Obesity and obesity-related conditions are the leading causes of death in the United States [1], making them a significant public health concern. On the basis of several recent national reports [2,3], there are disparities in the health and well-being of sexual minority individuals (eg, gay, lesbian, and bisexual) as compared with their heterosexual peers in a range of health conditions and health behaviors. In particular, large-scale studies suggest that sexual minority women are at least twice as likely to be obese as heterosexual women [4,5], and studies of young women have similarly found higher rates of being overweight or obese in young lesbian women (35.2%) than in young heterosexual women (22.8%) [6]. There is also evidence that rates of binge eating, defined as overeating with loss of control of eating, are similarly higher in sexual minority individuals. For example, in a study of nearly 14,000 people, young adult lesbian women were twice as likely to binge eat as heterosexual women [7]. Despite these documented disparities, relatively little is known about the contributing factors, which is a critical first step in developing interventions to address these health concerns among sexual minority women.

In working toward increased understanding of factors that contribute to higher obesity and binge eating rates in sexual minority women, the goal of this study is to use ecological momentary assessment (EMA) to identify general and minority-specific real-world predictors of binge eating. EMA involves the use of repeated assessments of people’s current behaviors, states, and experiences as they are going about their lives [8]. EMA approaches provide unique advantages over other research methods (eg, cross-sectional and traditional longitudinal methods), including greater real-world generalizability and reduced concerns regarding memory biases associated with retrospective recall that are generally present in traditional measures. In addition, as EMA uses repeated assessments (typically multiple times throughout the day), researchers can examine dynamic within-person processes and answer research questions that cannot be addressed using other methods [8,9]. For example, most previous research on binge eating among sexual minority women has used cross-sectional designs, which can address questions regarding between-person processes (eg, do women who are more depressed also binge eat more?). EMA approaches provide an opportunity to answer questions regarding within-person processes (eg, when women experience a more depressed mood, are they more likely to binge eat?). This study—called the Health and Experiences in Real Life (HER Life) Project—will examine within-person associations between a range of general predictors of binge eating (ie, affect, body dissatisfaction, social processes, and health behaviors) in young heterosexual and sexual minority women as well as sexual minority–specific factors (eg, minority stress and discrimination) among young sexual minority women. The existing empirical evidence and theoretical framework supporting this study are described in the following sections.

The factors influencing disordered eating are complex, and researchers have highlighted the potential utility of EMA to study disordered eating, including binge eating, in natural settings [10]. EMA research shows that disordered eating, including binge eating, is associated with affective states (eg, negative mood, stress, and body dissatisfaction) [11,12], social experiences (eg, interactions and media use), and health behaviors (eg, alcohol use and physical activity) in studies of young women where sexual orientation was not assessed or reported (hereafter referred to as general samples). However, with few exceptions [13], EMA studies have yet to consider how affective states, social processes, and health behaviors in daily life are associated with sexual minority women’s binge eating, which is a primary aim of this study.

The role of negative emotions (eg, negative affect and stress) in binge eating is often examined as a way of evaluating the affect regulation model of eating disorders [14]. A meta-analysis of EMA studies evaluating the affect regulation model of binge eating found that negative emotions increased leading up to a binge eating episode, with more mixed findings regarding negative affect after binge eating episodes [12]. Negative feelings about one’s body, or body dissatisfaction, have also been examined via EMA in general samples of young women but, with few exceptions [15], studies have generally focused on associations between body dissatisfaction and social processes, not disordered eating behaviors [16-18]. Although these studies show that negative mood and body dissatisfaction may be associated with binge eating in daily life, this work is largely based on general samples of young women, and whether and how these processes operate in sexual minority women is less clear. There is some preliminary evidence demonstrating that negative affect is associated with binge eating at the day level in a daily diary study of 30 lesbian women [13], but there are no EMA studies of body dissatisfaction in sexual minority women. Cross-sectional studies designed to answer between-person research questions have produced mixed findings, with some suggesting that body dissatisfaction is similar in sexual minority and heterosexual women [19] and others finding that larger body sizes are more acceptable among sexual minority social groups and, thus, body dissatisfaction may be lower [5]. However, as described previously, this study will fill gaps in the literature by assessing negative affect, stress, and body dissatisfaction in daily life to examine how they are associated with binge eating among both young sexual minority and heterosexual women.

There is a long history of research demonstrating the contribution of sociocultural processes to disordered eating behaviors [20,21]. In fact, the sociocultural model of disordered eating suggests that there are social pressures to be thin (and, increasingly, to be fit or toned) in Western countries, which are communicated directly and indirectly through the media, peers, and family and can become internalized (ie, thin ideal internalization) [22]. In recent years, several EMA studies have examined how social experiences are associated with body dissatisfaction and eating behaviors in daily life. For example, in a general sample of 121 women aged 18 to 40 years, women reported greater body dissatisfaction in social situations (ie, when they had had an interaction in the previous 30 minutes) than when alone and particularly when they perceived that the interaction quality was lower [18]. To our knowledge, studies on the effect of social processes in daily life on eating behaviors have been conducted only among general samples of young women and, thus, it remains unclear whether sexual minority women experience the same sociocultural influences as heterosexual women in daily life. Cross-sectional research suggests that larger body sizes may be more acceptable in sexual minority women’s social groups and relationships [5,23] and, thus, they may be less susceptible to thin ideals that they are exposed to in social interactions and the media. However, as described by the dual identity framework, there may also be commonalities between sexual minority and heterosexual women [24]. According to this framework, sexual minority women are influenced, as women, by the mainstream (heterosexual) community and, as sexual minority individuals, by the sexual minority community. Thus, sexual minority women’s eating behaviors will likely be influenced by general social processes but potentially to a lesser extent than heterosexual women. The HER Life Project will explore social influences on both sexual minority and heterosexual women in daily life, thus allowing for a better understanding of how these dual influences may operate for young sexual minority women.

In addition to the demonstrated importance of sociocultural considerations for understanding disordered eating, cross-sectional and meta-analytic studies have shown that hazardous alcohol use is associated with disordered eating in general samples of young adults [25] and among lesbian women. For instance, a cross-sectional survey found that lesbian women with obesity were more likely to report heavy drinking compared with lesbian women who were not overweight or obese [26]. Research on the associations between binge eating and physical activity are more equivocal [27], but 1 study with a general sample of young women found that those with binge eating disorder reported less physical activity compared with weight-matched women who did not have binge eating disorder [28]. Even less is known about physical activity and binge eating in sexual minority women, but 1 study of lesbian and bisexual women found that being overweight or obese was associated with less frequent exercise [29]. It is clear that there is a dearth of research on the associations among alcohol use, physical activity, and binge eating, especially in sexual minority women. This is particularly notable because of the documented disparities in alcohol use and physical activity, with sexual minority women engaging in more hazardous drinking [30-32] and less physical activity [33,34] than heterosexual women. This study will assess alcohol use and physical activity in daily life, which will allow for a real-world examination of how these health behaviors may operate among sexual minority and heterosexual young women, and explore potential associations with binge eating.

In addition to the aforementioned general factors that may influence binge eating, sexual minority women face unique experiences and stressors because of their marginalized and stigmatized status in society. This phenomenon has been described in minority stress theories by Meyer [35] and Hatzenbuehler [36]. These theories complement each other in describing how general and minority-specific stressors can affect mental health and well-being. For example, the minority stress theory by Meyer argues that sexual minority individuals experience stressors related to their minority status. These stressors include chronic and acute external events (discrimination, harassment, invalidations, and vicarious experiences through other sexual minority individuals) and internal stressors (concealment of sexual identity, expected rejection, and internalization of negative societal messages about sexual minorities [internalized heterosexism]) that can negatively affect mental health. The psychological mediation framework by Hatzenbuehler extended these ideas to suggest that general (eg, coping, emotion regulation, and social support) and minority-specific (eg, distal and proximal minority stressors) experiences operate together to influence sexual minority individuals’ well-being. These theories have been extended from their initial focus on understanding mental health to also consider the impact of minority stressors on physical health and health behaviors.

In recent years, minority stress theories have been used to identify potential mechanisms to explain health disparities between sexual minority and heterosexual individuals, including disordered eating. Cross-sectional studies of sexual minority women have found between-person associations between heterosexist experiences (eg, harassment, rejection, and sexual orientation discrimination) and disordered eating [37-39]. More recently, there have been several daily diary and EMA studies that have assessed within-person associations between minority stressors and disordered eating behaviors at the day or moment level. For example, in a sample of 30 young lesbian women who reported binge eating, researchers found that, on days when the women reported more discrimination, they also reported more binge eating (and this effect occurred in part through negative affect) [13]. In an EMA study of 55 sexual minority women who were overweight or obese, Panza et al [40] similarly found that those with higher baseline levels of internalized homophobia and sexual orientation concealment reported more binge eating and overeating during a 5-day EMA period. Taken together, sexual minority stress theories and empirical research suggest that sexual minority stress is associated with poor mental health, with substantially fewer studies considering the effects on binge eating. The goal of this study is to examine the within-person direct and indirect (through negative affect) effects of sexual minority stressors on binge eating in a subsample of sexual minority women. In other words, this study allows us to consider whether, at times when sexual minority women experience a minority stressor, that experience is either concurrently or subsequently associated with binge eating in daily life. Such information may improve the understanding of the potential mechanisms whereby sexual minority women experience more disordered eating behaviors as compared with heterosexual women.

Given that theories on binge eating and sexual minority health and health disparities are complex, with factors influencing these behaviors at various levels, an additional aim of this study is to consider how between-person factors may strengthen or weaken the aforementioned within-person associations. Three broad groups of moderators will be considered in this study: race, body- and eating-related factors, and sexual minority–specific factors. Considering these cross-level moderators can be useful for clarifying previous conflicting findings and provide a more complete picture of both when and for whom various real-world binge eating predictors operate.

Although cross-sectional research shows that sexual minority women are more likely to report binge eating compared with heterosexual women [7,41,42], many of these studies were conducted in samples that were not racially diverse. There is some conflicting research regarding potential racial differences in binge eating between Black and White women in the United States. Some studies in general samples of young women have found no differences in disordered eating (including binge eating) between Black and White women [43,44] but, in a general sample of women with binge eating disorder, Black women reported 1.5 times as many binge eating episodes per week compared with White women [45]. However, taken together, these studies fail to consider how multiple minority identities (ie, being a Black sexual minority woman) might interact to exacerbate health behavior risks, and there is scant research examining the effect of the intersection of sexual and racial identities on binge eating. In fact, we found only 1 study that assessed binge eating in a small, racially diverse sample of lesbian and bisexual women. This cross-sectional study found that 1.6% of White heterosexual women (n=63) reported binge eating at least twice per week compared with rates of 4.5% for White lesbian and bisexual women (n=67) and 3.1% for Black lesbian and bisexual women (n=64) [46]. Importantly, this study was cross-sectional and had a small sample. Therefore, to extend these past findings, in the HER Life Project, we powered to test for Black-White race moderator effects of the within-person associations between the general and minority-specific predictors of binge eating outlined in the first 2 study aims, allowing us to examine how intersecting minority statuses (ie, being Black and of a sexual minority) may uniquely shape eating behavior in daily life.

In addition to considering how women’s racial identity may moderate within-person associations, this study will consider 4 body- and eating-related variables: BMI, person-level (baseline) body dissatisfaction, disordered eating, and thin norm internalization. These variables were selected based on past research demonstrating that individuals higher or lower in these constructs may exhibit different associations between our predictors of interest and binge eating [17].

In addition, this study considers sexual minority–specific moderators of EMA associations, which will be considered among the subsample of sexual minority women. In particular, we will evaluate 3 protective factors (positive sexual minority identity; lesbian, gay, bisexual, transgender, and queer [LGBTQ+] community connection; and identification with the LGBTQ+ community) and one risk factor (lifetime sexual minority discrimination). Previous cross-sectional research suggests that connection to and identification with the sexual minority community may buffer against the deleterious effects of sexual minority stress [47], including perceived stigma and depressive symptoms [48]. Furthermore, in a cross-sectional study of lesbian, gay, and bisexual young adults, social support (eg, from the sexual minority community) attenuated the associations between minority stress and emotional distress [49]. Conversely, lifetime discrimination may be a risk factor for maladaptive health behaviors. In a daily diary study of Black college students, perceived lifetime racial discrimination moderated the associations between negative mood and alcohol use [50], providing evidence that (person-level) lifetime racial discrimination influences daily health behaviors. Taken together, the third aim of this study is to consider a variety of person-level moderators of the within-person associations explored in the first 2 study aims.

The broad objective of the HER Life Project is to enhance the understanding of the real-world predictors of binge eating among young sexual minority and heterosexual women. To address this objective, this study is organized around 3 aims.

The HER Life Project is an ongoing EMA study of young sexual minority and heterosexual women who binge eat. Young women aged 18 to 30 years are the focus of this study given that they report high levels of binge eating [4,5] and, thus, represent an at-risk group. We selected women aged 18 to 30 years as, although “young adulthood” has no clear definition, there is evidence that sexual identity continues to develop during this period, and data suggest associations among sexual identity, negative affect, and disordered eating up to the age of approximately 30 years [51]. Participants are being recruited from across the United States, and data collection is occurring entirely remotely as we are using web-based surveys and a smartphone app to deploy EMA surveys. Participants receive and review the written and video study procedure descriptions before beginning the study. They complete a web-based screening and baseline survey before beginning the 2-week EMA data collection period and conclude their participation with an end-of-study survey. During the EMA period, participants complete daily morning and before-bed surveys, 5 prompted surveys each day, and a user-initiated binge eating survey after binge eating episodes.

The Old Dominion University Institutional Review Board approved all the study procedures (project 1362990).

A 3-step power analysis was conducted to determine sample size. First, a regression power analysis using G*Power (version 3.1.9) [52] was conducted, powering the hardest-to-detect effects (aim 1 hypothesis 1c: sexual identity moderation; aim 3: race moderation) and the examinations using only sample subsets (aim 2: direct effects, sexual minority women only). To achieve a power of 0.80 using an α of .05, 227 independent observations are required for a small to medium effect (f²=0.035) for moderation examinations with 5 predictors (to allow for relevant covariates). Specifying the same power level, α, and effect size with 3 predictors for aim 2 main effects (to allow for relevant covariates), 227 independent observations among sexual minority women only would be necessary. Second, we applied a formula to account for the multiple correlated observations for each person and identify the number of actual participants needed [53]. This equation accounts for level-1 (assessments or moments) and level-2 (participants) sample sizes and the intraclass correlation coefficient (ICC; within-person relatedness). On the basis of past EMA studies of similar constructs among young women [54,55], a conservative estimate of the average number of assessments per person (n_L1) is 25, reflecting constructs only asked once or twice per day (as opposed to all prompts), approximately 70% compliance, and the reporting of approximately 6 binge eating episodes. If 100 sexual minority and 100 heterosexual women are recruited, this would result in approximately 5000 total observations (2500 from sexual minority women). We expect ICC values from 0.33 to 0.43 for binge eating. Using the West et al [53] formula and past studies, n_L2 of 100 sexual minority and 100 heterosexual women would result in an N-effective of 280 within sexual identity if the ICC is 0.33 and an N-effective of 221 within sexual identity if the ICC is 0.43. This means that, although there are 2500 assessments from sexual minority women, because of the degree of relatedness of observations within each individual, these are equivalent to 221 independent pieces of information. Thus, recruiting 200 women should provide sufficient power for these aims. In addition, to detect small to medium mediation effects (aim 2) using empirical bias-corrected bootstrap CIs, 148 independent observations should be sufficient [56]. Finally, to power for aim 3, half of the 100 participants in each sexual identity group (n=50) will need to identify as Black to optimize power. However, given the demographic representation in previous work examining eating behaviors in young women using EMA [54,55], it is expected that we will need to enroll >100 women of each identity to have 50 Black women in each group even if deliberately oversampling. To ensure that we are sufficiently powered to test all aims and have similar numbers in both groups, we aim to enroll 150 sexual minority women (50 Black) and 150 heterosexual women (50 Black).

Participants are being recruited using various web-based sources, including from web-based research panels, social media, and lists of people who participated in past studies and expressed an interest in future research opportunities. We are working with several market research firms who have access to web-based panels of people who are interested in participating in web-based research. Community Marketing & Insights is an LGBTQ+ market research firm that manages a proprietary panel of individuals who identify within the LGBTQ+ community and are interested in participating in web-based studies. Marketing Systems Group and Qualtrics International Inc are also assisting with accessing existing panels of individuals who regularly participate in web-based studies, primarily to recruit heterosexual women. We have also placed advertisements on social media (eg, Facebook and Instagram) using the Meta advertising platform and through advertisements on our laboratory social media page.

To be eligible to take part in the study, participants must (1) identify as women (from a multi-select question, selecting woman or woman and some combination of genderqueer, nonconforming, or nonbinary and also indicating that they were assigned female at birth), (2) be aged between 18 and 30 years (inclusive), (3) report binge eating (ie, overeating with a feeling of loss of control of eating) at least two times in the past 2 weeks, (4) not be currently receiving treatment for an eating disorder, and (5) have a schedule that allows for answering surveys during daytime hours. In addition, the participants must also meet either the sexual minority or heterosexual criteria. Women are included in the sexual minority sample if they report that they are (1) only or mostly attracted to women or (2) equally attracted to men and women or attracted to people regardless of their gender identity and also have a current or most recent romantic partner who identifies as a woman. Women are included in the heterosexual sample if they report that they are only or mostly attracted to men.

The measures in the baseline survey are described in Table 1 and are grouped together by construct. Participants generally complete the same baseline survey, with 2 exceptions. First, there are several specific questionnaires for sexual minority women regarding minority stressors (eg, discrimination and rejection), identity, and LGBT community connection. To balance survey length and content, heterosexual women complete questionnaires regarding their identity as women and their general social support. Second, as we are interested in assessing alcohol use but did not require participants to drink alcohol to participate in this study, we ask about alcohol use only of those women who drank in the past 30 days and administer a series of questionnaires regarding reasons for not drinking and general coping to nondrinkers to similarly balance for time.

The measures in the 4 different EMA surveys (morning, before bed, after binge eating, and prompted) are described in Table 2. Items were either drawn directly from existing EMA scales or past studies or were adapted from non-EMA measures and developed for the purposes of this study in instances where no EMA measure of a construct existed.

The end-of-study survey collects information about participants’ experiences in the study, interest in and willingness to be involved in future mobile health interventions regarding physical and mental health, and COVID-19–related questions. Table 3 provides a description of the constructs included in the end-of-study survey.

EMA data are inherently hierarchical, with surveys (level 1) nested within a person (level 2). As such, we will use multilevel modeling to account for similarities between responses for each person. Before hypothesis testing, the data will be examined for normality, outliers, and nonlinear relationships between continuous variables. Model comparison procedures will be used to determine the fixed versus random effects for each model.

Multilevel modeling is robust to missing data (eg, skipped EMA surveys). For variables skipped within a completed assessment, expectation maximization imputation will be used to replace missing values, an approach demonstrated to minimize bias [122]. Participants with complete data will be compared with those with missing data to identify potential attrition biases, and significant predictors will be included as covariates in all analyses.

To test our main hypotheses, level-1 associations between relevant predictors (negative emotions, social factors, and health behaviors) and outcomes (binge eating and loss of control of eating) will be examined using a series of multilevel models: isolated models with each predictor modeled separately first and then a final model examining all predictors simultaneously to explore unique effects (hypotheses 1a-b). Sexual identity will then be included as a level-2 predictor, and cross-level interactions between sexual orientation (level 2) and general factors (level 1) will indicate whether momentary associations are different across groups (hypothesis 1c). Aim 2 analyses will use only sexual minority women’s data. Level-1 associations between sexual minority stress and eating outcomes will be examined using the same approach as in aim 1: a series of isolated then simultaneous multilevel models (hypothesis 2a). To assess hypothesis 2b, a multilevel structural equation model will be used to examine if negative affect mediates these relationships (1-1-1 model) [123] using Monte Carlo CIs to assess the significance of indirect effects [124]. For aims 1 and 2, we will conduct the aforementioned analyses for momentary associations (random surveys) and event-lagged associations (event and random surveys), allowing for the examination of time-lagged associations (eg, negative affect before the event). Aim 3 analyses will examine race- and person-level body- and eating-related factors (BMI, body dissatisfaction, disordered eating, and thin norm internalization), and sexual minority–specific factors (lifetime discrimination, positive sexual minority identity, and community identification and connection) will be included as level-2 predictors in a series of models. Cross-level interactions will indicate whether person-level factors moderate level-1 associations between general factors and binge eating (moderation of aim 1) or between minority-specific factors and binge eating (moderation of aim 2). Three-way interactions among level-1, level-2 sexual orientation, and other level-2 moderators will indicate whether the differential associations detected in the second component of aim 1 (ie, whether associations are stronger or weaker across sexual identity) are moderated by other person-level factors.

The goal of the HER Life Project is to examine general and sexual minority–specific predictors of disordered eating, in particular binge eating, among sexual minority and heterosexual young women in their daily lives. This study uses an EMA approach to assess participants at fixed times (morning and before bed), at semirandom times throughout the day, and following self-reported binge eating episodes, thus providing rich information regarding the affective, social, behavioral, and minority-specific experiences of these women. Although there is a growing body of literature considering real-world associations among affective processes (eg, negative affect, stress, and body dissatisfaction), social processes (eg, interactions and social comparisons), and binge eating in general samples of young women where sexual identity is not known or reported, it remains unclear whether and how these processes operate for sexual minority women and if there are minority-specific predictors (eg, sexual minority stressors) for these women as well. This study includes samples of both sexual minority and heterosexual young women in an effort to explore potential similarities and differences in binge eating experiences and predictors. We expect that there will be many similar affective, social, and health behavior predictors of binge eating across sexual minority and heterosexual women but that there may also be some differences, particularly with regard to body dissatisfaction and unique sexual minority stressors. The findings from this study can help improve our understanding of potential processes that may contribute to health disparities between sexual minority and heterosexual young women’s disordered eating behaviors and can be used to inform culturally tailored interventions for binge eating for young sexual minority women. In the following sections, we discuss some of the questions we considered and the challenges we have experienced when designing and carrying out the HER Life Project.

The HER Life Project primarily focuses on exploring potential similarities and differences in predictors of binge eating between sexual minority and heterosexual young women given the documented health disparities in obesity rates and binge eating between these groups. However, as described in this protocol, part of aim 3 is to explore potential disparities that Black women may experience and consider how racial and sexual minority identities may intersect for Black sexual minority women. As described previously, there is very limited previous research on disordered eating among people who identify as both racial and sexual minority women [46] and, thus, this is a critically understudied area. Recent reports have highlighted the need for more research using an intersectional perspective and reiterated the importance of such work for providing a more complete understanding of minority health more broadly. Although this study will provide some insight on these topics, it is only a first step, and future work to evaluate the role of other identities (eg, other racial or ethnic group identities, location [urban or rural], and socioeconomic status) and their interaction with sexual orientation is needed to advance the health and well-being of all sexual minority individuals [132]. Continuing to expand and incorporate an intersectional framework will enrich both research and intervention, with ultimate benefits to education, health care, and public policy.

Finally, we view the HER Life Project as 1 step in a large research program aimed at developing culturally sensitive interventions to address binge eating and obesity in sexual minority women. Data from this study can help inform future work in 2 ways. First, as we are including both sexual minority and heterosexual young women, this study will allow us to determine whether there are differential associations between the hypothesized predictors and binge eating or if sexual minority–specific factors (eg, experiencing sexual minority stressors and discrimination) are associated with binge eating. These findings could help in identifying intervention content for disordered eating or body image interventions that require tailoring for sexual minority women. For example, it is plausible that intervention components teaching strategies for coping with sexual minority stress (instead of engaging in binge eating) are needed. The HER Life Project is also uniquely designed to inform the development of ecological momentary interventions (EMIs). Binge eating EMIs that use mobile apps or SMS text messaging can deliver treatment in women’s everyday lives and at specific times when most in need (ie, when likely to engage in binge eating) [133,134]. To develop EMIs, EMA data regarding within-person associations with the target population are first needed [135-137], and this study provides such data. In summary, the HER Life Project will provide foundational information needed to develop culturally tailored treatments and mobile technology–based interventions for sexual minority women with the ultimate goal of reducing binge eating and obesity disparities.