Predictors of duty sex frequency in women

Abstract

Background

Duty Sex—the act of engaging in sex out of a sense of duty or obligation to a partner—is a commonly reported reason why women have sex, with studies indicating associations between Duty Sex and sexual dysfunction, as well as nonconsensual sexual experiences (NSEs).

Aim

This study sought to examine the associations between Duty Sex frequency, sexual function, and NSEs using comprehensive, validated measures of sexual function and NSE histories in a large community sample of women.

Methods

Six hundred and fifty-eight women with (n = 293) and without (n = 365) NSE histories completed measures of sexual function and sexual motivations.

Results

Women with NSE histories reported more sexual pain, lower sexual satisfaction, and more frequent Duty Sex. The relation between NSE history and Duty Sex remained after controlling for the variance in sexual function. The type of NSE behavior, but not the timing of the first NSE nor the perceived impact of the experience, predicted Duty Sex frequency.

Clinical implications

Clinicians should consider sexual motivations, particularly Duty Sex, when treating women with NSE histories or who experience sexual dysfunction.

Strengths and Limitations

To our knowledge, this was the first study to examine how NSEs relate to all domains of sexual function and sexual motives, revealing a link between NSEs, sexual function, and Duty Sex. However, we were unable to account for what percentage of sexual encounters are considered Duty Sex. Therefore, our measurement of Duty Sex lacks specificity, which may explain the relatively low variance accounted for by the models in our analyses.

Conclusions

There are consequences of NSEs, including difficulties with communicating sexual boundaries and impairment of sexual function, that may contribute to the increased likelihood of engaging in Duty Sex.

In a series of studies, Meston and Buss¹ delineated 237 distinct reasons why humans have sex. Factor analyses of these reasons revealed four overarching factors (physical, emotional, goal attainment, and insecurity) and 13 subfactors. One commonly endorsed subfactor that emerged within the insecurity factor was duty/pressure motivations for sex. In their study, Meston and Buss¹ characterized “Duty Sex” by the following reasons for having sex: “I felt obligated to,” “I felt like it was my duty,” “I felt like I owed it to the person,” “It was expected of me,” “I didn’t want to disappoint the person,” and “It was just part of the relationship routine.” Earlier, Cooper et al.² described a similar construct they termed “Partner Approval,” defined by the following four items: how often do you have sex… “out of fear that your partner will not love you,” “to prevent your partner from being angry with you,” “because you worry that your partner will not want to be with you,” and “because you are afraid that your partner will leave you if you do not have sex.” This notion of engaging in sex that is not primarily driven by sexual pleasure or desire but, rather, out of a sense of duty or obligation to a partner or relationship has gained little research attention to date. This is surprising given that “Duty Sex” is a commonly reported sexual motivation among both heterosexual and non-heterosexual women,^1,3–5 with its prevalence peaking among women aged 55–83.⁵

Another similar construct to Duty Sex is sexual compliance. Sexual compliance refers to unwanted or undesired but consensual sex.⁶ While there is some overlap, the definition of sexual compliance emphasizes the degree of “wantedness” of the sexual encounter, while Duty Sex is one of many empirically derived motives¹ that motivate a sexual exchange. Individuals may acquiesce to unwanted sex for many reasons other than a sense of duty or obligation, making sexual compliance a broader definition. Sex may be unwanted due to lack of sexual desire or arousal, fear of pregnancy, feeling as though sexual activity is premature in the relationship,⁷ lack of attraction to the partner, or contextual factors like the location, timeframe, and privacy.¹ In contrast, someone may feel a sense of duty to have sex despite not wanting to because they experience sexual pain but also want to have sex to feel closer to their partner. Thus, it is more difficult to characterize the degree of “wantedness” present in this sexual exchange than to characterize the motivations—being duty and intimacy—that underly it.

Several questions relating to Duty Sex have important clinical implications and warrant research attention: (1) How does engaging in Duty Sex impact the sexual relationship—both long-term and short-term—and are there factors that mediate these outcomes? (2) How does engaging in Duty Sex impact the individual in terms of personal variables, such as self-esteem, well-being, negative versus positive affect, and emotionality? (3) Are there factors that predict which women are more likely to engage in Duty Sex? The current study focuses on this third question by examining whether previous nonconsensual sexual experiences (NSEs) and sexual function predict Duty Sex engagement.

While a link between sexual dysfunction and Duty Sex in women seems intuitive, few researchers have examined this relationship using psychometrically valid measures.⁸ In one of the few studies using empirical scales, Watson and colleagues⁹ examined associations between sexual motivations using the YSEX?¹ and the Female Sexual Function Index (FSFI).¹⁰ They found that, compared to women without sexual dysfunction, women with sexual dysfunctions, especially orgasm dysfunction, were more likely to endorse insecurity reasons (including Duty Sex) and less likely to endorse physical reasons like desire and pleasure.⁹ Other work points to a potential link between low sexual desire and Duty Sex, such that women with low sexual desire report a higher frequency of unwanted but consensual sex (ie, sexual compliance).⁶ Muise et al.¹¹ employed a 14-day daily diary study and found that participants experienced lower sexual desire on the days they engaged in Duty Sex.

Duty Sex frequency has also been linked to sexual dissatisfaction.^11–13 Poovey et al.¹⁴ found Duty Sex was associated with lower sexual pleasure, particularly among women who were dissatisfied with their bodies. That is, body dissatisfaction correlated with increased rates of Duty Sex, which subsequently led to distraction during sex (related to appearance, performance, and sexual embarrassment) and, consequently, lower sexual pleasure.¹⁴ Other work suggests that women with sexual pain often engage in Duty Sex.^15,16 Therefore, Duty Sex may be a precursor to sexual dysfunction for some women and a consequence of low sexual function, among others.

Emerging research suggests that a history of NSEs may also relate to a higher frequency of engaging in Duty Sex.^17,18 One study found that most women who reported engaging in Duty Sex also reported past experiences of sexual pressure.¹⁹ It is possible that women with histories of NSEs, particularly childhood NSEs, develop maladaptive schemas and cognitions surrounding sex. Indeed, childhood NSEs have been connected with trauma-related cognitions, such as that one is unlovable or unwanted, as well as feelings of isolation, loneliness, and rejection by partners in adulthood.²⁰ Previous research has demonstrated that disconnection/rejection schemas associated with childhood NSEs predict engagement in Duty Sex.¹⁸ Other research suggests that those who report engaging in unwanted consensual sex, particularly if they had previous NSEs, report less perceived ability to navigate unwanted sexual advances.^21,22 Thus, some women may engage in Duty Sex due to difficulties with or hesitancy to communicate their sexual boundaries, while others may do so to protect themselves from coercion, pressure, or manipulation from their sexual partner.²³

An alternative explanation for the link between a history of NSEs and Duty Sex is that women with a history of NSEs may experience less sexual satisfaction, which, in turn, increases Duty Sex engagement. In a series of studies, Meston and colleagues have shown that individuals with a history of NSEs, particularly childhood NSEs, view sex differently than those without a history of NSEs, such that they experience less sexual and romantic satisfaction.^24,25 Thus, women with a history of NSEs may find sex less satisfying overall, making them less likely to engage in sex for pleasure and more likely to engage in sex out of a sense of duty or obligation. In line with this thinking, Layh et al.¹² found that among college students, greater Duty Sex motives mediated the relation between NSE history and lower sexual satisfaction. Moreover, women with a history of NSEs tend to report lower global sexual function (of which includes sexual satisfaction)^26,27 therefore contributing to the association between having a history of NSEs and Duty Sex frequency.

The present study aims to clarify the relationships between sexual dysfunction, NSEs, and Duty Sex in a large community sample of women using validated psychometric scales. Based on previous literature, we hypothesized that women with histories of NSEs, relative to those without, would report lower sexual function and a higher frequency of Duty Sex. In this study, we examined whether the relation between NSE history and Duty Sex can be explained by distinct facets of sexual function/dysfunction. Analyses were also run using other sexual motives, namely pleasure, intimacy, and self-affirmation, to serve as a proxy control measure to ascertain whether a history of NSEs predict variations across all motivations for engaging in sexual activity, or if there is a unique association specifically with Duty Sex.

Methods

Participants

To be eligible, participants needed to be at least 18 years of age, reside in the United States or Canada, be fluent in English, and identify as a cisgender woman. Participants who did not complete items pertaining to NSE histories (n = 2), sexual motives (n = 6), or those who had not been sexually active in the past month (n = 154) were excluded from analyses, resulting in an analytic sample of 658 participants.

Participants ranged in age from 18 to 68, with a mean age of 34.34 (SD = 10.19) and were primarily heterosexual (n = 522; 79.3%), although a range of sexual identities were reported. Most participants were married (n = 288; 43.8%) or in a committed relationship with one partner (n = 152; 23.1%). Participants who reported at least one experience of nonconsensual oral sex or vaginal or anal penetration were considered to have an NSE history (n= 293; 44.5%); those who did not endorse these experiences comprised the no NSE history group (n = 365; 55.5%). Demographic variables are reported in Table 1.

Table 1.

Demographics.

Continuous variables	M	SD
Age	34.34	10.19
First consensual intercourse	17.54	3.39
First consensual oral sex	18.09	6.41
Discrete variables	N	%
Race/ethnicity
Caucasian/White	501	76.14
African American/Black	61	9.27
Hispanic/Latin American	38	5.78
Asian	37	5.62
Pacific Islander/Hawaiian Native	2	0.30
Native American/Aboriginal/First Nations	8	1.22
Other	11	1.67
Education (highest completed)
Some high school or less	3	0.46
High school graduate/GED	74	11.25
Some college/university	253	38.45
College diploma/university degree	256	38.91
Advanced degree (PhD, MS, MD, JD)	71	10.79
Sexual orientation
Heterosexual	522	79.33
Homosexual/Lesbian/Gay	27	4.10
Bisexual	78	11.85
Pansexual	13	1.98
Queer	2	0.30
No labels preferred	6	0.91
Other	10	1.52
Relationship status
Single (not dating)	45	6.84
Single (casually dating)	76	11.55
Committed relationship (one partner)	152	23.10
Committed relationship (multiple partners)	12	1.82
Living with my partner	84	12.77
Married	288	43.77
Divorced (lifetime)	136	20.67

N = 658. One participant did not report relationship status or education.

Measures

Demographics

An experimenter-derived questionnaire was used to assess various demographic factors, including age, race and ethnicity, education, sexual orientation, sexual debuts (age of first consensual penetrative [vaginal or anal] intercourse and oral sex), and relationship status.

NSEs inventory

The Nonconsensual Sexual Experience Inventory (NSEI)^26,28 is a comprehensive assessment of various NSEs, including vaginal or anal penetration by fingers, objects, genitals, giving or receiving oral sex, and genital or breast fondling or touching. Participants indicated whether they have ever experienced each of the NSE experiences. For each endorsed NSE, participants also indicated the age at the time of the NSE, relationship to the perpetrator, frequency of the experience, and the impact of the experience. The NSEI was used to determine a series of scores to categorize the nature of the NSE experience (ie, lifetime exposure, severity of the NSE behavior, impairment from the NSE, and timing of first NSE). Participants were grouped into a Lifetime Exposure variable, dummy coded as NSE history: (1) or no NSE history (0). The severity variable reflected the most severe sexually aggressive NSE behavior participants experienced: (1) No NSE, (2) sexual touching or fondling, (3) oral sex, or (4) anal or vaginal penetration. Participants also indicated how traumatic each NSE was, ranging from 1 (not at all traumatic) to 5 (extremely traumatic), and the impact the event had on their overall life, ranging from 1 (extremely positive impact) to 7 (extremely negative impact). These two ratings were summed to compute a total impairment score that ranged from 2 (low trauma and impact) to 12 (high trauma and negative impact) for each reported NSE experience. If participants reported multiple NSEs, their highest reported impairment score was used in analyses. The timing of the NSE variable reflects the age of participants during their first NSE.

Sexual function

The Female Sexual Functioning Index (FSFI)¹⁰ is a validated measure used to assess sexual function in females. The scale consists of 19 items and assesses sexual function over the past four weeks in the following six domains: desire (2 items), arousal (4 items), lubrication (4 items), orgasm (3 items), satisfaction (3 items), and pain (3 items). The FSFI also produces a total score as an indicator of overall sexual functioning, with higher scores indicative of better sexual function for both the subscale domains and the total score. The FSFI was scored in line with scoring recommendations,¹⁰ such that subscale scores were computed by summing the domain items and multiplying the sum by the domain factor so that each score ranges from 1 to 5. Total scores were calculated by summing the six domain scores, with scores ranging from 1 to 36. FSFI scores below 26.55 are considered clinical cutoffs and are indicative of sexual dysfunction. The internal consistency in this current sample was high for the total score (α = .93) and the subscales: desire (α = .90), arousal (α = .91), lubrication (α = .83), orgasm (α = .89), satisfaction (α = .89), and pain (α = .93).

Sexual motivations

The Sex Motives Scale (SMS)² assesses the frequency of motivations for engaging in sex with a partner. Participants indicate how often they have sex for each listed reason on a five-point Likert scale ranging from 1 (almost never/never) to 5 (almost always/always). The sexual motives assessed include: enhancement motives (five items; eg, “to satisfy your sexual needs”), intimacy motives (five items; eg, “to become more intimate with your partner”), self-affirmation motives (five items; eg, “to help you feel better about yourself”), and partner approval motives (four items; eg, “worry your partner will not want to be with you if you don’t”). Partner approval motives were used to assess Duty Sex in this current study (hereafter referred to as Duty Sex), given that the items assess a sense of obligation to engage in sex unrelated to intimacy or pleasure. Enhancement motives assess sexual desire or pleasure motives and, therefore, are referred to as Pleasure motives hereafter. Each subscale has shown strong convergent and divergent validity and is intended to be used alone as individual subscales.^2,29 The SMS demonstrated strong internal consistency in this current study: pleasure (α = .86), intimacy (α = .94), self-affirmation (α = .93), and Duty Sex (α = .92).

Procedure

As part of a larger study, participants were recruited using Amazon’s Mechanical Turk (MTurk), an anonymous online crowdsourcing platform, to take part in a 45-minute survey that investigated “consensual and nonconsensual sexual experiences and sexual well-being.” MTurk participants were prohibited from taking the survey more than once and were compensated $1.50 after completion, in line with payment recommendations at the time of the study.

The survey began with a consent form outlining the nature of the survey, potential benefits and risks, and data confidentiality. After providing informed consent, participants completed screener questions in which only participants who were 18 or older, spoke English, and resided in Canada or the United States were eligible to participate. Participants who met eligibility criteria in the screener were then shown the study measures as well as other measures reported elsewhere.^26,28 At the end of the survey, participants were provided with a debrief form that provided contacts if they had questions, additional explanations explaining the purpose of the study, and sexual violence support resources. All procedures were approved in full by the authors’ institutional ethics review board prior to the implementation of the study.

After data collection was complete, researchers reviewed responses to ensure they met data integrity. Participants were only included in analyses if they passed 50% (4 of 8) of attention check items (eg, “please select strongly disagree”), answered at least half of the survey items, and answered written response items with adequate detail to ensure data integrity. Researchers reviewed open-ended responses to ensure they were coherent and naturally phrased to detect non-human (ie, bot) responses.³⁰ A total of 29.1% (N = 335) of participant responses were excluded from analyses as they did not meet the data integrity guidelines. We then removed participants who were not sexually active in the past 4 weeks and those who did not answer any of the SMS items. A small percentage of participants had missing items on the SMS, which appeared to be random. If participants missed an item on the SMS, we were unable to compute a sum score for that motivation the item assessed. Thus, 5.6% (n = 37) of participants were missing one of four sex motives, and 0.3% (n = 2) were missing two motives. Specifically, 0.6% (n = 4) of participants were missing a Duty Sex score, 2.0% (n = 13) a self-affirmation score, 1.7% (n = 11) a pleasure score, and 2.0% (n = 13) an intimacy score.

Statistical analyses

The aim of this study was to examine differences in sexual motives, particularly Duty Sex, among women with and without histories of NSEs, and to examine whether NSE history predicts sexual motives even after accounting for the variance in sexual function. First, we ran a series of analyses to investigate whether women with histories of NSEs, relative to those without, differed in their frequencies of each sexual motive. Next, we examined how each of the six domains of sexual function predicted each of the four sexual motivations. This analysis aimed to examine how each domain contributes to sexual motivations and whether these contributions differ across motivations, as suggested by previous research.⁸ Then we examined which, if any, motives were predicted by NSE history. If NSE histories significantly predicted any of the sexual motives, we conducted follow-up analyses to determine: (1) whether NSEs still predicted that specific motivation after accounting for sexual function and (2) whether the characteristics of the NSE (age, type, and impairment of the experience) influenced Duty Sex frequency.

Data was analyzed using R version 4.3.1.³¹ First, we grouped NSE history into a binary variable such that participants who reported a history of nonconsensual sexual intercourse (oral, anal, or vaginal sex) were considered to have NSEs and were coded as one, while those without NSE histories were coded as zero.

Prior to conducting study analyses, the data was examined for outliers, normal distribution, and equal variance for items measuring sexual motives and sexual function. No studentized residuals had Bonferroni-adjusted P-values <.05, and therefore, there were no outliers in the data. Durbin–Watson Tests yielded non-significant results, and thus, the data met the independence assumption. However, the data was particularly skewed and was not normally distributed, as indicated by significant Shapiro–Wilk normality tests. In addition, the duty, pleasure, and self-affirmation subscales of the SMS and orgasm, desire, and lubrication scores on the FSFI had significant (P < .05) Levene’s test results, indicating that they violated the homogeneity of variance assumption.

Given that the data violated assumptions of normality and equality of variance, non-parametric methods were used to assess group differences in sexual motives and sexual function for women with and without NSEs. Mann–Whitney U tests were used to compare the distribution of samples of two independent groups (ie, women with and women without NSEs) to determine if the sample has the same median and distribution.

Data was analyzed using cumulative link models (CLM) using the Ordinal Regression Package in R³² to examine the relationship between sexual function, NSE history, and sexual motives. Sexual motives were considered the dependent variable, with sexual function and NSE history as the independent variables. The scores for each subscale of the SMS were summed to create a total score for that motive. These motive scores were treated as categorical variables with discrete levels of meaningful order. The Duty Sex subscale consisted of four items with scores ranging from 4 to 20, and higher values corresponded to higher levels of Duty Sex engagement. The intimacy, pleasure, and self-affirmation motives consisted of five items each, and therefore ranged from 5 to 25. Considering the predictor variables, NSE history was treated as a categorical binary variable that was dummy coded such that the reference group comprised of women without NSEs. Domains of sexual function (FSFI total scores) were treated as continuous variables and entered into the model as covariates. Model fit was assessed using the Akaike Information Criterion (AIC) derived from the CLM package. Lower AIC values reflect better model fit. In addition to the significant testing using P values, results are considered significant if the confidence intervals of the unstandardized regression coefficient do not cross zero.

Follow-up analyses were also conducted to examine how characteristics of the NSE influence sexual motivations. First, we examined whether the frequency of the sexual motive differed depending on the most severe NSE behavior participants experienced. Group differences were compared using Kruskal–Wallis tests, a non-parametric alternative to a one-way ANOVA, to examine differences among multiple independent groups when the dependent variable violates assumptions of normality. Post-hoc comparisons were performed using the Bonferroni correction for multiple comparisons. In addition, we conducted two ordinal logistic regression analyses using the same process described in the aforementioned paragraph to examine how the perceived impairment of the NSE and the timing (age) of the first NSE predicted sexual motivations.

It is important to note that we treated sexual function as a predictor of the sexual motive so we might evaluate how each domain individually contributes to each sexual motivation. However, sexual function may also be an outcome of the sexual motive. For example, sexual satisfaction is an outcome of the motivation if someone feels less sexually satisfied after having Duty Sex, but a predictor of the motivation if someone has low sexual desire and has sex only because they feel a sense of duty. We treated sexual function as a predictor on the premise that women reference their sexual function in making decisions about having sex more than they evaluate their sexual function based on their sexual motivations. This is especially true in evaluating lubrication, arousal, pain, and orgasm domains of one’s sexual function, which require reflection on the sexual encounter itself as opposed to their motivations prior. In other words, someone might consider the quality of their last orgasm and how aroused they felt when deciding if and why they want to have sex again.

Results

Descriptive statistics

See Table 1 for all descriptive statistics and Table 2 for characteristics of the NSEs. Ordinal logistic regression equations were used to determine whether there were differences in Duty Sex frequency and various demographic variables. Duty Sex frequency did not significantly differ by participant age, age of first sexual intercourse nor oral sex, relationship status, age of first NSE, or history of divorce (all P > .2).

Table 2.

Characteristics of participant NSEs.

Continuous variables	M	SD	Range
Age of first NSE	15.27	7.79	1-50
Highest impairment score across all NSEs	8.93	2.29	2-12
Discrete variables	N	%
Type of NSE behavior *
Sexual fondling	272	41.34
Oral sex	113	17.17
Vaginal sex	244	37.08
Anal sex	94	14.29
Most severe NSE behavior reported
Sexual fondling	91	23.45
Oral sex	20	5.15
Penetration (vaginal or anal sex)	277	71.39
Number of NSE behaviors reported
Sexual fondling only	91	23.45
One type of NSE (excluding fondling)	79	20.36
Two types of NSEs	126	32.47
Three types of NSEs	67	17.27
All four (fondling, oral, vaginal, anal)	25	6.44
Age of first NSE **
Less than 12	118	30.41
Between 12-17	115	29.64
18 or older	145	37.37
Timing of first NSE relative to consensual sex
Before first consensual experience	176	45.36
After first consensual experience	201	51.80

Type of NSE behavior frequencies were divided by total sample size (N = 658) and remaining categories divided by the sample that endorsed some type of NSE (n = 388).

^*Type of NSE behaviors frequencies are greater than the sample size as participants could endorse multiple types of NSEs.

^**Some participants did not provide age of first NSE.

Group differences in sexual motives and sexual function

A series of asymptotic Wilcoxon–Mann–Whitney Tests were used to examine whether there were differences between women with and without NSE histories on measures of sexual motives and sexual function. Women with NSE histories (MDN = 8) compared to those without (MDN = 6), reported a higher frequency of Duty Sex, z = −3.35, P < .001, r = 0.13. There were no other significant differences in sexual motives (ie, pleasure, intimacy, and self-affirmation) between women with and without NSE histories (see Table 3).

Table 3.

Sexual motives and sexual function by NSE history.

	NSE history (N = 293)		No NSE history (N = 365)		z
	M	SD	M	SD
Duty sex	9.12	5.28	7.57	4.30	−3.35**
Pleasure	18.19	4.67	17.93	4.21	−1.03
Intimacy	20.30	5.00	20.45	4.42	−.11
Self-affirmation	12.06	6.00	11.69	5.40	−.48
FSFI total	26.76	6.08	27.96	5.42	−2.39*
Orgasm	4.33	1.54	4.65	1.37	−2.53*
Desire	3.72	1.34	3.80	1.17	−.40
Arousal	4.48	1.38	4.61	1.23	−.71
Pain	4.93	1.49	5.00	1.50	−.87
Lubrication	4.78	1.29	5.05	1.01	−2.12*

^*P < .05; ^**P < .001. Higher sexual motive scores indicate more frequent sex for that motive. Higher FSFI scores indicate better sexual function.

Asymptotic Wilcoxon–Mann–Whitney Tests were also used to examine differences in sexual function between women with and without histories of NSEs. Women with NSE histories (MDN = 27.5) reported lower global sexual function scores than women without NSE histories (MDN = 29.2), although the effect was small z = −2.39, P = .02, r = 0.093. Comparing women with and without histories of NSEs revealed that, despite both groups having a median orgasm score of 4.8, the distribution of data differed such that women with histories of NSEs had lower orgasm function scores, z = −2.53, P = .01, r = 0.099. Women with histories of NSEs also had lower vaginal lubrication scores (MDN = 5.1) compared to those without NSE histories (MDN = 5.4), z = −2.12, P = .03, r = 0.083. There were no significant group differences in measures of desire, arousal, or sexual pain (see Table 3).

Sexual function and sexual motives

Next, we conducted a series of ordinal logistic regressions to examine whether sexual function predicted sexual motives. To evaluate model fit, we compared the performance of two models in the prediction of each sexual motive: one with the FSFI domains as predictors and one without any predictors (ie, the null model). We then examined whether there was a significant difference in the model AIC values. All models demonstrated significant improvement relative to the null model, such that including all six FSFI domains significantly predicted each of the four sexual motives (see Table 4). Experiencing more sexual pain (B = −0.26, SE = 0.055, P < .001) and lower sexual satisfaction (B = −0.25, SE = 0.075, P = .001) significantly predicted higher Duty Sex frequency. Notably, including all six FSFI domains as predictors provided a significantly better model fit (AIC = 2916.9) than the null model (AIC_null = 3060.7; X² = 155.82, df = 6, P < .001), suggesting that each of the domains of sexual function are important to Duty Sex frequency.

Table 4.

Ordinal regression analyses of sexual function to predict sexual motives.

Model	B	SE	z	P	CI	Model fit (X 2 )
Duty sex
Arousal	−0.16	0.097	−1.69	.091	[-0.35, 0.026]	155.82**
Lubrication	−0.16	0.088	−1.81	.070	[-0.33, 0.013]
Desire	−0.003	0.077	−.037	.97	[-0.15, 0.15]
Orgasm	−0.10	0.066	−1.53	.13	[-0.23, 0.028]
Pain	−0.26**	0.055	−4.84	<.001	[-0.37, -0.16]
Satisfaction	−0.25*	0.075	−3.36	<.001	[-0.40, -0.11]
Pleasure
Arousal	0.25*	0.099	2.56	.010	[0.060, 0.45]	277.74**
Lubrication	0.19*	0.087	2.14	.032	[0.016, 0.36]
Desire	0.65**	0.080	8.07	<.001	[0.49, 0.80]
Orgasm	0.13*	0.063	2.08	.037	[0.0076, 0.26]
Pain	0.13**	0.053	2.54	.0011	[0.031, 0.24]
Satisfaction	−0.023	0.073	−.32	.75	[-0.17, 0.12]
Intimacy
Arousal	0.069	0.097	.71	.48	[-0.12, 0.26]	86.79**
Lubrication	−0.086	0.089	−.97	.33	[-0.26, 0.088]
Desire	−0.080	0.075	−1.07	.29	[-0.23, 0.067]
Orgasm	−0.018	0.066	−.28	.78	[-0.15, 0.11]
Pain	0.16**	0.056	2.83	.0046	[0.049, 0.27]
Satisfaction	0.47**	0.074	6.37	<.001	[0.33, 0.62]
Self-Affirmation
Arousal	−0.22*	0.097	−2.31	.02	[-0.42, -0.034]	67.45**
Lubrication	−0.067	0.087	−.76	.44	[-0.24, 0.10]
Desire	0.48**	0.077	6.27	<.001	[0.33, 0.64]
Orgasm	−0.039	0.065	−.61	.55	[-0.17, 0.088]
Pain	−0.19**	0.054	−3.45	<.001	[-0.29, -0.081]
Satisfaction	−0.081	0.073	−1.10	.27	[-0.22, 0.064]

Model fit determined by comparing the model with predictors to the null model containing no predictors. df = 6. CI = confidence intervals of the estimate (2.5%-97.5%). Lower scores on the FSFI domains indicate lower sexual function.

^* P < .05.

^** P < .01.

The model examining pleasure motives (AIC = 3356.1) also provided a significantly better model fit than the null model (AIC_null = 3621.8; X² = 277.74, df = 6, P < .001), indicating that greater arousal, lubrication, desire, and orgasm scores, as well as less issues with sexual pain, predicted more frequent pleasure motives. The six FSFI domains also provided a significantly better model fit (AIC = 2941.8) in the prediction of intimacy motives relative to the null model (AIC_null = 3016.6; X² = 86.79, df = 6, P < .001), such that less sexual pain and greater satisfaction predicted greater intimacy motives. FSFI domain scores also significantly predicted self-affirmation motives (AIC = 3530.6) in comparison to the null model (AIC_null = 3586.0; X² = 67.45, df = 6, P < .001), indicating that lower arousal scores, scores indicating more issues with sexual pain, and higher desire scores predicted more frequent self-affirmation motives. All estimates and significance levels are reported in Table 4.

In sum, sexual function is an important predictor of all sexual motives. Duty Sex, in particular, occurs more frequently among women who report lower sexual satisfaction and more problems with sexual pain.

History of NSEs and sexual motives

We examined a series of ordinal logistic regressions using the CLM functions to determine whether having a history of NSEs predicted sexual motives. Considering Duty Sex as an outcome, the final AIC model (AIC = 3053.3) was a significantly better fit than the null model without any predictors (AIC_null = 3062.7; X² = 11.39, df = 1, P < .001), indicating that those with NSEs tend to report a higher frequency of Duty Sex (B = 0.48, SE = 0.14, P < 0.001). NSE history did not significantly predict physical, intimacy, or self-affirmation motives (see Table 5).

Table 5.

Ordinal regression analysis of NSE history to predict sexual motives.

Model	B	SE	z	P	CI	Model Fit (X 2 )
Duty sex
NSEs	0.48*	0.14	3.37	<.001	[0.20, 0.75]	11.39*
Pleasure
NSEs	0.14	0.14	1.04	.30	[-0.13, 0.41]	1.08
Intimacy
NSEs	0.016	0.14	0.11	.91	[-0.26, 0.29]	0.013
Self-affirmation
NSEs	0.067	0.14	0.49	.63	[-0.20, 0.34]	0.24

Model fit determined by comparing the model with predictors to the null model containing no predictors. df = 1. CI = confidence intervals of the estimate (2.5%-97.5%).

^* p < .01.

History of NSEs, sexual function, and duty sex

We examined whether NSE history still predicted sexual motives after accounting for the variance in sexual dysfunction. The model comprised of both NSE history and each of the FSFI domains (AIC = 2913.1) was a significantly better fit than the model consisting of only FSFI domain scores (AIC_null = 2916.9; X² = 5.83, df = 1, P = .02), indicating that NSE history adds to the predictive utility of the model over and above that of sexual functioning alone. There was a significant association such that NSE history predicted Duty Sex frequency (B = 0.35, SE = 0.15, P = .02), as did sexual pain (B = −0.27, SE = 0.054, P = <.001) and sexual satisfaction (B = −0.24, SE = 0.075, P = .001). Thus, women with NSE histories were more likely to report more frequent Duty Sex motives, even after accounting for variability in sexual function (Table 6). In terms of model fit, the model including all six FSFI domains as well as NSE history provided the best model fit (as opposed to excluding non-significant FSFI domains). This suggests that each of the six FSFI domains contribute, at least to some extent, to the frequency one engages in Duty Sex, even if not to a level that reaches statistical significance.

Table 6.

Ordinal regression analysis of NSE history and sexual function to predict duty sex.

Model	B	SE	z	P	CI	Model Fit (X 2 )
Duty Sex
History of NSEs	0.35*	0.15	2.42	.016	[0.066, 0.64]	5.83*
Arousal	−0.18	0.097	−1.90	.057	[-0.38, 0.0057]
Lubrication	−0.14	0.089	−1.61	.11	[-0.32, 0.031]
Desire	−0.0011	0.077	0.014	.99	[-0.15, 0.15]
Orgasm	−0.088	0.066	−1.33	.18	[-0.22, 0.042]
Pain	−0.27**	0.054	−4.94	<.001	[-0.38, -0.16]
Satisfaction	−0.24**	0.075	−3.19	.0014	[-0.39, -0.093]

Model fit determined by comparing the model with predictors to the null model that includes only FSFI domains and excludes NSE histories as predictors. df = 1. CI = confidence intervals of the estimate (2.5%–97.5%). Lower scores on the FSFI domains indicate lower sexual function.

^* P < .05.

^** P < .01.

Characteristics of the NSE and duty sex

Follow-up analyses were conducted to examine how the characteristics of the NSE influenced Duty Sex frequency. First, we used a Kruskal–Wallis H test to determine if there were statistically significant differences between the medians of Duty Sex frequency among each of the four independent groups. Participants were grouped based on the most severe type of assault behavior they reported—no NSEs, fondling, oral sex, or vaginal/anal penetration (Table 2). The results indicated a statistically significant difference in the median scores between the groups: X²(3) = 15.83, df = 3, P = 0.001. Post-hoc pairwise comparisons using Bonferroni corrections revealed that the penetration group (MDN = 8) reported a significantly higher frequency of Duty Sex than the group without NSE histories (MDN = 5), z = −3.93, P < .001. There were no other significant differences between groups with regard to the fondle only group (MDN = 7) and the oral sex group (MDN = 4). These results seem to suggest that vaginal or anal penetration are the NSE behaviors most likely to increase Duty Sex frequency, whereas sexual fondling NSEs do not seem to increase Duty Sex frequency. It is important to note here that the oral sex NSE group was relatively underpowered, which may account for the lack of significant differences in this group.

Next, we utilized ordinal regression analyses to examine whether the impairment (comprising trauma and impact) that individuals experienced after the NSE predicted their Duty Sex frequency. The final model (AIC = 1853.8) was not a significantly better fit than the null model (AIC_null = 1851.8; X² = 0.011, df = 1, P = .92), indicating that perceived impairment does not predict Duty Sex frequency. The timing (age) of the first NSE also did not significantly predict Duty Sex frequency (AIC = 1842.3) better than the null model without predictors (AIC_null = 1840.3; X² = 0.0015, df = 1, P = .97).

Taken together, these results suggest that neither the timing nor the individual’s perceived trauma/impact of the NSE explain the variance in Duty Sex frequency among women. In contrast, the type of behavior may better predict the frequency of Duty Sex, given that women who experienced penetrative assault were the only group that differed from women without NSE experiences.

Discussion

To our knowledge, this study was the first to examine the potential link between NSEs, each of the domains of sexual function, and Duty Sex. Our findings revealed that having a history of NSEs positively predicted the frequency of engaging in Duty Sex, and this relationship could not be accounted for by variations in sexual function. Also noteworthy, having a history of NSEs was unrelated to other categories of sexual motivation studied (ie, pleasure, intimacy, self-affirmation), suggesting a unique relationship between having a history of nonconsensual sex and frequency of engaging in sex out of a sense of obligation or duty.

Consistent with past findings,^6,9,11,16 women with greater levels of sexual dysfunction reported engaging in more frequent Duty Sex. Specifically, there was a significant relationship such that women who reported more Duty Sex frequency tended to report more problems with sexual pain. Qualitative interviews suggest that women with sexual pain may engage in Duty Sex because of their belief, or their partner’s assertion, that their male partner needs sex and/or because they perceive sexual activity as crucial to their gender identity as a woman.¹⁵ Brauer et al.¹⁶ reported that women who experience painful intercourse often endure intercourse despite discomfort, especially if their male partners react negatively to their pain. As such, women with sexual pain appear to engage in Duty Sex to accommodate their partner’s sexual needs, whether perceived or explicitly expressed.

Also consistent with past findings, sexual dissatisfaction emerged as a significant predictor of more frequent Duty Sex.^33,34 Women experiencing low sexual satisfaction may engage in Duty Sex to fulfill their partner’s desire for sex. However, emerging research seems to suggest that low satisfaction is a consequence of Duty Sex. For example, women who experienced body dissatisfaction¹⁴ or who derived their self-worth from their ability to maintain relationships¹³ tended to engage in more Duty Sex, which in turn led to lower sexual satisfaction. Sanchez et al.¹³ demonstrated, through statistical analyses, that this relationship was strongest when sexual satisfaction was considered an outcome as opposed to a predictor of Duty Sex. Muise et al.¹¹ elucidated the potential direction of this association by measuring participants’ event-specific sexual motivations and sexual satisfaction each day after couples had sex. Participants reported lower sexual satisfaction when they reported that their Duty Sex motives were greater that day than compared to their average motivation to engage in Duty Sex across the 14-day study.¹¹ Taken together, these results suggest that low satisfaction is an outcome of Duty Sex, possibly because participants were not seeking sexual satisfaction in the first place. In addition, other factors that tend to lead to less satisfying sex, such as sexual pain, may simultaneously increase the frequency of Duty Sex and contribute to lower sexual satisfaction.

In contrast to the strong relation between sexual pain and Duty Sex, difficulties with orgasm, desire, arousal, and lubrication did not significantly contribute to Duty Sex frequency. Watson et al.⁹ used the same validated measure of sexual desire that we used in the present study (ie, FSFI Desire domain) and also reported that sexual desire levels did not predict a tendency to engage in Duty Sex. Results from studies that have used measures of sexual desire different from the current study have yielded conflicting results, with some studies finding a positive relation between Duty Sex and low desire¹¹ and others finding no relation.³⁴

Unlike the findings from the current study, Watson et al.⁹ found that women with orgasm problems were more likely to endorse engaging in Duty Sex. Although Watson et al.⁹ also measured orgasm function using the FSFI, differences in the categorization of orgasm function between Watson et al.’s⁹ study and the current study may explain the differences in findings. Watson et al.⁹ grouped participants into two groups, such that those who reported “always” or “almost always” to each of the items were in the orgasm difficulty group, which comprised 15% of the total sample.⁹ In contrast, we maintained continuous scores for each FSFI domain to allow for more significant variability in our data and to capture nuances in female orgasm function. Watson et al.⁹ also examined the entire Insecurity subfactor comprising duty/pressure, self-esteem boost, and mate guarding,¹ while we exclusively examined Duty Sex. While our results indicated that orgasm functioning scores were not associated with Duty Sex, other motives within the Insecurity factor may be more likely to motivate sex among women with orgasm difficulties.

It is also important to note that although only the FSFI domains of pain and satisfaction significantly contributed to Duty Sex frequency, model fit analyses suggested that the model with all six FSFI domains was the best predictor of Duty Sex. As such, it can be said that each of the six domains of sexual function contributed to Duty Sex frequency, at least to some extent. Indeed, the current research adds to the literature suggesting sexual function across multiple domains is a critical predictor of women’s frequency of engaging in sex for multiple motivations and that poorer sexual functioning is likely reciprocally associated with engaging in sex for motivations other than pleasure (eg, out of duty and obligation motives).

The second aim of the study was to examine how NSE history influences sexual motives. Notably, NSE histories were only associated with Duty Sex frequency and did not predict the frequency of self-affirmation, intimacy, or pleasure, indicating a unique relationship between NSE histories and engaging in sex for duty and obligation motivations. Women often report engaging in Duty Sex in response to their partner’s discontent with sexual (in)frequency.^6,35 Therefore, women with NSE histories may find it particularly difficult to decline sex out of fear of disappointing their partner or a reluctance to deny partner advances.^21,22 Women with a history of childhood NSEs are more likely to develop a fear of rejection or isolation, referred to as disconnection/rejection schemas.^20,36 Such schemas, in turn, predict engagement in unwanted but consensual sex in adulthood.¹⁸ Women with NSE histories may have an increased risk of abusive/coercive relationships in which they engage in Duty Sex to avoid escalated coercion or pressure to engage in sex.^23,37 Other studies have found that sexual autonomy, that is, an individual’s capacity to communicate sexual preferences and make decisions during sexual interactions, is lower among those who engage in Duty Sex.¹³ Although speculative, the violation of autonomy during a NSE may have lingering effects that impact women’s perceived ability to make and communicate sexual decisions, leading them to engage in sex out of a sense of duty that is not physically or psychologically desired.

Given that past research has shown that NSE histories are associated with sexual dysfunction,^26,27 we investigated whether sexual dysfunction accounts for the relationship between NSEs and Duty Sex. Our findings revealed that Duty Sex remained higher among women with NSE histories after controlling for the variance in sexual dysfunction. Thus, there is something unique to having NSE histories that is associated with increased Duty Sex frequency beyond that attributable to sexual dysfunction. Some work has suggested that Duty Sex mediates the relationship between NSE history and sexual dissatisfaction,^12,37 although the mechanism that explains the association remains unclear. Perhaps there is a dual pathway: (1) NSEs contribute to a higher likelihood of sexual dysfunction^26,27 that decreases the desire to engage in sex and in turn increases the frequency of Duty Sex, and (2) NSEs lead to maladaptive beliefs about sex and their sexual self^20,26 and one’s ability to communicate their sexual boundaries^21,22 that then make it more challenging to deny a partner’s advances. NSEs could therefore impair the sexual experience itself by impacting sexual function, as well as the communication of consent, desire, and sexual preferences before and during sex, which cumulatively may increase the likelihood of engaging in Duty Sex. Our findings may suggest that it is penetrative NSEs, as opposed to other types of NSE behaviors, that drive increases in Duty Sex, which may be explained by both increases in sexual dysfunction and maladaptive beliefs about sex and consent that develop after penetrative NSEs.

It is notable that the developmental timing of the NSE did not predict Duty Sex frequency. There is a considerable amount of research that suggests that NSEs that occur in childhood, particularly before the age of consent, tend to cause more significant impairment on well-being, sexual schemas, and daily functioning than those that occur in adulthood.^24,27 Yet, age did not significantly influence Duty Sex frequency. It is possible that regardless of age, individuals who experience NSEs may also adopt beliefs that they cannot refuse sex. Perhaps women with penetrative NSEs internalize beliefs that it is safer to engage in sex out of a sense of obligation than to refuse sex and risk experiencing another NSE. Regardless of age, NSEs can cause future sexual pain,^27,38 which in turn might increase Duty Sex frequency and therefore explain the lack of age-related differences. Still, more research is needed to replicate these findings and to better explain why the timing of the first NSE does not impact Duty Sex frequency.

There are limitations with the current study that warrant mention. In terms of sexual motivations, we did not measure how frequently individuals have sex nor how often they believe they have Duty Sex. Instead, we broadly measured how frequently participants have sex for each of the motivations on a Likert scale ranging from never to always. These retrospective reflections on their motives, as opposed to event-specific motivations, may not capture the nuances of sexual motivations. For example, women can be motivated by multiple motives at once—they may feel a sense of duty but also want to feel closer to their partner. In addition, some participants may not define their encounters as Duty Sex even if they endorse items on the SMS that are reflective of Duty Sex. It remains unclear what percentage of our participants’ sexual encounters are exclusively or primarily motivated by duty or obligation and if they would choose to identify their encounters as Duty Sex.

We also did not capture how Duty Sex engagement directly influences the participant’s sexual satisfaction and overall well-being. One future area of research to account for this discrepancy is to provide a definition of Duty Sex to participants and then measure the percentage of sexual encounters one has that they consider to be Duty Sex. We suspect that women with NSEs who experience sexual dysfunction would report a higher percentage of sexual encounters motivated by a sense of duty or obligation. Structured interviews may also provide insight into why individuals feel a sense of duty or obligation, how they feel after Duty Sex, and their perceptions of the long-term personal and interpersonal consequences of engaging in Duty Sex.

The second limitation of the study is the low variance accounted for by the independent variables in the prediction of Duty Sex. Investigating the percentage of sexual experiences that constitute Duty Sex may remedy our second limitation by improving the specificity of our measurement. Moreover, a minority of women in our sample (40%) met the clinical cutoff for sexual dysfunction. Including a larger sample of women who experience clinically diagnostic levels of sexual dysfunction in future research would increase statistical power and possibly result in more variability accounted for in our models. In addition, there are likely other factors, such as age, relationship length, sexual experience, partner performance, religious beliefs, self-esteem, and sociosexuality, that contribute to one’s frequency of Duty Sex that warrant future investigation.

We also did not have enough variability in participants’ sexual orientation and gender identity to provide adequate power to conduct subgroup analyses. However, Duty Sex is not limited to cis-gender heterosexual women. In one of the few studies examining sexual orientation, Rosen et al.¹⁰ found that lesbian women report engaging in Duty Sex with both their current female partners and previous male partners. Men also report engaging in Duty Sex, although at a lower rate than women.¹⁹ There remains a number of questions that need to be addressed in future research; how does one’s gender and sexual orientation shape their experiences with Duty Sex, how might Duty Sex maintain or challenge traditional gender roles and the heteronormative script within relationships, and does Duty Sex manifest differently among non-binary folks and transgender men and women?

Another limitation of our study is that we did not account for whether our participants have sought out therapy or treatment for their NSEs or sexual dysfunction. Research demonstrates that treatment seeking for NSEs, particularly those that occurred in childhood, can improve sexual function and other adverse outcomes, such as anxiety, guilt, and post-traumatic stress disorder.^39,40 Therefore, women who have sought treatment after their NSEs may be less likely to engage in Duty Sex by nature of increasing their sexual function or adopting more positive views about sex.

Conclusions

It is important to note that we are unable to say for certain whether Duty Sex is inherently problematic, as the extent to which Duty Sex causes distress and greater impairment is still unclear. For some women, what begins as a dutiful response to a partner’s sexual advances might end in genuine enjoyment, pleasure, and intimacy.^34,35,41 Still, our findings highlight how women with NSE histories and lower sexual function, both of which may be distressing experiences, are more likely to engage in Duty Sex. We also found that sexual satisfaction is lower among those who engage in Duty Sex. Therefore, if future research points to additional deleterious consequences of Duty Sex, such as relationship conflict and distress, our work highlights which groups are most at risk and may benefit from efforts to enhance sexual function and satisfaction, as well as more pleasure and self-determined motivations for sex.

Author contributions

K.M.: Conceptualization-lead, Formal analysis-lead, Writing—original draft-lead. C.K.: Conceptualization-supporting, Data curation-lead, Investigation-lead, Methodology-lead, Writing—original draft-supporting. C.M.: Conceptualization-supporting, Data curation-supporting, Investigation-supporting, Methodology-supporting, Supervision-lead, Writing—original draft-supporting.

Funding

This work was supported in part by the Canadian Institutes of Health Research (Funding Reference Number: 152269, awarded to Chelsea Kilimnik in May 2017).

Conflicts of interest

We have no known conflicts of interest to disclose.

Disclosure

Dr. Cindy Meston is a Pharmaceutical Advisory Board Member.