Figures & data
Figure 1. Theoretical model linking marginalization and loneliness.
![Figure 1. Theoretical model linking marginalization and loneliness.](/cms/asset/7f74f679-0d43-458e-8a98-db30063cb0c8/hjsr_a_2339511_f0001_oc.jpg)
Table 1. Demographic characteristics.
Figure 2. Age differences in adjusted means with robust 95% confidence intervals. Note: For everyday discrimination/harassment, microaggressions, family rejection, and internalized homonegativity, 95% CIs may not be symmetric around the mean because they were back-transformed from log-10 values. All group means were adjusted for gender identity, sexual orientation, ethnoracial identity, geographic region, and urbanicity.
![Figure 2. Age differences in adjusted means with robust 95% confidence intervals. Note: For everyday discrimination/harassment, microaggressions, family rejection, and internalized homonegativity, 95% CIs may not be symmetric around the mean because they were back-transformed from log-10 values. All group means were adjusted for gender identity, sexual orientation, ethnoracial identity, geographic region, and urbanicity.](/cms/asset/4c05861b-9e91-4043-929f-8a1222547b53/hjsr_a_2339511_f0002_oc.jpg)
Table 2. Total and direct associations from final partially invariant multigroup Model G.
Figure 3. Gender differences in adjusted means with robust 95% confidence intervals. Notes: CG = cisgender. For everyday discrimination/harassment, microaggressions, family rejection, and internalized homonegativity, confidence intervals may not be symmetric around the mean because they were back-transformed from log-10 values. All group means were adjusted for age (continuous), sexual orientation, ethnoracial identity, geographic region, and urbanicity.
![Figure 3. Gender differences in adjusted means with robust 95% confidence intervals. Notes: CG = cisgender. For everyday discrimination/harassment, microaggressions, family rejection, and internalized homonegativity, confidence intervals may not be symmetric around the mean because they were back-transformed from log-10 values. All group means were adjusted for age (continuous), sexual orientation, ethnoracial identity, geographic region, and urbanicity.](/cms/asset/3fb03626-8494-4f4d-acef-a8bcc3411966/hjsr_a_2339511_f0003_oc.jpg)