Perceived parental reactions to coming out, attachment, and romantic relationship views

Abstract

Coming out as lesbian, gay, or bisexual (LGB) to one's parents can be a challenging experience and may lead to acceptance or rejection. Attachment theory can help predict parents' reactions to coming out and consequences for romantic attachment. In a cross-sectional study of 309 LGB individuals, we found that those who perceived their mother as accepting in childhood were more likely to have come out to her. Moreover, parents perceived as accepting and independence-encouraging in childhood were reported to react more positively to their child's sexual orientation. Mothers' positive reactions were associated with lower romantic attachment anxiety for men. The links between parent–child relationship quality and optimism and trust in romantic relationships were mediated by romantic attachment patterns. Findings support the contention that LGB pair bonds are attachment relationships, and underline the importance of prior parent–child relationships for predicting LGB individuals' experience of coming out and romantic relationships.

Keywords:

• attachment
• sexual orientation
• coming out
• relationship views
• parent–child relations

Acknowledgements

This research was supported in part by a grant from the Faculty Research Support Fund, University of Southampton. We would like to thank Caroline Gamble for assistance with data collection and thank all of our participants.

Notes

1. Some measures are based on retrospective reports. Unfortunately, it is difficult to examine these issues longitudinally (and collect pre-disclosure measures) due to stigma associated with identifying as LGB (Heatherington & Lavner, 2008). For example, it would be very difficult to recruit a sample of adolescents who had not yet come out and have them report perceived quality of parent–child relationship before and after coming out to parents. Therefore, the retrospective nature of self-reports is a typical limitation in this area (Heatherington & Lavner, 2008).

2. Participants were recruited from organizations for LGB individuals (e.g., Breakout, Gay West, university LGB organizations, Lesbian/Gay Christian Movement, FFLAG [Friends and Family of Lesbians and Gays], Outrage, Outzone, Stonewall), and for HIV positive individuals (e.g., Body Positive, Terrence Higgins Trust). Twenty-two participants were HIV positive.

3. Sample 1 comprised 31 women and 120 men (M _AGE = 32.31, SD = 12.17; 109 out to mother, 95 out to father). Sample 2 comprised 42 women and 61 men (M _AGE = 24.15, SD = 6.92; 91 out to mother, 75 out to father). Sample 3 comprised 35 women and 20 men (M _AGE = 22.85, SD = 3.96; 54 out to mother, 35 out to father). Partly to avoid burdening participants with long questionnaires and thus increase the response rate, and partly because studies involved several investigators who included other variables not the focus of this study, researchers collected a subsample of items. All participants reported demographics, information about coming out, attachment dimensions, parents' reactions to coming out, and parent–child relationship while growing up. In addition, participants in sample 1 reported romantic relationship optimism, whereas those in samples 2 and 3 reported romantic relationship trust and employment status.

4. Supplementary analyses were also conducted with the initial and current reactions in separate models. Some of the reported results were weaker when using only one of the indices, but no clear meaningful differences arose between the two models, and the overall pattern of results was unchanged.

5. Except for the Mother, Father, Peer scale and romantic optimism scale, these measures have been used with LGB samples in prior research.

6. There was no systematic pattern of missing data, except where parental relationship or parental reactions data were missing due to not having grown up with a parent or a parent being deceased. All scales had less than 5% missing data (after accounting for deceased parents and not being out) apart from trust, whereby missing data was slightly higher among participants who were not in a current romantic relationship compared to those who were.

7. Endogenous binary variables do not meet parametric assumptions of SEM. Because being out to mother and father were observed (as opposed to latent) variables, recommendations to use Bayesian estimation or to provide additional constraints to make the model identified (Long, 1997) were not appropriate. Another recommended strategy for binary variables is to use maximum likelihood estimation with bootstrap resampling (Bollen & Stine, 1993). When inspecting the confidence intervals for our paths generated by bootstrapping analysis, all of the significant paths in Figure 2 remained significant. We therefore report parametric path estimates in Figure 2 for clarity and ease of reading.

8. In order to test associations with trust and optimism in the same SEM, we used a stochastic regression method to impute missing values (Little & Rubin, 2002). Although Full Information Maximum Likelihood estimation can be used in AMOS without imputing values, it is not possible to fit the saturated model and thus generate fit indices because no participants completed both the trust and optimism scales. Stochastic regression imputation fits a regression model for each observed variable using maximum likelihood estimation, and then imputes a value for each missing case by drawing at random from the distribution of values based on the nonmissing scores for that case. This random element means that stochastic regression does not artificially increase model fit in the way that standard regression imputation might. Fit indices and path estimates were very similar to those obtained when trust and optimism were examined in separate models with no imputation.

9. Age was controlled for in all analyses and the pattern of results remained the same.