to do this reliably in this sample. The present
respectively) compared to women who were younger, while study included country variables to encourage further investi-
men in the 55–59 year old age group showed increases (to gation of potential differences at these age and relationship
54.8 and 57.5%, respectively) on both variables. In the present duration levels but solving the sampling and survey delivery
study, men’s relationship happiness increased linearly with mode problems is a substantial task. This is especially true for
each category of relationship length, while women’s did not. enrolling couples,ratherthanindividuals,toparticipateinresearch
Women were more likely to report less relationship happiness since the issues of representative sampling for one gender may
in the first 15 years, with greater levels starting at 20–50 years not fit the other and protocols for sampling relationships have
together. During the later years, however (relationship years not been developed to our knowledge.
20–40), women’s relationship happiness was significantly The issue of couple recruitment also applies to the low
lower than men’s. Sexual satisfaction for the women showed response rates in the present study. It is possible that lower
a very different pattern. Women in relationships up to about response rates create sample selection biases by including
10 years reported less sexual satisfaction and in later years individuals who are either more interested or more confident in
(25–50) more sexual satisfaction than men at the same dura- the topic or who have more time to answer questions. In this
tion points. These data, perhaps because they involve couples study, it may have selected people who were happier with their
rather than just individuals, contrast with Liu’s (2003). Liu, relationships and were willing, once informed of the sexual
using the Laumann et al. (1994) data base, found that duration content, to continue (though the loss due to sexual content was
did have a small and negative effect on the quality of marital less than expected). Other studies of international samples on
sex and women were less satisfied than men with their marital sexuality have faced similar obstacles (Laumann et al., 2006;
sexual lives. Rosen et al., 2004) due to recruitment burden and expense,
It is important to note that the overall levels of relationship cross-cultural variation in acceptance of different research
happiness were high in these couples, so the changes are better approaches, and content sensitivity. Participation rates in the
seen as reflecting greater or lesser degrees of happiness rather current study were not out of line with other cross-cultural
than a distinct difference between being happy and unhappy. studies in which couples have been enrolled (Rosen et al.,
The patterns of results fit partner choice and lifespan event 2004). While there is no obvious solution to this problem, one
considerations. It is likely that women selected committed approach is to regard the present findings as descriptive and
partners based on relationship quality more than sexual satis- exploratory in nature, to be confirmed in further larger epi-
faction and that, for women in childbearing years, dealing with demiological studies of this type. By examining relationships
avoiding pregnancies or having children among other life among variables and evaluating predictor models in the con-
stressors, all take a toll on sexual satisfaction. Women who were text of the descriptive information about the sample, we can at
in the 25 and beyond years of a relationship were at least 50 years least begin to formulate ideas about how patterns of sexuality
old (with partners at least 5 years older) and thus in physical fit into enduring relationships, knowing they will need further
(menopausal) and role transitions with their partner. Since 90% refinement and testing.
of the couples had children, the burdens and distractions of In conclusion, this is the first international study of indi-
parenting, falling heavily on women in their 30 s and 40 s, begin viduals in committed relationships, where the focus was on
to shift. This change, along with the freedom from reproductive middle and older-aged men and women in relationships with a
worries, may facilitate greater levels of sexual satisfaction, yet median of 25 years duration. The goal was to describe key

123
Arch Sex Behav (2011) 40:741–753 753

demographic, health, sexual behavior and function, and physi- global study of sexual attitudes and behaviors. Archives of Sexual
cal intimacy factors that might predict sexual satisfaction and Behavior, 35, 145–161.
Liu, C. (2003). Does quality of marital sex decline with duration? Archives
relationship happiness. The findings that sexual and relation- of Sexual Behavior, 32, 55–60.
ship satisfaction were to some degrees independent and dif- Long, J. S., & Freese, J. (2005). Regression models for categorical depen-
ferentially impacted by gender and years in a relationship are dent variables using Stata (2nd ed.). College Station, TX: Stata
useful as an initial framework to begin to examine couples in Press.
Peterson, B. D., Pirritano, M., Christensen, U., & Schmidt, L. (2008).
more depth as to how sexuality and relationship variables The impact of partner coping in couples experiencing infertility.
interact with different life phases. Human Reproduction, 23, 1128–1137.
Raymo, J. M., Iwasawa, M., & Bumpass, L. (2004). Marital dissolution
Acknowledgments This study was supported by an independent inves- in Japan: Recent trends and patterns. Demographic Research, 11,
tigator-initiated grant from Bayer-Schering. Sampling and data collection 395–420.
fieldwork were performed by Synovate Healthcare. The design, concep- Remien, R. H., Stirratt, M. J., Dolezal, C., Dognin, J. S., Wagner, G. J.,
tualization, analysis, and interpretation of the results were the sole product Carballo-Dieguez, A., … Jung, T. M. (2005). Couple-focused
of discussions among the co-authors, represent the consensus of the co- support to improve HIV medication adherence: A randomized
authors, and have not been subject to editorial revision by Bayer-Scher- controlled trial. AIDS, 19, 807–814.
ing. Johanna Salazar provided valuable assistance with article prepa- Rosen, R., Brown, C., Heiman, J., Leiblum, S., Meston, C., Shabsigh, R.,
ration. … D’Agostino, R. (2000). The Female Sexual Function Index
(FSFI): A multidimensional self-report instrument for the assess-
ment of female sexual function. Journal of Sex and Marital Ther-
apy, 26, 191–208.
References Rosen, R., Fisher, W., Eardley, I., Niederberger, C., Nadel, A., & Sand,
M. (2004). The Multinational Men’s Attitudes to Life Events and
Andersson, G., & Philipov, D. (2001). Life-table representations of Sexuality (MALES) Study. I. Prevalence of erectile dysfunction
family dynamics in 16 FFS countries (MPIDR Working Paper and related health concerns in the general population. Current
Series WP 2001-024). Rostock, Germany: Max-Planck Institute Medical Research Opinion, 20, 607–617.
for Demographic Research. Rosen, R. C., Riley, A., Wagner, G., Osterloh, I. H., Kirkpatrick, J., &
Benevolo, M., Mottolese, M., Marandino, F., Carosi, M., Diodoro, M. Mishra, A. (1997). The International Index of Erectile Function
G., Sentinelli, S., … Vocaturo, A. (2008). HPV prevalence among (IIEF): A multidimensional scale for assessment of erectile dys-
healthy Italian male sexual partners of women with cervical HPV function. Urology, 49, 822–830.
infection. Journal of Medical Virology, 80, 1275–1281. Smith, K. B., Pukall, C. F., Tripp, D. A., & Nickel, J. C. (2007). Sexual
Fisher, W. A., Rosen, R. C., Eardley, I., Sand, M., & Goldstein, I. (2005). and relationship functioning in men with chronic prostatitis/
Sexual experience of female partners of men with erectile dysfunc- chronic pelvic pain syndrome and their partners. Archives of Sexual
tion: The Female Experience of Men’s Attitudes to Life Events Behavior, 36, 301–311.
and Sexuality (FEMALES) study. Journal of Sexual Medicine, 2, Spanier, G. (1976). Measuring dyadic adjustment: New scales for assess-
675–684. ing the quality of marriage and similar dyads. Journal of Marriage
Fisher, W. A., Rosen, R. C., Mollen, M., Brock, G., Karlin, G., Family, 38, 15–28.
Pommerville, P., … Sand, M. (2005). Improving the sexual quality Sprecher, S. (2002). Sexual satisfaction in premarital relationships:
of life of couples affected by erectile dysfunction: A double-blind, Associations with satisfaction, love, commitment, and stability.
randomized, placebo-controlled trial of vardenafil. Journal of Journal of Sex Research, 39, 190–196.
Sexual Medicine, 2, 699–708. Sprecher, S., & Cate, R. (2004). Sexual satisfaction and sexual
Hassebrauk, M., & Fehr, B. (2002). Dimensions of relationship quality. expression as predictors of relationship satisfaction and stability.
Personal Relationships, 9, 253–270. In J. Harvey, A. Wenzel, & S. Sprecher (Eds.), Handbook of sexu-
Heiman, J. R., Talley, D. R., Bailen, J. L., Rosenberg, S. J., Pace, C. R., & ality in close relationships (pp. 235–256). Mahwah, NJ: Lawrence
Bavendam, T. (2007). Sexual function and satisfaction in hetero- Erlbaum.
sexual couples when men are administered sildenafil citrate (Viagra) Turner, A. N., Morrison, C. S., Padian, N. S., Kaufman, J. S., Behets, F.
for erectile dysfunction: A multicentre, randomized, double-blind, M., Salta, R. A., … Miller, W. C. (2008). Male circumcision and
placebo-controlled trial. British Journal of Obstetrics and Gynecol- women’s risk of incident chlamydial, gonococcal, and trichomonal
ogy, 114, 437–447. infections. Sexually Transmitted Diseases, 35, 689–695.
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of U.S. Census Bureau. (2005). Number, timing, and duration of marriages
marital quality and stability: A review of theory, methods, and and divorces: 2001. Current Population Reports, P70-97. Wash-
research. Psychological Bulletin, 118, 3–34. ington, DC: U.S. Government Printing Office.
Laumann, E. O., Gagnon, J. H., Michael, R. T., & Michaels, S. (1994). Yeh, H. C., Lorenz, F. O., Wickrama, K. A., Conger, R. D., & Elder, G.
The social organization of sexuality: Sexual practices in the United H. (2006). Relationships among sexual satisfaction, marital qual-
States. Chicago: University of Chicago Press. ity, and marital instability in midlife. Journal of Family Psychol-
Laumann, E. O., Paik, A., Kang, J.-H., Wang, T., Levinson, B., Nicolosi, ogy, 20, 339–343.
A., … Gingell, C. (2006). A cross-national study of subjective
sexual well-being among older women and men: Findings from the

123