Adult Attachment, Differentiation of Self, and Relationship Satisfaction in Lesbians and Gay Men

We examined adult attachment styles, differentiation of self, and relationship satisfaction in a sample of 298 Italian lesbians (48%) and gay men (52%), all of whom were ages 19–71 years (M = 36.1; SD = 11.8) and in a couple relationship for at least six months. Participants were recruited via the Internet and completed a web-based survey. We tested the hypotheses that attachment insecurity and differentiation of self would predict relationship satisfaction, as well as that differentiation of self would mediate the relationship between adult attachment and relationship satisfaction. Results supported the hypotheses, thus indicating that attachment insecurity and differentiation of self were correlated with and predicted relationship satisfaction. Furthermore, results supported the hypothesized mediating role of differentiation of self. Finally, results indicated that younger participants, lesbians, and participants in civil unions reported higher levels of relationship satisfaction. Implications for counselors and therapists working with LG populations are discussed.


Introduction
Since the pioneering research of Hazan and Shaver (1987;Shaver and Hazan 1988), which conceptualized romantic love as an attachment process, several studies have investigated adult romantic attachment. These studies have hypothesized that attachment patterns found in infant-caregiver relationships (e.g., Ainsworth 1989;Bowlby 1969Bowlby /1982 may apply to adult love experiences (e.g., Feeney 2016; Fraley and Roisman 2019; Hainlen et al. 2016;Muşdal Çelebi̇ 2018). Two dimensions have been identified as those most useful for such a conceptualization: avoidance, related to the minimization of attachment needs and which indicates individuals' discomfort in close relationships; and anxiety, related to the hyperactivation of the attachment system (e.g., Mikulincer and Shaver 2016). Moreover, research has shown the relevance of these two dimensions in predicting relationship satisfaction (e.g., Campbell and Stanton 2019;Gleeson and Fitzgerald 2014;Mikulincer et al. 2002;Li and Chan 2012;Sandberg et al. 2017;Vollmann et al. 2019). They do so by generally indicating that high levels of anxiety and avoidance are negatively associated with satisfaction in romantic relationships, while secure attachment is positively associated with the quality of romantic relationships. In particular, research findings indicate that individuals who report higher levels of anxiety and avoidance are characterized by more dysfunctional relationship thoughts, feelings, and behaviors. In turn, this leads individuals to be less satisfied in their couple relationship, regardless of whether they are in a same-sex couple or heterosexual couple (e.g., Roisman et al. 2008;Wright 2020).
Mary Ainsworth was the first to show that same-sex and heterosexual romantic attachments are likely to function in the same way (1985). She also highlighted that the main difference between same-sex and heterosexual couples is that the first are socially sanctioned, while the second are subject to discrimination. Despite the fact that the international literature on attachment in same-sex couples is relatively sparse, several studies have found evidence of the positive association between relationship satisfaction and lower levels of anxiety and avoidance (e.g., Sommantico et al. 2020;Fingerhut and Peplau 2013;Guzmán-González et al. 2019; Mohr and Jackson 2016;Mohr et al. 2013;Starks et al. 2015;Starks and Parsons 2014). Furthermore, research findings have indicated a positive association between attachment insecurity and negative identity, as well as a strong connection between stress-related variables, such as attachment insecurity and internalized sexual stigma (e.g., Sommantico et al. 2020;Fassinger 2003, 2006;Popa-Velea et al. 2019;Šević et al. 2016).
Several researchers (e.g., Hardy and Fisher 2018;Lampis and Cataudella 2019;Ross et al. 2016;Skowron and Dendy 2004), have shown the importance of studying adult attachment together with differentiation of self, due to a shared focus on critical interactional and relational constructs. Differentiation of self is the central concept of Bowen's model (1978) of individual developmental trajectories. It constitutes a fundamental developmental target, which is located in between the need for autonomy and the ability to create and maintain intimate connections with others (e.g., Knerr and Bartle-Haring 2010;Titelman 2014). This construct is comprised of the following: intra-psychic dimensions, such as emotional reactivity and the I-position, conceptualized as self-regulation strategies; and interpersonal dimensions, such as emotional cutoff and fusion with others, conceptualized as interactive regulation strategies (Skowron and Schmitt 2003).
In particular, in analyzing a couple's functioning, partners' differentiation of self appears to be fundamental to the ability to achieve intimacy and mutuality in an intimate relationship (e.g., Gubbins et al. 2010). Indeed, several studies conducted on heterosexual populations have highlighted positive associations between good differentiation of self and couple relationship satisfaction and adjustment (e.g., Gubbins et al. 2010;Işık et al. 2020;Lampis 2016;Peixoto-Freitas et al. 2020;Peleg 2008;Rodríguez-González et al. 2020;Skowron 2000).
Both adult attachment and differentiation of self focus on the balance between intimacy and autonomy, recognizing the central role of emotions within close relationships. They also focus on the continuity of relationship quality across generations, thus creating a dialogue between intra-psychic and relational processes (e.g., Stinson 2016). Furthermore, as indicated above, research on both constructs has indicated significant associations with romantic relationship satisfaction.
Research on the functioning, quality, and satisfaction of couple relationships in lesbian and gay (LG) populations has produced significant findings (e.g., Sommantico et al. 2020;Calvillo et al. 2020;Lampis et al. 2020). Contrary to discriminatory stereotypes, Peplau and Fingerhut (2007) have noted that gay and lesbian couples' relationships did not differ in relationship quality and satisfaction in comparison to heterosexual couples' relationships. Their relationship satisfaction seems to be specifically characterized by a higher similarity between partners in values and attitudes (Kurdek and Schmitt 1986), as well as by higher equality in power and decision-making between partners (Peplau and Spalding 2000). In particular, research findings indicate that, at the initial stage of the couple formation, gay couples have sex more often, while lesbian couples generally have less sex (e.g., Peplau et al. 2004). Furthermore, lesbians tend to be more satisfied with their couple relationship (e.g., Sommantico et al. 2019;Mohr et al. 2013;Ridge and Feeney 1998;Roisman et al. 2008) and more sexually exclusive, while, in samples of gay men, a correlation was not found between sexual exclusivity and relationship satisfaction (e.g., LaSala 2004; Whitton et al. 2015).
In terms of the Italian context, in 2012 the Italian National Institute of Statistics (ISTAT) published data (collected in 2011 from a sample of 59.433.744 respondents) from a survey on attitudes toward the LG population. Results indicated that some important forms of discrimination toward lesbians, gay men, and same-sex couples still exist, despite the majority of respondents condemning discriminatory behaviors toward the LG population and finding it acceptable for a man or woman to have emotional and sexual relationships with the same sex. Lesbians' and gay men's perception of discrimination also remains high, especially in areas ranging from housing, healthcare access, public services, and community relations, to the workplace or school/university environment, and the search for employment.
These data, together with data emerging from other Italian research in the field (e.g., Sommantico et al. 2020;Lingiardi et al. 2005Lingiardi et al. , 2016Scierri and Batini 2020), reveal some important specificities in the Italian context. These include the influence of the Catholic Church, the recent legislation regarding civil unions (Law May 20, 2016, n.76), as well as a "don't ask, don't tell" attitude toward the LG population, all of which together render the Italian context more ambivalent than friendly.
Furthermore, despite the growing social visibility of and acceptance toward lesbians and gay men, Italy is mid-ranking among European countries in terms of acceptance of homosexuality (e.g., Pew Research Center 2013). Individuals in same-sex relationships are still often stigmatized and marginalized, via an increased internalization of society's negative ideologies. In turn, as several studies show (e.g., Sommantico et al. 2018Sommantico et al. , 2019Baiocco et al. 2014;Lorenzi et al. 2015;Petrocchi et al. 2020), internalized sexual stigma strongly interferes with Italian lesbians' and gay men's psychological well-being, as well as with their couple relationship satisfaction.
Taken together, these research findings indicate the relevance of deeply understanding the association between adult attachment and relationship satisfaction in Italian sexual minority populations, taking into account the stated influence of minority stress on adult attachment and relationship satisfaction (e.g., Sommantico et al. 2018;Cao et al. 2017;Cook and Calebs 2016;LeBlanc et al. 2015;Gonçalves et al. 2020;Fassinger 2003, 2006;Popa-Velea et al. 2019;Šević et al. 2016;Wright 2020). We can do so by focusing in particular on the role played by differentiation of self, a variable whose role has never been analyzed in sexual minority populations.

The Present Study: Aims and Hypotheses
The main aim of the present study is to evaluate the significance of associations between adult attachment styles, differentiation of self, and romantic relationship satisfaction in a sample of Italian lesbians and gay men.
Based upon findings from previous studies (e.g., Sommantico et al. 2020;Calvillo et al. 2020;Guzmán-González et al. 2019;Lampis et al. 2019a, b;Mohr et al. 2013;Roisman et al. 2008; Parsons 2014), we tested the following hypotheses: (H1) attachment insecurity is negatively correlated with differentiation of self and relationship satisfaction; (H2) differentiation of self is positively correlated with relationship satisfaction; (H3) attachment insecurity and differentiation of self predict relationship satisfaction; (H4) differentiation of self mediates the relationship between attachment insecurity and relationship satisfaction; and (H5) lesbians and participants in civil unions report higher levels of relationship satisfaction.

Procedure and Participants
Participants were recruited online by advertisements placed on social media (LGBT Associations, LGBT listservs, LGBT discussion boards, and LGBT research centers), according to the following criteria: (i) participants must identify as lesbians or gay men; (ii) they must be over the age of 18 years; and (iii) they must have been in a stable relationship for at least six months. No remunerative rewards were given. This web-based sampling is one of the most-utilized sampling strategies for LGBT psychological research, in particular for the study of theoretically-driven hypotheses. It is implemented for its ability to reach dispersed and otherwise hard-to-contact individuals (Binson et al. 2007).
To improve upon this community-based sampling, we also used snowball sampling (Meyer and Wilson 2009). We asked participants who were initially recruited to identify other potential respondents from their social network who, in turn, were asked to nominate individuals from their own social network, and so forth.
Data were collected through self-report instruments using an Internet-based survey (Riggle et al. 2005). On the first page, participants were informed about the purpose, rationale, and procedure of the study, and they were asked to give informed consent to participate in the survey, which took approximately 25 min to complete. A basic demographic questionnaire was completed on the second page, collecting information regarding: age, gender, relationship status (partnered/cohabiting/in a civil union), relationship length, and level of education. The following pages of the survey consisted of a presentation of three instruments: (a) Experiences in Close Relationships Revised (ECR-R; Fraley et al. 2000); (b) Differentiation of Self Inventory Revised (DSI-R; Skowron and Friedlander 1998;Skowron and Schmitt 2003); and (c) Gay and Lesbian Relationship Satisfaction Scale (GLRSS; Belous and Wampler 2016;Belous et al. 2020). For a detailed description of the measures, see the "Measures" section.
The sample consisted of 298 participants (lesbians = 48%; gay men = 52%). They had been in a stable relationship for an average of 82.1 months (SD = 82.4), and their mean age was 36.1 years (SD = 11.8). Cohabiting participants were 45.3%, and 19.1% of the participants had been in a civil union. The sample had a high educational level, with 57.3% of the participants having completed a university degree or a post-university degree, and 38.2% having completed secondary school.

Measures
A basic demographic questionnaire collected information regarding age, gender, relationship status (civil union vs. relationship/cohabitation), relationship length, and level of education.
The Experiences in Close Relationships Revised (ECR-R) (Fraley et al. 2000;Italian validation: Picardi et al. 2000) is a 36-item self-report questionnaire which provides a dimensional evaluation of current attachment strategies in romantic relationships on two subscales, which correspond to the two main dimensions of romantic attachment: (1) Avoidance (AVO-18 odd-numbered items); and (2) Anxiety (ANX-18 even-numbered items). The questionnaire was completed by expressing feelings regarding the couple relationship according to a Likert-type scale ranging from 1 ("Totally false") to 7 ("Totally true"). Low scores on both dimensions are indicative of relatively secure attachment. In our study, Cronbach's α was 0.90 for AVO and 0.84 for ANX.
The Differentiation of Self Inventory Revised (DSI-R; Skowron and Friedlander 1998;Skowron and Schmitt 2003;Italian validation: Lampis et al. 2017a, b) is a 46-item selfreport questionnaire which measures intra-psychic and interpersonal aspects of differentiation of self on four subscales: (1) Emotional Reactivity (ER-11 items); (2) I-Position (IP -11 items); (3) Emotional Cutoff (EC -12 items); and (4) Fusion with Others (FO-12 items). The questionnaire was completed by expressing agreement according to a Likert-type scale ranging from 1 ("Totally disagree") to 6 ("Totally agree"). Higher scores on the four dimensions indicate greater differentiation of self. In our study, Cronbach's α was 0.79 for ER, 0.86 for IP, 0.78 for EC, 0.73 for FO, and 0.85 for the total score.
The Gay and Lesbian Relationship Satisfaction Scale (GLRSS) (Belous and Wampler 2016;Belous et al., 2020;Italian validation: Sommantico et al. 2019) is a 24-item selfreport questionnaire which measures relationship satisfaction in lesbian and gay respondents on two subscales: (1) Relationship Satisfaction (RS-16 items); and (2) Social Support (SS-8 items). The questionnaire was completed by expressing agreement according to a Likert-type scale ranging from 1 ("Strongly disagree") to 6 ("Strongly agree"). Scores above the mean indicate higher levels of couple relationship satisfaction/support, while lower scores indicate possible deficits in those two areas. In our study, Cronbach's α was 0.79 for RS, 0.80 for SS, and 0.79 for the total score.

Statistical Analyses
All statistical analyses were carried out using SPSS 23.0 (IBM Corp. Released 2015). The reliability analysis was computed using Cronbach's α, and was considered to be satisfactory if the values were greater than 0.70 (Nunnally and Bernstein 1995). Correlations analyses were conducted by means of Pearson's coefficient (r; between 0.10 and 0.29 = small association; between 0.30 and 0.49 = medium association; and > 0.50 = large association; p-value < 0.05). Group differences were verified through ANOVA (p-value < 0.05). Effect sizes were measured through Etasquare (η 2 ; small ≥ 0.01; medium ≥ 0.059; large ≥ 0.138) (Cohen 1988). Multiple regression analyses were conducted, using standardized β coefficients and R 2 coefficients (p < 0.05), to determine the contribution of each predictive variable to the regression model. For mediation analyses, both direct and indirect effects were examined using bootstrapping methods in order to estimate bias-corrected asymmetric confidence intervals (CIs) with 5000 resamples with replacement, through the PROCESS macro tool for SPSS (Hayes 2018). A CI not inclusive of zero indicates significant effect (Hayes and Rockwood 2017).
Ethical Standards.
The study complied with the American Psychological Association (APA) ethical standards in the treatment of human research participants and conformed to the provisions of the Declaration of Helsinki in 1995 (as revised in Edinburgh 2000). Furthermore, the study was approved by the Ethical Committee of Psychological Research of the Department of Humanities of the University of Naples Federico II.

Descriptive Statistics, Correlations, and Group Differences
Means, standard deviations, and Cronbach's α for the instruments are presented in Table 1. The means of avoidance and anxiety were, respectively 2.0 (SD = 0.9) and 3.6 (SD = 1.0). The mean of differentiation of self was 3.9 (SD = 1.0), and the mean of relationship satisfaction was 97.8 (SD = 16.4).
Zero-order correlations between the demographic variables and the instruments, as well as among the instruments, are shown in Table 2. Results indicated small to large 1 3 significant associations: age was significantly positively correlated with relationship length and avoidance (r respectively 0.56 and 0.21; p < 0.01). Age was also significantly negatively correlated with anxiety, relationship satisfaction, and the GLRSS total score [(r respectively − 0.27, − 0.24 (p < 0.01), and − 0.12 (p < 0.05)]. Relationship length was significantly positively correlated with avoidance and social support (r, respectively, 0.17 and 0.23; p < 0.01). It was also significantly negatively correlated with anxiety and relationship satisfaction (r, respectively, − 0.17 and − 0.16; p < 0.01). Avoidance was significantly negatively correlated with the DSI-R subscales and total score, as well as with the GLRSS subscales and total score (with r ranging from − 0.16 to − 0.59; p < 0.01). Anxiety was significantly negatively correlated with the DSI-R subscales and total score, as well as with the GLRSS subscales and total score (with r ranging from − 0.19 to − 0.33; p < 0.01). Finally, the DSI-R subscales and total score were significantly positively correlated with the GLRSS subscales and total score [with r ranging from 0.12 (p < 0.05) to 0.71 (p < 0.01)]. These findings support Hypotheses 1 and 2, indicating the following: participants with more insecure attachment reported lower levels of differentiation of self and relationship satisfaction, and participants with higher levels of differentiation of self also reported higher levels of relationship satisfaction.
Regarding relationship status, the ANOVA omnibus test and Tukey post-hoc tests showed small to large statistically significant differences. Participants in civil unions reported lower levels of anxiety (F(2, 295) = 6.49, p < 0.01; These findings support Hypothesis 5, thus indicating that lesbians and participants in civil unions reported lower levels of attachment insecurity, as well as higher levels of differentiation of self and relationship satisfaction. Legal recognition of couples' relationships has positive effects in terms of greater differentiation of self and relationship satisfaction, as well as in terms of less attachment insecurity.

Regression Analysis
Based on previous results, a hierarchical multiple regression analysis was conducted in order to determine the extent to which each predictor variable contributed to the model predicting lesbians' and gay men's relationship satisfaction above and beyond the others (see Table 3). After controlling for differences in age and gender, the addition of relationship status contributed significantly to the regression model by increasing the accounted variance by 11.9% (β = 0.370; R 2 = 0.157, p < 0.01), indicating that being in a civil union predicts higher relationship satisfaction. The addition of avoidance contributed significantly to the regression model by increasing the accounted variance by 20.6% (β = − 0-0.474; R 2 = 0.363, p < 0.001), indicating that lower levels of avoidance predict higher relationship satisfaction. The addition of anxiety contributed significantly to the regression model by increasing the accounted variance by 7.9% (β = − 0.297; R 2 = 0.442, p < 0.001), indicating that lower levels of anxiety predict higher relationship satisfaction. Finally, the addition of differentiation of self contributed significantly to the regression model by increasing the accounted variance by 33.9% (β = 0.725; R 2 = 0.781, p < 0.001), indicating that higher levels of differentiation of self predict higher relationship satisfaction. Age, gender, relationship status, adult attachment styles, and differentiation of self, combined, accounted for 78.1% of the variance in relationship satisfaction, and 77.7% of the variance when adjusted for sample size and the number of predictors.
We also conducted hierarchical regression analyses, separately for lesbians and gay men, in order to determine, in both subsamples, the extent to which each predictor variable contributed to the overall model predicting lesbians' and gay men's relationship satisfaction above and beyond the others (see Table 4).
For the lesbian subsample, age, relationship status, adult attachment styles, and differentiation of self, combined, accounted for 75.2% of the variance in relationship satisfaction, and 74.3% of the variance when adjusted for sample size and the number of predictors. For the subsample of gay men, age, relationship status, adult attachment styles, and differentiation of self, combined, accounted for 81.7% of the variance in relationship satisfaction, and 81.1% of the variance when adjusted for sample size and the number of predictors.
These findings strongly support Hypotheses 3 and 5, indicating that, in the total sample, as well as in the two subsamples, civil union status, lower levels of adult attachment styles, and higher levels of differentiation of self positively influence perceived relationship satisfaction, and predict its scores.

Mediation Analysis
Based on previous results, we explored the direct and indirect effects of adult attachment styles on romantic relationship satisfaction, through the variable of differentiation of self. We found direct and indirect effects, as reported in Tables 5 and 6.
These findings support Hypothesis 4, thus indicating that adult attachment styles are associated with relationship satisfaction and differentiation of self. Moreover, the negative coefficients between adult attachment and relationship Table 3 Hierarchical regression analysis for total sample (N = 298) DSI-R DSI-R total score; GLRSS = GLRSS total score * p < 0.01; **p < 0.001 Step and predictor variable β R 2 ΔR 2 satisfaction (avoidance = − 2.83; anxiety = − 0.99), as well as between adult attachment and differentiation of self (avoidance = − 0.53; anxiety = − 0.32), support Hypotheses 1 and 2, indicating that participants who report higher levels of attachment insecurity show lower relationship satisfaction and differentiation of self.

Discussion
In line with previous studies (e.g., Sommantico et al. 2019Sommantico et al. , 2020Mohr and Fassinger 2006;Mohr et al. 2013;Roisman et al. 2008), our findings indicate that lesbian participants showed lower levels of anxious attachment styles than gay men, while also reporting higher levels of relationship satisfaction. We can hypothesize that, especially in the Italian context, which is characterized by strong sexism and by an ambivalent attitude toward the LGBT population (e.g., Lingiardi et al. 2005Lingiardi et al. , 2016Petrocchi et al. 2020;Scierri and Batini 2020), it is possible that lesbians are less exposed than gay men to sexual stigma, which more negatively influence gay men's well-being and relationship satisfaction (e.g., Sommantico et al. 2018Sommantico et al. , 2020Cao et al. 2017;Gonçalves et al. 2020;LeBlanc et al. 2015). Contrary to previous studies (e.g., Stearns and Sabini 1997), younger people reported higher levels of relationship satisfaction in our sample. We can hypothesize that, in the Italian context, young lesbians and gay men are less influenced by the historical social stigma and minority stress that greatly affect older lesbians and gay men, thereby accounting for their higher levels of relationship satisfaction.
Regarding relationship length, consistent with previous studies and reviews (e.g., Kurdek 1998;Peplau and Fingerhut 2007), our results indicate that the longest-lasting relationships are characterized by higher levels of perceived social support, as well as that the early stages of cohabitation are those in which there is more relationship satisfaction.
Previous research on LG individuals and couples have often shown the significant effect of relationship status on romantic relationship satisfaction (e.g., Sommantico et al. 2018Sommantico et al. , 2020Riggle et al. 2017;Todosijevic et al. 2005). Our results confirm the hypothesis that civil unions have a positive effect on relationship satisfaction. Indeed, legitimization of same-sex unions seems to strengthen the couple relationship, as well as increasing relationship satisfaction, particularly due to broader support networks and greater visibility. In this light, the result that indicates that participants in civil unions perceived significantly higher social support is of special interest.
In line with the literature's findings on adult attachment and same-sex relationships' well-being (e.g., Sommantico et al. 2020;Fingerhut and Peplau 2013;Guzmán-González et al. 2019;Mohr et al. 2013;Popa-Velea et al. 2019; Starks and Parsons 2014;Wright 2020), the results of our study showed a positive correlation between attachment insecurity and lower levels of relationship satisfaction. In fact, participants with insecure attachment patterns showed lower levels of relationship satisfaction, thus confirming the hypothesis regarding the important role of the attachment system in relationship satisfaction. Previous investigations of heterosexual individuals and couples (e.g., Hardy and Fisher 2018;Lampis and Cataudella 2019;Ross et al. 2016;Skowron and Dendy 2004) have shown a positive association between attachment insecurity and lower levels of differentiation of self. Our results confirmed this association in lesbians and gay men, indicating that lesbians and gay men with insecure attachment patterns show lower levels of differentiation of self.
In line with research findings on heterosexual populations (e.g., Gubbins et al. 2010;Işık et al. 2020;Knerr and Bartle-Haring 2010;Lampis 2016;Lampis et al. 2017a, b;Lampis et al. 2019a, b;Peixoto-Freitas et al. 2020;Peleg 2008;Rodríguez-González et al. 2020;Skowron 2000), the results also confirmed the positive association between differentiation of self and relationship satisfaction in lesbians and gay men. Indeed, participants with higher levels of differentiation of self also showed higher levels of relationship satisfaction.
Finally, our results confirmed the hypothesis that differentiation of self, in its intra-psychic and interpersonal dimensions, plays an important mediating role between attachment insecurity and lesbians' and gay men's relationship satisfaction.

Strengths and Limitations
This study is the first in the Italian context to explore adult attachment styles, differentiation of self, and romantic relationship satisfaction in lesbians and gay men, thus providing data to add to the literature on the same variables among heterosexual individuals. By not simply assuming similarity between lesbians and gay men, the study also has the strength of distinguishing and comparing the two groups.
The first general limitation of the study is related to sampling strategy. The literature suggests that the LG population is difficult to define conceptually, as well as to reach. This is probably due to the minority stress that may impede lesbians' and gay men's disclosure to researchers and, thus, their participation in studies. In this way, online surveys become an important and convenient method of collecting data from LG populations, although often with not very large samples (e.g., Hartwell et al. 2017). Furthermore, sampling in the LG community limits the generalization of results. Indeed, community-based sampling and snowball sampling imply specific possible biases. One bias in particular concerns the special characteristics of individuals who voluntarily participate in a study, and who therefore may differ significantly from other populations in terms of adult attachment styles, differentiation of self, and romantic relationship satisfaction. Another possible bias regards the mono-method. Assessing all variables by using self-report questionnaires could have led to inflation in observed associations. To address this, future investigations could integrate quantitative data with qualitative data. Finally, we can imagine future longitudinal research designs, which would also allow for causal inferences to be carried out.

Conclusions
The results of our study contribute to understanding the implications of adult attachment styles, differentiation of self, and relationship satisfaction in lesbians and gay men. It seems that lesbians and gay men who are more securely attached report higher differentiation of self, as well as higher relationship satisfaction. Furthermore, our results highlighted the indirect effect of attachment insecurity on relationship satisfaction, via differentiation of self.
In conclusion, this study provides important evidence about the role of adult attachment styles, differentiation of self, and romantic relationship satisfaction for lesbians and gay men. This is useful information for counselors and therapists working with LG populations, who must be aware of the role of these variables on LG individuals' and couples' well-being. They must also take into account the differences between lesbian and gay populations, as well as differences dependent upon couples' particular stage in the course of life.
Author contributions All authors contributed in the same way to conceive the study and its design, to draft the manuscript, and to interpret the data. All authors read and approved the final manuscript.
Funding Open Access funding provided by Università degli Studi di Napoli Federico II.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.

Ethical approval
The study complied with the American Psychological Association (APA) ethical standards in the treatment of human research participants and conformed to the provisions of the 1964 Helsinki declaration and its later amendments. Furthermore, the study was approved by the Ethical Committee of Psychological Research of the Department of Humanities of the University of Naples Federico II.
Informed consent Informed consent was obtained from all individual participants included in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.