Introduction

Clinical eating disorders and subclinical disordered eating behaviors, which can include binge eating, compulsory eating, purging, restrained eating, and excessive dieting, are a common public health concern as well as a health disparity for LGBTQ + populations (Parker & Harriger, 2020). Prevalence estimates of clinical eating disorders are as high as 9% in the U.S. with marginalized groups uniquely affected (Deloitte Access Economics, 2020). Not only are sexual minority men more likely to experience health concerns when compared with heterosexual men (Operario et al., 2015), but sexual minority men, specifically, are at an increased risk of eating disorder symptoms (Brown & Keel, 2015). While some research suggests that being in a relationship can be a protective factor among bisexual and gay men, other literature suggests that dynamics within the relationship, including dietary-related interactions and lower relationship quality, can be harmful to sexual minority men and individuals (Brown & Keel, 2015; Novak et al., 2021). Better understanding these relationship factors in connection with the eating behaviors of sexual minority individuals could inform practice to better support this population and their partners; the use of a dyadic sample gives insight into the context of relationship factors rather than simply relationship status itself.

Intuitive eating, a non-restraining approach to eating based on listening to one’s body cues, is increasingly shown to reduce disordered eating behaviors (Hazzard et al., 2021), but has yet to be investigated in the context of a romantic relationship. Higher scores of intuitive eating are associated with improved food intake/diet quality and several other health indicators (Van Dyke & Drinkwater, 2014). A meta-analysis of interventions designed to increase intuitive eating shows associations between these interventions and positive health changes on outcomes like body appreciation and body image (Babbott et al., 2022). Among sexual minority adults, intuitive eating has been associated with greater positive body image (Souillard & Vanderwall, 2019), but studies of intuitive eating among sexual minority populations are rare (Babbott et al., 2022) and nonexistent among dyads of partners. The present study sought to psychometrically explore the Intuitive Eating Scale-2 (IES-2) and examine the dyadic associations among partner’s mental health, relational quality, and interpersonal eating interactions in a sample of 228 married male sexual minority couples.

Literature Review

Sexual Minority Individuals, Eating Patterns, and Eating Beliefs and Practices

As mentioned, sexual minority men experience eating disorders and disordered eating behavior at greater rates than heterosexual men (Feldman & Meyer, 2007). This pattern is seen with bulimic symptoms, frequency of dieting, and drive for thinness (Brown & Keel, 2012). Among sexual minority men, processes of objectification (Fredrickson & Roberts, 1997) such as body surveillance, internalization of appearance ideals, and appearance-based social comparisons, are evidenced as risk factors for clinical eating disorders, while negative body image, sociocultural norms and attitudes about weight, and external motivation to work out/pressure to diet have been considered risk factors for disordered eating behavior (Parker & Harringer, 2020). These processes of objectification may be a part of the connection between eating behavior and body image, prompting further investigation into eating attitudes, beliefs, and practices of sexual minority men and individuals (the present sample includes male individuals in marital relationships with male individuals). In addition, disparities in health (such as disordered eating) among sexual minority populations are explained by systemic issues of minority stress and stigma (Hatzenbuehler et al., 2013; Meyer, 2003). When it comes to intervention efforts, there is one important construct closely related to disordered eating that has been highlighted in the literature as modifiable and sensitive to intervention, which is the construct of intuitive eating.

Intuitive Eating

Intuitive eating is the practice of eating with respect to one’s internal hunger and satiety cues while focusing on pleasure and wellness and not restricting specific foods (Tylka & Kroon Van Diest, 2013). Higher scores of intuitive eating are associated with increased positive psychological and better mental health outcomes, specifically concerning disordered eating, body image, self-esteem, and depressive symptomology (Hazzard et al., 2021; Linardon et al., 2021). In terms of physical health, participants of health promotion and mindfulness-based interventions that incorporated intuitive eating principles showed increased moderate physical activity and decreased rates of low-density lipoprotein cholesterol levels and systolic blood-pressure at 2-year follow up as well as decreased c-reactive protein–a biomarker of inflammation and cardiovascular risk (Bacon et al., 2005; Dalen et al., 2010). These improvements in physical health indicators could result from the acknowledged relationship between intuitive eating and diet quality (Tabatabai et al., 2021).

However, these studies have not validated intuitive eating among sexual minority individuals, nor have they examined intuitive eating within the context of romantic relationships, with past work highlighting the interdependence between eating behaviors and relational processes, especially among gay partnered men (August et al., 2022). Validation of the IES-2 among samples including racially, ethnically, or culturally diverse individuals and individuals with low income has shown differences from the original scale in factor structure (Khalsa et al., 2019; Saunders et al., 2018; Swami et al., 2020); the sociocultural aspects of being a part of a sexual minority couple may also affect the factor structure of this scale and thereby measurement variance and bias. Studies of intuitive eating are common among women (Linardon et al., 2021); however, intuitive eating levels are high among men (Gast et al., 2012), making this demographic cross-section useful for understanding gendered implications and for targeted intervention.

Eating Behaviors and Health Influence Among Partners

Importantly, relationship status may be a protective factor among sexual minority individuals, as research comparing single and partnered cisgender gay men found that being single predicted greater appearance intolerance (Nagata et al., 2022) and greater drive for thinness (Brown & Keel, 2013). The quality of the relationship, however, paints a more nuanced picture among partnered sexual minority men, with some research suggesting greater relationship satisfaction is protective against bulimic symptoms, but lower relationship satisfaction is a risk factor for drive for thinness (Brown & Keel, 2012, 2015). Thus, being in a relationship, whether one is a sexual minority or not, may not be only a source of strength in face of body image and eating challenges but could exacerbate these troubles, depending on the ways partners influence each other’s eating behaviors.

Eating behaviors (and eating regulation strategies) may be one of the more significant manifestations of dyadic processes of health influence, as couples can experience dietary convergence—in which the eating habits and regulation strategies used by their partner affect their own eating habits—for good or ill (Bove et al., 2003). A burgeoning literature has found that significant others often seek to influence their partners through health behavior modification, a phenomenon of both social support and control (Umberson et al., 2018). Attempts to control one’s partner, as opposed to support, have been shown to result in feelings of guilt or compromised mental wellness (August et al., 2016). In addition, partners can influence each other’s eating behaviors in more indirect ways beyond support and control. For example, Novak et al. (2021) found that gay married men may participate in mutual dietary undermining (e.g., eating foods in front of the other that the other is trying not to eat), which was associated with more conflict, higher depressive symptoms, and worse diet quality for both partners. Related to intuitive eating, past work has found that the interpersonal style of partner control influences controlled eating regulation directly and intuitive eating indirectly (Carbonneau et al., 2015). And yet, the concept of intuitive eating has rarely been explored in relationships even though it fits nicely with both individual theories on health behavior and systemic theories of relational health influence.

Theoretical Orientation

Dyadic Health Influence Model and Self-Determination Theory

The Dyadic Health Influence Model suggests that one’s health beliefs and behaviors, one’s beliefs about partners’ health, each partner’s relational beliefs, and the target’s health beliefs and behaviors, all play a part in the behaviors modeled in relationships, the behaviors used to support relationships, as well as direct influence of partners’ behavior (Huelsnitz et al., 2022). As such, utilizing the Dyadic Health Influence Model allows us to examine key theoretically supported relationships between intuitive eating behaviors and interpersonal eating regulation strategies (partners’ health behaviors, health beliefs, and influence strategies; Huelsnitz et al., 2022). Importantly, the Dyadic Health Influence Model also allows us to examine established intuitive eating relationships.

In addition, an important part of intuitive eating is the absence of restraints from any specific foods, typically allowing oneself to eat any food unconditionally (Tylka & Kroon Van Diest, 2013). In this way, intuitive eating may increase autonomy in choices regarding health, thereby increasing feelings of intrinsic motivation (Ryan & Deci, 2000) as well as feelings of self-efficacy and potential likelihood of behavioral change (Bandura, 1977). Partner influence on intuitive eating can also be viewed through the lens of Self-Determination Theory, which describes how greater autonomy corresponds to greater intrinsic motivation in which the benefit is inherent enjoyment and satisfaction rather than satisfaction derived from meeting a standard (Ryan & Deci, 2000). Intuitive eating holds a higher level of autonomy for individuals than does eating following restrictions of typical diets; when looking specifically at the influence of others on intuitive eating, partner autonomy support influences intuitive eating behaviors through autonomous eating regulation (Carbonneau et al., 2015). The potential connections between intuitive eating and greater intrinsic motivation are seen in that higher overall intuitive eating scores are significantly related to higher intrinsic regulation, an indicator of intrinsic motivation (Gast et al., 2015). Among men, overall higher intuitive eating scores are significantly predictive of lesser external and introjected regulation, indicating less extrinsic motivation (Gast et al., 2012). While greater intuitive eating did correspond to greater identified and intrinsic regulation, associations were not significant (Gast et al., 2012). Testing partners’ health support and control in relation to intuitive eating will add to our understanding of Self-Determination Theory and interpersonal styles of control and autonomy as well as our understanding of relatedness (Carbonneau et al., 2015; Ryan & Deci, 2000). Further, Self-Determination Theory describes relatedness to the behaviors modeled by others around oneself (this pathway is described among couples in the Dyadic Health Influence Model; Huelsnitz et al., 2022) as a process by which internalization of unmotivating behaviors, such as eating with respect to internal cues of hunger and satiety, become motivating and valued (Ryan & Deci, 2000).

The Present Study

Given the higher risk of some sexual minority individuals for disordered eating and health concerns (Nagata et al., 2020), as well as the evidence-based health benefits of intuitive eating (Linardon et al., 2021), we propose that intuitive eating is a useful measure to assess individual and dyadic health behaviors among this particular group. Despite the existing evidence base the IES-2, a widely used instrument for measuring intuitive eating behaviors, has not been examined among sexual minority couples. Though individual eating behaviors and food choices are highly influenced by significant others (Markey et al., 2008), most studies rarely consider using a dyadic or systemic lens to evaluate individuals’ eating behaviors. Exploring the IES-2 in a sexual minority population and understanding the connections between eating behaviors and other factors of interpersonal and relational functioning may provide useful insights for intervention.

The present study has four aims: to examine the measurement model of the IES-2 in a sample of married male sexual minority couples using factor analysis (Aim 1); to test the construct validity of the IES-2 by examining the correlations among the overall IES-2 score, subscale scores, and other empirically validated mental health measures (Aim 2); to explore the dyadic associations between individuals’ IES-2 scores and their partner’s individual and relational functioning (Aim 3); and, finally, to examine the correlations between both partner’s IES-2 (Aim 4).

Methods

Procedure and Participants

The data of the current study were collected in February 2018 via Qualtrics Panels (Qualtrics, Provo, UT, USA). While online panel sampling using Amazon Mechanical Turk has demonstrated low validity (Burnette et al., 2022), a meta-analysis of many different panel sample outlets including Mechanical Turk and Qualtrics Panels showed criterion validity in comparison with data collected in more traditional methods (Walter et al., 2019). Inclusion criteria included that participants must be male, married to a maleFootnote 1partner with whom they live, at least 18 years old at data collection, able to complete the survey in English, and commit to both partners’ completion of the surveys. Out of the 950 individuals who expressed interest, 228 couples were eligible to participate in the study. Informed consent was obtained from both partners in each couple. The average completion time per couple was 50.62 min. This study was approved by the Institutional Review Board at Utah State University. Participants provided informed consent before participating.

The analytical sample of the current study was 228 male sexual minorityFootnote 2 couples (456 individuals) who reported to be in a married relationship. Average age at data collection was 52.58 years old. Relationship length ranged from 1 to 52 years with an average of 19.08 years (SD = 12.13). Eighty percent of the participants identified as white, 8.8% identified as Latino or Hispanic, 5.3% Identified as African American or Black, and 5.9% identified to be in other ethnicity/racial groups. The sample was relatively well-educated; 35.5% reported having a graduate or professional degree, 32% reported having a bachelor’s degree, 17.8% reported having some college education but did not graduate, 8.1% reported having an associate degree, and 6.6% reported as high school graduates or less than high school graduates.

Measures

Intuitive Eating Scale-2 (IES-2)

Couples’ intuitive eating behaviors and attitudes were measured using the 23-item IES-2 (Tylka & Kroon Van Diest, 2013). The IES-2 has shown differences in measurement across cultures and populations (Camilleri et al., 2015; Nejati et al., 2021; Vintilă et al., 2020; Swami et al., 2020; Tylka & Kroon Van Diest, 2013). Previous research has established good internal consistency, test–retest reliability, as well as construct validity of this measure (Bas et al., 2017; Carbonneau et al., 2016; Tylka & Kroon Van Diest, 2013). The IES-2 consists of four subscales each capturing a specific aspect of intuitive eating attitudes and behaviors: Unconditional Permission to Eat, Eating for Physical Rather than Emotional Reasons, Reliance on Internal Hunger and Satiety Cues, and Body-Food Choice Congruence (Tylka & Kroon Van Diest, 2013). Items in the IES-2 are on a five-point Likert scale ranging from 1 strongly disagree to 5 strongly agree; items 1, 2, 4, 5, 9, 10, and 11 are reversed coded. The overall intuitive eating score is calculated by averaging each of the completed items, with higher scores indicating higher levels of intuitive eating; scores can also be calculated for each of the subscales (Tylka & Kroon Van Diest, 2013). Item text is available in the appendix of the original publication of the IES-2 (Tylka & Kroon Van Diest, 2013).

Mental Health Symptoms

Participants’ mental health symptoms were assessed using the four-item Patient Health Questionnaire-4 (PHQ-4; Lowe et al., 2009) with two items each for anxiety and depressive symptoms. Example items are: “Over the last two weeks, how often have you been bothered by the following problems, (1) Not being able to stop or control worrying, (2) Feeling down, depressed or hopeless.” Reponses are on a four-point scale ranging from 0 not at all to 3 nearly every day. Cronbach’s Alpha for the current sample was α = 0.91.

Couple Satisfaction Index (CSI)

The four-item Couple Relationship Index (Funk & Rogge, 2007) was used to assess couple satisfaction (sample item: “In general, how satisfied are you with your relationship?”). Three of the items are on a six-point Likert scale ranging from 1 (not at all) to 6 (completely), and one item is on a seven-point scale ranging from 1 (extremely unhappy) to 7 (perfect). The three items that are on a six-point scale were adjusted to be seven-point, and the mean score was calculated using all four items. Cronbach’s alpha coefficient for the current sample was α = 0.94.

Health Support and Health Control

This 8-item Spouses’ Provision of Health-Related Support and Control Scale (Franks et al., 2006) was used to measure the frequency of the participants’ use of health-related support or control. An example item for health-related support is, “how often have you: listened to your spouse’s concerns about protecting her/his health”; an example item for health-related control is, “how often have you: tried to influence your spouse’s choices about protecting her/his health.” All items ranged from 0 never to 4 every day. Cronbach’s alpha coefficient was α = 0.86 for health-related support and α = 0.91 health-related control.

Encouragement to Diet, Encouragement to Change Dietary Habits, and Diet Undermining

The ten-item Social Support and Eating Habits Survey (Sallis et al., 1987) assessed how their partner influenced their dietary behaviors over the past three months. Three items represented encouraging their partner to diet (example item is “encouraged me not to eat ‘unhealthy foods’ (cake, salted chips) when I’m tempted to do so.”), two items reflected the partner’s encouragement to change dietary practices (example item is “discussed my eating habit changes with me”), and five items reflected a partner’s diet undermining strategies (example item is “brought home foods I’m trying not to eat”). Answers ranged from 1 = none to 5 = very often. Summed scores are used. All three of the subscales had good reliability in the current sample (encouragement to diet α = 0.87, encouragement to change dietary practice α = 0.78, diet undermining α = 0.85).

Analytic Plan

Because of the non-independent and indistinguishable nature of our dyadic data, we conducted dyadic Confirmatory Factor Analysis (CFA) where we set item intercepts, factor loadings, residual variances, and residual covariances to be equal for both partners in order to investigate how well the original factor structure of the IES-2 fits the current sample (Olsen & Kenny, 2006). Analyses were conducted in R Studio (Rstudio Team, 2020) using the lavaan package (Rosseel, 2012; Aim 1). Comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), and standardized root mean residual (SRMR) are used to evaluate models’ goodness of fit; values of CFI and TLI greater than 0.90 as well as values of SRMR and RMSEA below 0.08 were used as the cut-off values to determine the model fit (Hair et al., 2010; Hu & Bentler, 1999). Based on the inadequate fit of the CFA test with the original factor structure (see Results section), we proceeded with a series of Exploratory Factor Analyses (EFA; Aim 1). Before conducting the EFA, we conducted Kaiser–Meyer–Olkin (KMO) test to examine the adequacy of our sample for an EFA test. We also used eigenvalues (> 1) and scree plot to help determine the number of EFA models conducted. Based on the results that there were five factors that had eigenvalues > 1 and that the scree plot showed a steep drop at factor 5 (Eigenvalues for the first five factors ranged from 6.02- 1.12; see Supplemental Fig. 1 for scree plot and Supplemental Table 1 for full Eigenvalues), EFA models with 1–5 factors were conducted. We followed Tabachnick and Fidell’s (2021) recommendations to determine item retention or omission (e.g., omit below 0.32). Internal consistency of the final factor structure was examined using Cronbach’s alpha. For all the CFA and EFA analyses, we collapsed respondents’ dyadic data into individual level data resulting in a full sample of 456 individuals.

Table 1 Dyadic CFA with the original IES-2 scale

After conducting CFA and EFA analyses, we restructured the data to reflect the couple pairs and separated respondent data by Partner 1 and 2. To investigate the construct validity of the scale, we rescored each of the IES-2 subscale scores and the overall score based on our EFA results and conducted bivariate correlations among the overall IES-2 score, the subscale scores, and mental health symptoms at the individual level using the paired data (Aim 2). Further, we proceeded with conducting bivariate correlations among the IES scores and relational functioning variables including couple relationship satisfaction, health support and control, as well as diet support and eating undermining strategies to explore the dyadic aspects of intuitive eating behaviors in dyads of married sexual minority male individuals (Aim 3). Then we proceeded to run bivariate correlations between the two partners to investigate couple-level correlations among the overall IES-2 score and the subscale scores (Aim 4). Materials and analysis code for this study are not available. This study was not preregistered.

Results

Aim 1: Exploration of IES-2 in Married Male Sexual Minority Couples

Confirmatory Factor Analysis with the Original Factor Structure

The dyadic CFA with the originally validated IES-2 factor structure indicated poor model fit (CFA = 0.69, TLI = 0.67, RMSEA = 0.1, SRMR = 0.13). Items’ loadings ranged from −0.04 to 0.95, and four of the items had a loading of < 0.4 across both partners (See Table 1 for full item loadings). The insufficient model fit was likely due to the number of items with low loadings. We proceeded with conducting a series of CFA models each dropping one item starting with the item that had the lowest loading (Khalsa et al., 2019). The items that were dropped in the alternative models were items 9, 5, 4, and 12 respectively. Model fit did not improve to an acceptable level in the four alternative models.

Exploratory Factor Analysis

Exploratory Factor Analysis (EFA) was conducted following the insufficient model fit results indicated by the dyadic CFA with the original IES-2. Models with 1–5 factors were conducted, resulting in the model with five factors being the most optimal model (See Supplemental Table 2). Although the four-factor and the five-factor model both demonstrated good fit with the data, we eventually selected the five-factor model to be the best fitting model for a list of evidence. First, based on Tabachnick and Fidell’s (2021) recommendation to omit items below the 0.32 cut-off, we were able to retain all 23 items on the scale with the five-factor structure, whereas we will have to drop item 4 (λ = 0.24) in the four-factor structure. Furthermore, according to the original IES-2, items 12–15 should load together with item 2, 5, 10, and 11. In the four-factor model, when loaded on the same factor, items 12–15’s loadings (λ = 0.26–0.43) were way lower than item 2, 5, 10 and 11 (λ = 0.92–0.96). Items 12–15 also cross loaded on two other factors with better factor loadings (λ = 0.32–0.62). The five-factor structure, however, does not have issues with cross loading. Items 12–15 that failed to emerge a clear pattern in the four-factor structure became the fifth factor in the five-factor structure with better factor loadings (λ = 0.67–0.80). Finally, the five-factor model did indicate a slightly better model fit (CFI = 0.999, TLI = 0.999, χ2 (148) = 211.03, p < 0.01, RMSEA = 0.031, SRMR = 0.031) than the four-factor model (CFI = 0.995, TLI = 0.993, χ2 (167) = 504.71, p < 0.01, RMSEA = 0.067, SRMR = 0.048). Full factor loadings of the five-factor structure can be found in Table 3, and the factor loadings of the four-factor structure can be found in Table 2.

Table 2 Four-factor exploratory factor analysis results
Table 3 Five-factor exploratory factor analysis results

The final five factors suggested by the EFA results are: Unconditional Permission to Eat (retained their original six items; α = 0.76; explains 7.07% of total variance), Body–Food Choice Congruence (retained their original three items; α = 0.865; explains 4.87% of total variance), Reliance on Hunger and Satiety Cues (retained their original six items; α = 0.885; explains 26.18% of total variance), Eating for Physical Rather Than Emotional Reasons (retained four original items; items 12, 13, 14, and 15 from this original subscale loaded on the new generated subscale, Eating Not as Coping; α = 0.946; explains 16.03% of total variance), and Eating Not as Coping (the new generated factor of 4 items; α = 0.837; explains 15.30% of total variance). Cumulatively, all five factors explain 69.45% of total variance. See Table 3 for the final five-factor EFA model results.

Aim 2: Construct Validity of IES-2 in Married Male Sexual Minority Couples

Psychological Constructs

The bivariate correlation results indicated that higher levels of one’s overall intuitive eating behaviors were associated with their own lower anxiety for both partners and depressive symptoms for one partner but not for both. Higher levels of Eating for Physical Rather than Emotional Reasons shows significant associated with one’s own lower anxiety and depressive symptoms and with their partner’s lower anxiety and depressive symptoms for both partners. See Table 4 for full results of the bivariate correlations between revised IES-2 full scale, subscales, and mental health symptoms.

Table 4 Bivariate correlations between IES full score, subscales, and mental health symptoms

Aim 3: Dyadic Association between IES-2 and Relational Constructs

Individually and dyadically, many intuitive eating behaviors were associated with one’s own relational outcomes and the interactions they have with their partners related to health. For individuals and crossing over between partners, higher relationship satisfaction was associated with higher overall intuitive eating, Body-Food Choice Congruence, Reliance on Hunger and Satiety Cues, and Eating Not as Coping. For health-related relational measures, higher Body-Food Choice Congruence was associated with more provision of health support individually and between partners, whereas lower Eating for Physical Rather than Emotional Reasons was associated with less provision of health control.

Higher scores of being encouraged to diet by one’s partner were associated with higher scores of Body-Food Choice Congruence and Reliance on Hunger and Satiety Cues of one’s partner and lower scores of Eating for Physical Rather than Emotional Reasons of oneself and one’s partner. Higher scores of being encouraged to change one’s diet by one’s partner were associated with higher scores of Body-Food Choice Congruence of oneself and lower scores of Eating for Physical Rather than Emotional Reasons of oneself and one’s partner and lower scores of Unconditional Permission to Eat of oneself and one’s partner. Higher scores of receiving dietary undermining from one’s partner were associated with lower scores of Eating for Physical Rather than Emotional Reasons for both oneself and one’s partner. See Table 6 for full results of the bivariate correlations between revised IES-2 full scale, subscales, relational constructs, and measures regarding eating behaviors.

Aim 4: Couple Level Correlation

Bivariate correlations were run among each of the subscales and the modified full intuitive eating scale to investigate couple-level correlations using paired data (identified by partner 1 and partner 2). Using the new factor structure, significant bivariate correlations ranged from r = 0.269 to 0.498 between partners’ reports. These results indicated low to moderate levels of correlations between each of the factors of intuitive eating behaviors at the couple level. See Tables 4, 5, and 6 for full results of couple level correlations.

Table 5 Bivariate correlations between IES full score, subscales, CSI, health support and control, diet encouragement, eating behavior change encouragement, and eating undermining strategy from partner
Table 6 Bivariate correlations between IES full score, subscales, CSI, health support and control, diet encouragement, eating behavior change encouragement, and eating undermining strategy from partner

Discussion

The present study aimed to investigate and examine the factor structure of the IES-2 scale and explored associations between both partner’s mental health, relational quality, and interpersonal eating interactions within a dyadic sample of 228 married male sexual minority couples. Results revealed some differences in factor structure in our sample of married male sexual minority couples compared to previous validation with more general populations (Tylka & Kroon Van Diest, 2013). Additionally, we found evidence supporting that one’s intuitive eating behaviors were associated with their own and their partner’s mental health symptoms, their own and their partner’s relationship satisfaction, as well as their health-related interactions with one another. These findings highlight important avenues for dyadic health promotion and intervention.

Factor Analysis (Aim 1)

Results of our dyadic CFA model indicated that the original four factor structure of the IES-2 scale did not fit the present sample of married male sexual minority couples. It is well-supported that intuitive eating behaviors are associated with improved outcomes regarding body image and pathological eating concerns (Hazzard et al., 2021) that are salient for sexual minority men (Brown & Keel, 2015). This finding furthers our understanding that sexual minority individuals may have a different experience when it comes to body satisfaction, and both disordered and intuitive eating-related behaviors in comparison to heterosexual individuals.

Specifically, our EFA analyses generated a new factor that was not captured in the original IES-2 scale, thus signaling some intuitive eating behaviors that may be unique to male sexual minority couples. This factor, Eating Not as Coping, consists of four items (12, 13, 14, 15) that loaded on the Eating for Physical Rather than Emotional Reasons subscale in the original validation of IES-2 (Tylka & Kroon Van Diest, 2013); these four items regard negative emotionality and are worded such that a greater endorsement of the item corresponds to not eating for physical reasons in the presence of negative emotions. The original Eating for Physical Rather than Emotional Reasons subscale was broader conceptually than the current model in including these four items in conjunction with items that are worded such that a greater endorsement of the item corresponds to finding oneself eating in the presence of negative emotions. There may be a meaningful difference in cognition when one passively finds oneself engaging in the undesired eating behavior rather than when one actively refuses the undesired eating behavior. This difference in factor structure may reflect differences between active coping skills and passive behaviors that are less so the result of the individual’s intentions and control. As some of the questions on the new subscale added in the present model ask about ability to cope and active coping skills aside from eating, one reason for this difference in the present factor structure and the original factor structure of the IES-2 may be that capacity for coping skills, or resilience, is notable among sexual minority individuals and could override the need to eat as coping (Handlovsky et al., 2018).

Construct Validity (Aim 2) and Dyadic Association Between IES-2 and Relational Constructs (Aim 3)

Individual Associations

The significant relationships that emerged between an individual’s overall score of intuitive eating and lower levels mental health symptoms evidence the potential of intuitive eating for improving mental health and understanding minority stress for married male sexual minority couples (Meyer, 2003). Among the subscales, an individual’s Eating for Physical Rather than Emotional Reasons showed an association with less anxiety and depressive symptoms of themself. Three subscales of intuitive eating (Body-Food Choice Congruence, Reliance on Hunger and Satiety Cues, and Eating Not as Coping) and overall intuitive eating showed a connection with greater individual relationship satisfaction. The connection between individual overall intuitive eating and greater individual relationship satisfaction is notable as relationship satisfaction plays a part in dyadic stress processes (Randall & Bodenmann, 2017).

While no overall effects were shown between overall individual scores of intuitive eating with health support/control and dietary encouragement/pressure to change/undermining, individual scores on intuitive eating subscales reveal that greater Eating for Physical Rather than Emotional Reasons corresponds to providing lesser health control to one’s partner. While the Dyadic Health Influence Model does not include a direct path from agent’s health behavior (e.g., Eating for Physical Rather than Emotional Reasons) to agent’s influence strategies (e.g., providing lesser health control to one’s partner), one’s Eating For Physical Rather than Emotional Reasons may also represent agent’s health beliefs, which impact agent’s beliefs about target’s health which subsequently impact agent’s influence strategies in the Dyadic Health Influence Model (Huelsnitz et al., 2022).

Greater Body-Food Choice Congruence, on an individual level, corresponded to providing greater health support for one’s partner. This may also reflect the indirect pathway from agent’s health beliefs to agent’s influence strategies described in the Dyadic Health Influence Model mentioned above (Huelsnitz et al., 2022). If one is able to not use food as a coping mechanism while also providing health support to their partner, it is possible that they may have relatively fewer stressors or greater coping resources in their life. One such coping resource may be trust in oneself, which could be reflected in the Body-Food Choice Congruence subscale. Further, perhaps those who have better eating regulation habits feel that they can better support their partner, whereas those with poorer eating regulation do not feel justified in providing support.

The Dyadic Health Influence Model describes a direct pathway from agent’s influence strategies to target’s health behaviors, which could explain the correlations seen between these subscales of eating behavior and the influence received from one’s partner (Huelsnitz et al., 2022). For example, individual Eating for Physical Rather than Emotional Reasons corresponds to receiving lesser dietary encouragement/pressure to change/undermining, and individual Body-Food Choice Congruence is connected to receiving greater encouragement to make dietary changes. Individual Unconditional Permission to Eat showed a connection with receiving less encouragement to change one’s diet, potentially reflecting a shared couple health belief (less restraint of eating behaviors) as seen in the pathway from agent’s health beliefs to target’s health beliefs via agent’s health behaviors (Huelsnitz et al., 2022).

Dyadic Associations

As mentioned, relationship satisfaction has implications for dyadic processes of stress (Randall & Bodenmann, 2017) which could hold potential for building resilience to minority stress. Overall IES score as well as scores on the Body-Food Choice Congruence, Reliance on Hunger and Satiety Cues, and Eating Not as Coping subscales of one partner were significantly associated with greater relationship satisfaction reported by the other partner, indicating that participating in intuitive eating may hold benefits for one’s partner(s) and relationship(s) with implications for improving health. Dyadic associations also reveal connections between one’s health behaviors and the relational behaviors of their partner(s) such as the connection between one’s own scores of Body-Food Choice Congruence and the health support provided by one’s partner. This also reflects the Dyadic Health Influence Model pathway between target’s health behaviors and agent’s influence strategies (Huelsnitz et al., 2022).

Attempts to encourage or undermine one’s partner in their diet and eating behaviors is related to some intuitive eating behaviors in oneself. Higher levels of Eating for Physical Rather than Emotional Reasons correspond to lower levels of encouragement to diet or to change one’s partner’s diet habits as well as undermining one’s partner’s diet habits, reflecting the effects of agent’s health beliefs on beliefs about target’s health on agent’s influence strategies (Huelsnitz et al., 2022). One’s own Unconditional Permission to Eat is related to less attempts to change one’s partners diet. Conversely, Greater individual Body-Food Choice Congruence is related to greater encouragement of one’s partner to diet or to change their diet habits. This increase in agent’s influence strategies could be related to the increase in the target’s intuitive eating behaviors, demonstrating that these influence strategies may be changing one’s partner’s eating behavior, although directionality is unclear (Huelsnitz et al., 2022). This connection between greater Body-Food Choice Congruence and encouraging one’s partner to make dietary changes stands in opposition to the interpretation of our other findings that greater intuitive eating behaviors reflect greater appreciation for autonomy and intrinsic motivation for oneself and for others (Ryan & Deci, 2000).

Dyadic Associations of Intuitive Eating and Subscales (Aim 4)

In support of the protective contexts of relationships as well as the influences of partners on each other’s health behaviors, our results show that for overall intuitive eating scores and all five subscales (Body-Food Choice Congruence, Reliance on Hunger and Satiety Cues, Unconditional Permission to Eat, Eating Not as Coping, and Eating for Physical Rather than Emotional Reasons), spouses shared similar eating behaviors and health beliefs with implications for mental health, relationship satisfaction, and health support of one’s partner. These findings are important for both a theoretical understanding of the interdependence and dynamic nature of partners’ health behaviors and beliefs as well as an applied understanding for both dyadic health interventions as well as the therapeutic practice with partnerships of sexual minority individuals, specifically for male sexual minority individuals who are married or in long-term relationships.

Clinical Relevance and Application

The results from our study have important implications for intervention. First, because we demonstrated some variance in the construct of intuitive eating in sexual minority clients, clinicians should discuss the unique ways in which their sexual minority clients view their own body image and their attitudes, beliefs, and behaviors around food, particularly when working with disordered eating or eating disorders. In addition, clinicians would do well to discuss larger sociocultural and societal factors related to the unique experiences of sexual minority status and how it impinges on or affects their eating attitudes, beliefs, and behaviors. Finally, our results demonstrate the interconnectedness between nutrition/diet and psychosocial functioning. Thus, relational therapists are in a unique position to systemically assess dynamics around nutrition/eating and the way in which in interacts with relational and psychological health (Novak et al., 2021). This may include situations in which clients present with nutrition-related concerns where relational dynamics may be impinging upon dietary change, such is the case with dietary undermining (Novak et al., 2021) or emotional eating, or when relational dynamics could be improved as a result of addressing nutrition-related concerns and challenges (e.g., low blood sugar, rigid rules or roles around food, and conflict around these domains).

Limitations, Constraints on Generality, and Directions for Future Research

Though this study has many strengths as it is the first known study examining the IES-2 in a partnered sexual minority population and highlights the critical role of dyadic health management in couple relationships using dyadic data, the results should be viewed in light of its limitations. First, the current study sample is not representative of the larger sexual minority community or communities of gay, bisexual, and sexual minority men or male individuals. The current study only included married male sexual minority couples—married same-sex couples only account for about 59% of all same-sex couples in the U.S. (U.S. Census Bureau, 2021). Thus, considering the potential protective effect that marriage has on health (Rendall et al., 2011), results of the current study may not be generalizable to non-married couples. Though sexual minority men continue to be at higher risk for eating disorders, sexual minority adults in general are more likely to be diagnosed with an eating disorder than heterosexual individuals, and the prevalence of diagnosis of eating disorders is higher in transgender men and women compared to cisgender individuals (Nagata et al., 2020), differences not presently delineated. Recent literature also points out that identifying as an ethnic minority individual is another risk factor for sexual and gender minority individuals when it comes to experience with disordered eating behaviors (Calzo et al., 2017). Notably, the current sample is not economically or ethnically diverse, which introduces another limitation on generalizability of the results. Future studies should further the effort of validating IES-2 by using a more inclusive sexual and gender minority sample, a comparison group of heterosexual partners, as well as considering the role that intersectionality plays in relation to risk of eating disorders. Without this additional testing, researchers and clinicians hoping to use the IES-2 with sexual minority populations should proceed with caution because of potential measurement invariance as evidenced here. Additionally, the current sample was not screened for eating disorders or eating disorder histories. Thus, the results are not generalizable to a clinical sample.

Another limitation is the lack of measures related to disordered eating behaviors or body image satisfaction/dissatisfaction. There is an established inverse link between intuitive eating behaviors and eating pathology as well as body image issues (Linardon et al., 2021), which makes these constructs important parts of validating IES-2. The current study only included brief mental health related measures (depressive symptoms and anxiety symptoms); future studies should consider including other mental health and wellbeing related constructs as part of testing the construct validity for IES-2.

Furthermore, participant self-selection into the study might indicate a level of interest in couple relationship functioning and health behaviors, which may result in selection bias. The self-reported nature of the data also introduced social desirability bias especially considering reporting on questions that are socially constructed to be “healthy” or “unhealthy.” Future studies may benefit from using other tools to measure eating behaviors (e.g., daily food diaries).

Conclusion

This is the first study of intuitive eating among married couples of sexual minority male individuals. Our findings demonstrate that sexual minority populations may interpret the IES-2 differently as reflected in the generation of a new factor (Eating Not as Coping). In addition, we found evidence supporting that intuitive eating behaviors were associated with their own and their partner’s emotional and relational health, as well as interpersonal interactions related to food. Thus, our findings not only underscore the importance of examining how sexual minority individuals experience cultural narratives, beliefs, and attitudes around food, but also on the ways interpersonal relationships can serve to protect or undermine them.