Sexual obligation and perceived stress: A national longitudinal study of older adults

Abstract

Objectives:

This study examines the relationship between sexual obligation and perceived stress among older adults in the United States.

Methods:

Using longitudinal data from three waves of the National Social Life, Health, and Aging Project (NSHAP), our sample included 1,477 partnered, sexually-active respondents aged 57 to 85 at the baseline survey. We estimated mixed-effects models to test how feelings of sexual obligation are related to changes in perceived stress score.

Results:

Sexual obligation was positively associated with perceived stress score. The positive relationship between sexual obligation and perceived stress score became stronger over the study period among older men, although it remained relatively stable among older women. Relationship quality only partially explains this relationship.

Conclusions:

Feeling more obligated to have sex had a significantly greater effect on older men’s perceived level of stress over time than older women’s. This association became marginally significant after relationship quality was controlled for, suggesting that relationship quality was a key explanatory factor for the gendered patterns in sexual obligation’s linkage to stress. These results highlight the importance of understanding gendered sexuality among aging older adults within the context of their relationship.

Clinical Implications:

Older adults’ feelings of sexual obligation can manifest in their daily stress experience. Clinicians seeking to lower older adults’, in particular older men’s, stress levels should address the context of their sexual life and if they feel obligated to have sex, along with the positive and negative aspects of their relationship, as these could elevate their stress levels over time.

Introduction

According to a recent national study, 36% of Americans ages 57 to 85 who are sexually active feel pressured or obligated to have sex (Nowakowski & Sumerau, 2019). Although emerging research indicates that sexual involvement is beneficial for older adults’ mental health (Zhang & Liu, 2019) — including from decreased stress through intercourse (Brody, 2010) — this may depend on the context in which the sex is occurring. For example, health benefits of sex may be less likely to accrue for those who feel pressured or obligated to have sex than those who desire sex. Yet, this proposition has not been tested. In fact, little research has been conducted to understand the problem of sexual obligation — defined as having sexual intercourse, even if it is unwanted, because of a feeling of obligation to do so (Impett et al., 2008) — particularly among older adults and how it may affect stress in late life.

Drawing upon gender socialization theory, we analyze data from three waves of longitudinal data from the National Social Life, Health, and Aging Project (NSHAP) to address three major research questions: 1) how is feeling obligated to have sex related to older adults’ perceived stress? 2) does this relationship vary by gender? and 3) is the association between sexual obligation and perceived stress explained by sexual challenges or relationship quality? The importance of this study is highlighted by the high prevalence of sexual obligation (Lee et al., 2016; Nowakowski & Sumerau, 2019; Waite et al., 2009) among older adults, perhaps higher than among younger adults due to increased sexual problems faced in later life (Lindau et al., 2007). Although stress tends to decline with age (Weissman et al., 2015), the negative health impact of stress may increase with age, especially for older adults who experience evoking stressors and cannot attenuate the stress or who become frail in late life (Charles, 2010; Desrichard et al., 2018). Levels of elevated stress are related to a variety of health problems for older adults, including cardiovascular disease, cancer, stroke, and poor immune response (Hamarat et al., 2001). Results from this study speak to clinical practices regarding both the sexual well-being and mental health of the aging U.S. population.

Sexual obligation among older adults

The desire to have sex tends to decline towards the end of life which interrupts older adults deriving pleasure from sex (Camacho & Reyes-Ortiz, 2005). As they age, older women have less desire than older men to continue the sexual relationship (Kontula & Haavio-Mannila, 2009). At the same time, while the physical changes that come from aging can make sex more difficult, older adults (more so for older women than older men) may still feel pressure to continue their sexual activity out of obligation in order to meet their partner’s sexual needs (DeLamater & Friedrich, 2002; Nowakowski & Sumerau, 2019; Traeen, 2008). They may have sex out of obligation even if it is painful or more difficult. This can come from cultural pressures to be more youthful with a healthy sex drive despite physical changes, or wanting to be a pleasing romantic partner.

A limited number of studies have quantitatively examined sexual obligation among older adults, with a focus on documenting differences in prevalence of sexual obligation across groups in heterosexual relationships. A national, cross-sectional study of 3,005 older adults found that older women were more likely than older men to report feeling obligated to have sex (Nowakowski & Sumerau, 2019). Another study of heterosexual couples in Norway between ages 22 and 67 found that while more women than men had sex out of obligation at all ages, those men who did have sex out of obligation were more likely to talk with their partner about their sexual desires compared to the women who had sex out of obligation (Træen & Skogerbø, 2009). Further, another study of a representative sample of U.S. older adults 62–91 years old examined how feeling obligated to have sex varies by cognitive status and found that reported sexual obligation was similar across cognitive groups for older women, but higher for older men with dementia compared to older men with normal cognition (Lindau et al., 2018).

Qualitative studies on this topic offer some insight into the decisions behind having sex out of obligation, mostly outside of the United States. Interviews of six women ages 60 and older living in rural Brazil revealed that these women perceived sexually pleasing their husband as part of their marital obligation (Baldissera et al., 2012), while another study of eighteen Taiwanese women ages 45–60 showed that both before and after menopause, women felt obligated to have sex with their partner, even if they did not desire sex at the time, in order to maintain the sex life in their own relationship (Yang et al., 2015). These studies indicate that obligatory sex is tied to one’s role as a partner and having sex is perceived as one characteristic of being a romantic partner or spouse (Culbertson & Dehle, 2001). Nevertheless, these limited studies, both qualitative and quantitative, inform little about the impacts that sexual obligation has on individuals’ well-being.

Sexual obligation and mental health: Limited empirical evidence

Previous studies on how sexual obligation is linked to mental well-being are limited and mostly focus on college-aged samples, suggesting that college students do have sex out of obligation and it is associated with different mentally-straining outcomes. For example, interviews of partnered college students ages 18–24 revealed that some students obliged to sex in order to avoid feelings of guilt about not wanting to have sex (Vannier & O’Sullivan, 2009). Another study of partnered college students found a positive association between having sex out of obligation and greater sexual conflict (Long et al., 1996). College women who had ever had sex out of obligation in their partnership were also shown to be less satisfied with their relationship compared to college women who did not (Katz & Tirone, 2009).

There is a paucity of evidence regarding sexual obligation linked to mental health in late life. For example, a cross-sectional analysis of adults ages 50 and older in England found that feeling obligated to have sex was associated with men’s greater worry about their sex life and women’s lower sexual satisfaction and greater depression (Lee et al., 2016). While our understanding of this topic is limited to cross-sectional evidence with little known about its ties to various mental health outcomes over time, previous studies on the linkages between general sexual activity and stress in late life are insightful. For example, Allen (2017) analyzed a representative sample of English adults who were 66 years old, on average, and found that more frequent sexual activity is related to reduced stress biomarkers. Zhang and Liu’s (2019) analysis of a national representative sample of older adults ages 57 to 85 found that dissatisfaction with sexual frequency was related to higher levels of psychological distress and unhappiness. Medical scientists view sex as a form of moderate exercise, arguing that having sex may reduce stress in the same way exercise does (Levin, 2007; Salmon, 2001). Furthermore, sexual activity can release hormones that promote feelings of social attachment which can decrease feelings of stress (Carter, 1998). However, if the sexual activity is obligatory rather than desired, those beneficial links to stress could be absent.

A gendered socialization perspective on sexual obligation and mental health

Despite a paucity of empirical evidence, there is a theoretical basis to expect that feeling obligated to have sex would be stressful for both older men and women in different ways. Gender socialization theory explains that from a young age, boys and girls learn how to behave in ways socially appropriate according to their gender, and these differences shape their identity as men and women which continue across the life course (Carter, 2014). This socialization explains how men and women act differently in sexual situations and thus experience different impacts on their well-being, as being socialized to be masculine or feminine influences the role they play in their romantic relationships (Impett & Peplau, 2003). For example, the current generation of older adults came of age during a time when traditional gender roles were common and seen as an ideal. These gender roles permeated into sexual relationships, where men had more authority and women were expected to be submissive. Women were socialized to follow men’s sexual desire (Stein, 1989), and a greater cultural expectation falls on women to respond to men’s sexual arousal, even if they do not want to do so, rather than men responding to women’s (Basson, 2000). Consistent with this view, research indicates that women often have sex because they feel it is their obligation or duty to their partner (Impett et al., 2008).

On the other hand, men face more social and cultural pressures regarding their sexual performance (Hirayama & Walker, 2011). Men are socialized to be sexually successful. A vital aspect of masculinity involves sexual performance, and sexual activity remains important to men’s self-identify even as they age (Thompson & Barnes, 2013). However, men’s testosterone can start decreasing in their 40s and they may begin to experience sexual dysfunctions as they get older (DeLamater & Friedrich, 2002). Still, with the increased popularity of erectile dysfunction medications to help alleviate that problem, older men may feel more pressure to continue to perform well sexually (Lodge & Umberson, 2012). Sexual performance and its tie to masculinity may make men feel obligated to have sex even if it is more difficult than when they were younger.

For both men and women, the physical changes they face in later life may make sex uncomfortable, and this physical challenge combined with feeling obligated to have sex could lead to greater feelings of stress through increasing the risk of sexual challenges and lowering relationship quality.

Sexual challenges.

Older adults experience a range of sexual challenges including a lack of interest in sex, men’s erectile dysfunction, or women’s lubrication problems (Laumann et al., 1999; Lindau et al., 2018). Feeling obligated to have sex may increase the risk of sexual challenges by decreasing the satisfaction and pleasure from one’s sex life (Lee et al., 2016). Moreover, individuals who experience a sexual dysfunction may still feel obligated to have sex despite their sexual challenge in order to please their partner, which may worsen their sexual challenges (Cacchioni, 2007; Traeen & Skogerbo, 2009). Experiencing sexual challenges increases stress. For instance, Laumann and colleagues (2008) found that lack of interest in sex for both older men and women, along with women’s lubrication difficulty, were related to perceived stress.

Relationship quality.

Relationship quality, which has both positive and negative dimensions that reflect assessments of the relationship in terms of aspects like satisfaction, happiness, strain, and conflict (Liu & Waite, 2014), may be another pathway through which sexual obligation shapes levels of stress. Older couples with better sexual lives, in terms of more frequent sex (Galinsky & Waite, 2014) and more sexual satisfaction (Schoenfeld et al., 2017), are often in relationships that have better overall relationship quality. Therefore, as feeling obligated to have sex may be indicative of worse sexual lives, this may promote worse relationship quality (Traeen & Skogerbo, 2009). Further, poor relationship quality has been shown to increase exposure to stress. Holt-Lunstad and colleagues (2008) found an inverse relationship between marital quality and perceived stress, suggesting that lower levels of marital adjustment and satisfaction predicted greater levels of perceived stress. Relationship conflict is harmful to the psychological well-being of married men and women (Williams, 2003), and adults in relationships characterized as unhappy had worse subjective well-being, including more depressive symptoms, than those in happy relationships (Dush & Amato, 2005).

Research hypotheses

• Hypothesis 1: Feeling obligated to have sex is related to greater perceived stress over time.

• Hypothesis 2: The association between sexual obligation and perceived stress is partially mediated through sexual challenges.

• Hypothesis 3: The association between sexual obligation and perceived stress is partially mediated through relationship quality.

Guided by the gender socialization theory, we also examine potential gender differences in this relationship.

Methods

Data and Sample

We utilized the first three waves of longitudinal data from the National Social Life, Health, and Aging Project (NSHAP) collected in 2005/06, 2009/10, and 2015/16, respectively. NSHAP was collected by the National Opinion Research Center (NORC) at the University of Chicago, and it is a nationally representative, population-based dataset of older adults in the United States. The first wave (2005/6) surveyed a national probability sample of 3,005 community-dwelling adults ages 57–85 in the United States (Waite, Laumann, et al., 2014). In the second wave (2010/11), 3,377 respondents completed the survey, including 2,261 Wave 1 respondents who were re-interviewed (Waite, Cagney, et al., 2014). The third wave (2015/16) survey was completed by 4,777 respondents, including 2,350 respondents from Wave 2 and 2,250 new refreshment respondents (Waite, 2017). To take advantage of the longitudinal nature of the data, we restricted our analytic sample to the 3,005 original respondents who participated in the survey starting at Wave 1.

The NSHAP questionnaire consisted of two parts, the in-person questionnaire (IPQ; the face-to-face interview) and the leave-behind questionnaire (LBQ; a self-administered survey). About five out of six respondents completed the LBQ. Because the key measures of perceived stress and sexual obligation were only asked in the LBQ, the current analysis was restricted to respondents who completed the LBQ. Results from preliminary analysis (results not shown but available upon request) suggested that for all waves, the LBQ sample was more educated, had more partnered respondents, and had more white respondents than the IPQ sample. We further restricted the sample to respondents who were partnered (i.e. either married, cohabiting, or had a romantic partner), who had had sex in the past 12 months, and who had no missing reports on measures of stress and sexual obligation. The final analytic sample included 1,477 sexually active heterosexual respondents who contributed to 2,484 person-period observations across the three waves.

Measures

Perceived stress (time-varying).

The NSHAP Perceived Stress Measure (NPSM) reflects Cohen and Williamson’s (1988) 4-item scale, which is a condensed form of the original 14-item Perceived Stress Scale (PSS) (Payne et al., 2014). The four questions asked respondents to indicate how often during the past week they (1) were unable to control important things in their life, (2) felt confident about their ability to handle personal problems, (3) felt that things were going their way, and (4) felt that difficulties were piling up so high they could not overcome them. The response categories for all four questions were 0=rarely or none of the time, 1=some of the time, 2=occasionally, and 3=most of the time. All items were coded with a high score indicating a greater level of perceived stress. These items were summed, with a possible range on the scale from 0 (lowest perceived stress) to 12 (highest perceived stress). The NPSM scale has been well used in previous studies (Birditt et al., 2014; Kotwal et al., 2012; Upenieks et al., 2016) despite its relatively low alpha value of 0.63. We note that Cronbach’s alpha may be misleading when estimating inter-item correlations in the case of a small number of indicators or categorical response options such as the NPSM scale (Raykov, 1997). Cronbach’s alpha is very sensitive to the number of items being estimated, and tends to be low when two to five indicators are being estimated, even though the internal consistency is high (Streiner, 2003).

Sexual obligation (time-varying).

NSHAP asked respondents how often they had sex mainly because they felt obligated or that it was their duty during the past 12 months. All three waves had Likert-type scale answer options ranging from never to always. We recoded sexual obligation as a dichotomous measure to distinguish between very low levels of sexual obligation that may be signs of a healthy relationship in contrast to higher levels of sexual obligation. Sexual obligation is coded as never/rarely (0) and sometimes/usually/always (1).

Sexual challenges (time-varying).

We tested two measures of sexual challenges which may be problematic for older adults as potential mediators in our model. First, NSHAP asked respondents if they lacked interest in sex (0=no, 1=yes). Second, respondents also reported if they experienced gender-specific sexual dysfunctions. Men indicated whether or not they had trouble getting or maintaining an erection, and women indicated if they had trouble lubricating. We combined these responses to create a measure of sexual dysfunction, with those reporting they did experience the dysfunction (1) versus those who did not (0).

Relationship quality (time-varying).

We followed previous studies of NSHAP data to create measures of relationship quality (Liu & Waite, 2014; Zhang & Liu, 2019) and tested them as potential mediators. There were six relationship measures consistently asked in all waves that we used to generate relationship quality scales. First, respondents reported their relationship happiness, with answers collapsed into unhappy (1), happy (2), and very happy (3). Second, respondents reported if they spent their free time together or apart from their partner, and answers were reversely coded to be mostly apart (1), some together and some apart (2), and mostly together (3). Third, respondents reported how often they could open up to their partner, with answers of never, hardy ever, or rarely (1), some of the time (2), and often (3). Fourth, respondents reported how often they could rely on their partner, with answers of never, hardy ever, or rarely (1), some of the time (2), and often (3). Fifth, respondents reported how often their partner made too many demands on them, with answers of never, hardy ever, or rarely (1), some of the time (2), and often (3). Sixth, respondents reported how often their partner criticized them, with answers of never, hardy ever, or rarely (1), some of the time (2), and often (3). Results from factor analysis indicated these six measures yielded two dimensions, which we refer to as positive and negative relationship quality. Table 1 reports the factor loadings for each measure used to create the factor scores for the two continuous relationship quality measures.

Table 1.

Relationship quality factor loadings, NSHAP Waves 1–3

	Wave 1		Wave 2		Wave 3
	POS	NEG	POS	NEG	POS	NEG
How happy is your relationship?	0.49	−0.17	0.55	−0.11	0.45	−0.22
How often can you open up to your partner?	0.69	0.05	0.65	−0.03	0.67	0.00
Do you spend your free time together?	0.37	−0.06	0.45	0.05	0.34	−0.10
How often can you rely on your partner?	0.63	0.02	0.59	0.03	0.70	0.05
How often does your partner criticize you?	0.05	0.78	−0.05	0.62	0.04	0.87
How often does your partner make too many demands on you?	−0.11	0.52	0.05	0.68	−0.13	0.47

Note: POS=positive relationship quality; NEG=negative relationship quality; bold values indicate factor loadings above the 0.35 cutoff

Sociodemographic covariates.

We controlled for both time-varying and time-invariant sociodemographic covariates which were suggested to be associated with both stress and sexual life (Aggarwal et al., 2014; Cohen & Williamson, 1988; DeLamater, 2012; Ezzati et al., 2014; Laumann et al., 1999; Payne et al., 2014). Time-varying covariates included age (in years, centered at the mean), marital status (0=married or cohabiting and 1=unmarried), self-rated health (from 1=poor to 5=excellent, centered at the mean), and employment status (0=not currently working and 1=currently working). Time-invariant covariates (all measured at Wave 1) included race/ethnicity (non-Hispanic white (reference), non-Hispanic black, Hispanic, and others), education (less than high school (reference), high school degree, some college but no degree, and college degree or above) and gender (0=men and 1=women).

Analytic Approach

We estimated multilevel mixed-effects models to handle the nested distribution of the longitudinal data. A major advantage of multilevel mixed-effects models compared to traditional regression analysis is the ability to distinguish the two types of heterogeneity (within- and between-individual) in estimates of the effects of the sexuality measures on changes in mental health. Mixed-effects models account for the unobserved heterogeneity related to the nested distribution of the longitudinal data by allowing random effects to vary across individuals. The linear mixed-effects model was specified as:

$$\text{Level}1:{\text{Y}}_{\text{ij}}={\pi }_{\text{0i}}+{\pi }_{\text{1i}}{\text{T}}_{\text{ij}}+\gamma S_{ij}+Z_{ij}’C+{\epsilon }_{\text{ij}}$$

$$\text{Level}2:{\pi }_{\text{0i}}={\beta }_{\text{00}}+X_i’A_0+{\delta }_{\text{0i}}$$

$${\pi }_{\text{1i}}={\beta }_{10}$$

where Y_ij is ith individual’s stress at time j. T_ij is the survey year (centered at the baseline year of 2005). π_0i is the initial stress for the ith individual at Wave 1 (random effect). π_1i is the slope of change in perceived stress score over time for the ith individual. For parsimony, we estimated π_1i as a fixed effect rather than a random effect because the preliminary analysis suggested little variation in this parameter and adding a stochastic component for π_1i did not improve the model. S_ij indicates the time-varying sexual obligation variable included in the model and γ is the corresponding coefficient—the focus of our interpretation. X_i is the vector of Wave 1 time-invariant covariates and Z_ij is the vector of time-varying covariates. ε_ij is the level-1 residual (within-individuals) and δ_0i is the level-2 residual (between-individuals).

We estimated six mixed-effects models. Each model controlled for all sociodemographic covariates. In Model 1, we tested the basic association between sexual obligation and stress. In Model 2, we added an interaction term of sexual obligation and analytic time (i.e., survey year) to test whether and how perceived stress score changed across time in response to sexual obligation. In Model 3, we tested potential gender differences in the effect of sexual obligation on stress by adding the interaction of gender and sexual obligation as well as the three-way interaction of sexual obligation, analytic time and gender. In Models 4 and 5, we added the sexual challenge and relationship quality factors separately to test if they explain the relationship between sexual obligation and perceived stress over time. Model 6 included all covariates. All analysis was conducted using Stata 15 (StataCorp, 2017).

Results

Table 2 shows the descriptive statistics of all analytic variables in the sample. Respondents reported low levels of perceived stress across all waves (on a scale ranging from 0 to 12, perceived stress was 1.53 at Wave 1, 3.01 at Wave 2, and 2.51 at Wave 3). At Wave 1, about 20.31% (28.91% for women and 14.71% for men) of the sample reported feeling obligated to have sex either sometimes, usually, or always, with the remaining 79.69% (71.09% for women and 85.29% for men) reporting sexual obligation occurring either never or rarely¹. Just over a quarter of the sample (25.47%) reported they lacked interest in sex at Wave 1, and this increased across waves. 30.78% of the sample reported experiencing sexual dysfunction at Wave 1, and this increased to 40.80% in Wave 3. Relationship quality was relatively stable across waves, with positive relationship quality ranging from 0.06 to 0.11 across waves and negative relationship quality ranging from −0.04 to −0.11 across waves. The majority of the sample was male and either married or cohabiting (92.42% at Wave 1). A majority of the respondents were non-Hispanic white, and most had at least a high school degree or above. On average, respondents reported their health to be between good and very good across all waves, and a majority were not currently working (66.91% at Wave 1).

Table 2.

Descriptive statistics, NSHAP waves 1–3 (N of person-period = 2484, N of respondents =1477)

	Mean(SD) or %			N
	Wave 1	Wave 2	Wave 3
Perceived Stress	1.53 (2.09)	3.01 (2.72)	2.51 (2.40)
Obligated to Have Sex
Never/rarely	79.69	76.06	84.57	1976
Sometimes/usually/always	20.31	23.94	15.43	508
Gender
Men	60.55	64.57	63.85	1549
Women	39.45	35.43	36.15	935
Marital Status
Married/cohabiting	92.42	92.20	90.91	2287
Not married/cohabiting	7.58	7.80	9.09	197
Race/ethnicity
Non-Hispanic White	76.56	76.61	75.48	1897
Non-Hispanic Black	10.78	9.71	11.84	265
Hispanic	9.92	10.26	10.36	251
Other	2.73	3.42	2.33	71
Education
Less than high school	15.55	12.59	12.05	348
High school degree	25.47	22.02	20.72	585
Some college	30.39	32.69	32.35	781
College degree or above	28.59	32.69	34.88	770
Age	66.91 (7.19)	70.35 (6.38)	74.17 (5.49)
Self-rated health	3.46 (1.06)	3.48 (1.00)	3.48 (0.97)
Employment Status
Not currently working	60.63	71.14	79.28	1671
Currently working	39.38	28.86	20.72	813
Lack sexual interest
No	55.23	63.47	63.64	1472
Yes	25.47	33.11	29.39	707
Missing	19.30	3.42	6.98	305
Experience sexual dysfunction
No	49.30	49.66	53.07	1245
Yes	30.78	44.19	40.80	910
Missing	19.92	6.16	6.13	329
Positive relationship quality	0.06 (0.78)	0.11 (0.75)	0.06 (0.77)
Negative relationship quality	−0.04 (0.81)	−0.11 (0.74)	−0.05 (0.85)

Table 3 shows the estimated regression coefficients from the linear mixed-effects models predicting perceived stress from obligation to have sex. Model 1 tested the basic association between sexual obligation and stress, and the results indicated a positive relationship between feeling obligated to have sex and perceived stress (β = 0.42, p = 0.001), holding all other covariates constant. Specifically, every one unit increase in sexual obligation was associated with a 0.42 unit increase in perceived stress score. Model 2 tested the interaction between sexual obligation and survey year in order to understand how the relationship between sexual obligation and perceived stress score changed across time, revealing no significant change in this association over time. However, our further examination of gender differences in perceived stress score in relation to sexual obligation over time in Model 3 suggested that there was an increasing association between sexual obligation and perceived stress score over time and it was more profound for men than for women, indicated by the significant three-way interaction effect of sexual obligation, survey year and gender (β = −0.13, p = 0.043). To better illustrate this three-way interaction effect, we graphically present the results of Model 3 in Figure 1. From Figure 1, we can clearly see some patterns. First, perceived stress score increased over time for older men and older women with and without reported sexual obligation. However, men who felt obligated to have sex showed much steeper increases in perceived stress than any other group. At the end of the study period, the stress score was more than a point (1.20 points) higher for men who felt sexual obligation than men who did not feel sexual obligation. Further, women with sexual obligation had lower perceived stress scores than women without sexual obligation. Although this difference was not in the expected direction, the gap in stress between women who felt sexual obligation and women who did not feel sexual obligation was small and remained quite stable over the entire study period.

Table 3.

Estimated regression coefficients from mixed-effects models predicting perceived stress from obligation to have sex, NSHAP waves 1–3

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Obligated to have sex (ref: never/rarely)	0.42***	0.33*	0.10	0.11	0.14	0.15
Obligated*Year		0.03	0.11*	0.10*	0.09*	0.09*
Obligated*Women			−0.40	−0.37	−0.31	−0.27
Year*Women			0.00	−0.00	−0.00	−0.00
Obligated*Year*Women			−0.13*	−0.12*	−0.12+	−0.12+
Year	0.14***	0.14***	0.14***	0.14***	0.14***	0.14***
Women	0.25*	0.25*	0.64*	0.58*	0.59*	0.54*
Lack sexual interest (ref: no)
Yes				0.22*		0.14
Missing				−0.19		−0.43
Experience sexual dysfunction (ref: no)
Yes				−0.07		−0.08
Missing				0.31		0.40
Positive relationship quality					−0.28***	−0.28***
Negative relationship quality					0.24***	0.24***
Race (ref: Non-Hispanic White)
Non-Hispanic Black	0.14	0.14	0.15	0.14	0.02	0.02
Hispanic	0.19	0.19	0.18	0.18	0.13	0.14
Other	0.14	0.14	0.15	0.17	0.15	0.16
Education (ref: <high school degree)
High school degree	−0.06	−0.06	−0.06	−0.04	−0.11	−0.09
Some college	−0.30+	−0.31+	−0.31+	−0.28	−0.35*	−0.32+
College or above	−0.67***	−0.68***	−0.69***	−0.67***	−0.74***	−0.72***
Age	−0.00	−0.00	−0.00	−0.00	−0.00	−0.00
Marital Status (ref: married/cohabiting)	0.22	0.22	0.22	0.26	0.16	0.18
Self-rated health	−0.42***	−0.42***	−0.42***	−0.41***	−0.40***	−0.40***
Employed	0.08	0.08	0.08	0.08	0.06	0.05
Intercept	1.84***	1.86***	1.87***	1.79***	1.94***	1.91***
Within individual variance	4.18***	4.17***	4.15***	4.15***	4.10***	4.09***
Between individual Variance	1.07***	1.07***	1.09***	1.07***	0.97***	0.96***
Number of observations	2,484	2,484	2,484	2,484	2,466	2,466
Number of respondents	1,477	1,477	1,477	1,477	1,467	1,467

^***p<0.001,

^**p<0.01,

^*p<0.05,

⁺p<0.10

Figure 1.

Predicated Levels of Perceived Stress and Sexual Obligation Interacting with Survey Year and Gender

The final three models tested the mediating variables. Model 4 added the sexual challenges measures of lack interest in sex and experience sexual dysfunction, and the results indicated little change in the estimated effects of sexual obligation (compared with Model 3). The three-way interaction effect remained significant in Model 4 (β = −0.12, p = 0.045) — suggesting that these sexual challenges did not mediate the relationship between sexual obligation and stress over time by gender. Model 5 examined relationship quality as potential mediating factors, and the results show that the three-way interaction effect became marginally significant (β = −0.12, p = 0.051, bootstrap 95% CI [−0.26, 0.02]) after relationship quality factors were controlled — suggesting that the relationship between sexual obligation and stress over time by gender as illustrated in Figure 1 was partially explained by relationship quality. Results in Model 5 also suggest that positive relationship quality was related to lower perceived stress score while negative relationship quality was related to higher perceived stress score. In Model 6, when all covariates were added together, the three-way interaction effect remained marginally significant (β = −0.12, p = 0.054, bootstrap 95% CI [−0.28, 0.05]).

Discussion

Sex remains important to the lives of a significant proportion of older adults (Lindau et al., 2007). Previous research suggests that sex can be beneficial for older adults’ physical (DeLamater, 2012) and mental (Zhang & Liu, 2019) health, although this may vary by gender (Liu et al., 2016). Leading scholars posit that it is important to consider not only if or how often sex occurs, but also the context of that sex—often overlooked in previous literature (Christopher & Sprecher, 2000; Rosen & Bachmann, 2008; Schoenfeld et al., 2017). The present study is the first to examine how feelings of sexual obligation affect perceived stress in older adults. Consistent with our hypothesis, we find that higher levels of sexual obligation were significantly associated with higher levels of perceived stress; and, this relationship became stronger over the study period among older men, although it remained relatively stable among older women – this pattern was partially explained by relationship quality. These results raise some important concerns for the sexual lives of older adults, in particular older men.

While few previous studies have examined whether and how sexual obligation matters for perceived stress levels, especially among older adults, our finding of a positive association between sexual obligation and perceived stress is consistent with a broader literature on the negative mental health impacts of sexual disorders (Atlantis & Sullivan, 2012). Sexual obligation is one type of sexual experience that can be linked to low sexual desire (Træen & Skogerbø, 2009) or changing sexual attitudes (Waite et al., 2009). Even though reports of sexual obligation are present into older age (Waite et al., 2009), these types of sexual issues may receive less attention in older populations due to stereotypes about older adults having little to no sex (Syme, 2014).

The association between sexual obligation and stress may result from cultural expectations that a relationship includes sexual intimacy (Muehlenhard & Cook, 1988). For example, sex is an expectation of marriage, but that implication can take excitement out of the sexual relationship and lead to sex feeling more obligatory (Sims & Meana, 2010; Wood et al., 2007). Partners are expected to desire sex, and when this is not met, a person may simulate sexual desire in order to meet their partner’s need. However, this can be seen as additional work that must be done in order to avoid relationship conflict or boost a partner’s self-esteem (Elliott & Umberson, 2008). If sex is viewed as work rather than for pleasure, that may be interpreted as a failure of the relationship and reflect poorly on the relationship quality (Kingsberg, 2002). The ramifications of needing to feign desire in order to keep the relationship alive or meet a partner’s sexual needs rather than one’s own could be distressing.

Moreover, while we find that the association between sexual obligation and elevated stress increased over time for all groups in the sample, it is men with high levels of sexual obligation who experience the steepest increase in stress. This result is supported by gender and age patterns in sexual obligation found in previous work on adults 50 and older where an analysis by decade showed that among women, rates of more frequent sexual obligation was highest in young-old age groups and lowest in the oldest-old group; for men the opposite results were seen, with sexual obligation highest among the oldest-old group (Lee et al., 2016). Still, research shows that more women than men report having sex out of obligation (Lee et al., 2016; Nowakowski & Sumerau, 2019) — a pattern confirmed in our sample, where a significantly greater percent of women than men reported high levels of sexual obligation. However, our results suggest that the mental health implications of that obligation are more prominent for older men than older women over time.

Sexual obligation may have a weaker but stable effect on older women’s stress levels because they are used to dealing with this pressure throughout their life course, so it may be more normalized and not induce elevated levels of stress over time (Yang et al., 2015). Indeed, the results show the difference in perceived stress score for women who did and did not experience sexual obligation was stable across time, so this aspect of their sexual lives may be more routine compared to older men. Our findings also indicated that perceived stress scores were slightly higher for women who did not experience sexual obligation compared to women who did. It may be that by this later life stage, women have become accustomed to sexual obligation because of gendered expectations in sexual behavior throughout their lives, including learning to appease their partner sexually or not voicing their sexual desires (Gagnon & Simon, 2005), so the link to stress is not as prominent.

While sex out of obligation may be more normalized for older women (Nowakowski & Sumerau, 2019), it may be new to men in later life. Men are socialized to be sexually successful, and they are expected to want sex (Muehlenhard & Cook, 1988). By integrating this gender role expectation into their own sense of identity, older men may either feel pressured or put pressure on themselves to have sex, although they may feel less pressured in partnerships with better relationship quality, as suggested by our results. While masculinity can change across the life course, our findings highlight that sexual activity is still a significant part of older men’s identity — which may create stress especially for older men who may feel sexually obligated.

Stress as a result of older men’s sexuality can come from several related sources. For example, men may feel pressured to be able to perform sex because medical interventions, such as Viagra or Cialis, are heavily advertised as solutions to sexual dysfunction (Lodge & Umberson, 2012). Physically, it is more difficult for men to pretend to be sexually aroused than women (Træen & Skogerbø, 2009), and pharmaceutical drugs do not always facilitate the desired results (Gledhill & Schweitzer, 2013). In contrast, cultural representations perpetuate masculinity, with advertisements pushing men’s virility into old age (Calasanti & King, 2005; Marshall, 2006). Therefore, older men may be distressed about upholding their earlier sexual ability — even if it means seeking medical intervention — and they may feel obligated to have sex as it is a reflection of their masculinity. This setting could lead to men having greater stress from their obligatory sex, especially if it perpetuates over time.

Finally, the significant interaction effect of gender, sexual obligation and time becomes marginally significant after relationship quality is controlled in the model, suggesting that relationship quality is a mediator in explaining the gendered linkages between sexual obligation and stress over time. Older adults, particularly older men, who experience sexual obligation are more likely to experience worse relationship quality, which is harmful for their level of perceived stress over time. It is likely that sexual obligation can lead to older men’s feelings of guilt, resentment, or displeasure toward their partner (Traeen, 2008), which is harmful for relationship quality. Poor relationship quality is found to be related to higher levels of stress (Kiecolt-Glaser & Glaser, 2001; Umberson et al., 2006). Additionally, if poor relationship quality continues over time, the greater stress level may become a chronic experience, which has consequences for multiple body systems, especially for older couples (Kiecolt-Glaser & Newton, 2001).

Limitations

This study is limited in several ways. Notably, we only have information from one partner. This is largely a limitation of the dataset, as partner data was not available across all three waves of data collection. Future studies should seek data from both partners across all waves to explore the dyadic relationships of one or both partners’ sexual obligation experiences as it relates to the mental well-being of the couple. Second, with only three waves of data, we are cautious in making causal claims. The upcoming release of the fourth wave of NSHAP data will enable additional work on a longer trajectory to test the robustness of our findings and better discern the causal direction. Third, NSHAP did not oversample sexual minority groups, and a vast majority of the respondents are in heterosexual partnerships. NSHAP only collected sexual identity in Wave 3. Therefore, there is a heteronormative assumption in the results. A growing number of research has shown that lesbian, gay, and bisexual adults suffer worse mental well-being compared to heterosexual adults (Hsieh & Liu, 2020). Future work should investigate whether and how sexual obligation relates to stress differently for same-sex and different-sex couples. Finally, we did not include information on sexual history. Previous qualitative research indicates that having sex out of obligation rather than desire occurs in long-term relationships (Elliott & Umberson, 2008; Sims & Meana, 2010), suggesting that this obligation can occur for years. Unfortunately, NSHAP only inquired about sexual obligation in the past 12 months, limiting our ability to consider sexual obligation which may have been occurring over a much longer time period or for the majority of the respondent’s relationship and could also affect stress levels. These different aspects would be important to pursue in future research.

In conclusion, a significant proportion of older adults do still feel obligated to have sex with their partner. Although these feelings may come from wanting to please their partner or to maintain their earlier sex life or identity, our results suggest that sexual obligation may manifest in the daily stress experience, partially through damaging their relationship quality. This is concerning for both the quality of the sexual relationship and for the health impacts that such stress can have, particularly in the older population. Additionally, as this is the first longitudinal study to examine sexual obligation and stress over time, we reveal an important gender difference that may be overlooked in cross-sectional analyses. Namely, this relationship becomes stronger over time among older men but not older women. Clinicians or therapists looking to manage older adults’ health levels, particularly their stress levels, should consider asking about the context of their sexual relationship. It should also not be assumed that sexual desire is universal for all men at any age, and clinicians or therapists instead should inquire about how sexual obligation manifests for older men as a risk factor for elevated stress levels. Interventions to improve the sexual lives of older adults, particularly older men, can be beneficial for their overall well-being.

Clinical Implications.

• Clinicians should inquire about the changing nature of older patients’ sexual experiences and if and how the partners are communicating with each other about their sexual desire.

• Because the relationship between higher sexual obligation and greater perceived stress increases over time, clinicians should be concerned earlier with the quality of their patients’ sexual relationship due to the harm that compounding high stress levels have on health.

• Clinicians wanting to lower older men’s stress levels in particular should address why they are feeling obligated to have sex, as this could elevate their stress levels as they age.