“Shame, Doubt and Sadness”: A Qualitative Investigation of the Experience of Self-Stigma in Adolescents with Diverse Sexual Orientations

Round 1

Reviewer 1 Report

I believe that this article is very relevant to the field of study of diverse sexuality in young people and adolescents. The methodology is congruent and the results make important contributions to better understand how young people experience the stigma of being members of the LGBTIQ community.

In order to enrich the critical apparatus of this research and broaden its analytical perspective, I suggest that the following Latin American research be reviewed.

https://www.redalyc.org/pdf/2933/293330166006.pdf

https://www.redalyc.org/pdf/363/36329481026.pdf

https://editorial.upnvirtual.edu.mx/index.php/publicaciones/colecciones/horizontes-educativos/389-a-look-into-masculine-identity-in-mexican-young-men

https://www.redalyc.org/journal/4763/476368270013/html/

https://www.redalyc.org/journal/4769/476960345007/476960345007.pdf

https://www.redalyc.org/jatsRepo/112/11252977009/html/index.html

https://www.redalyc.org/journal/815/81554612008/html/

https://www.redalyc.org/pdf/325/32550024003.pdf

https://www.redalyc.org/pdf/2933/293345349015.pdf

Author Response

Point 1: I believe that this article is very relevant to the field of study of diverse sexuality in young people and adolescents. The methodology is congruent and the results make important contributions to better understand how young people experience the stigma of being members of the LGBTIQ community.

In order to enrich the critical apparatus of this research and broaden its analytical perspective, I suggest that the following Latin American research be reviewed.

Response 1: Thank you to this reviewer for your kind response and review of our manuscript. We greatly appreciate the suggestion of this research to review, however, we note that these references are in Spanish without an available English translation. As such, we are unable to meaningfully read and reference this literature in the current paper. If this is a concern we are able to perhaps note that broader non-English literature exists, but was not incorporated, in the limitations section.

Reviewer 2 Report

The first sentence is an overgeneralization, particularly since the authors note later that stigmatization is context-dependent. 

I would recommend adding some discussion of the role of structural stress and violence into the introduction, since these may also play a role in the development of stigmatized identities. 

The term "sexuality-diverse" is awkward and not commonly used--suggest revising to avoid it. Sexuality and sexual orientation are not the same thing and should not be conflated this way. The interview questions suggest that the participants understood the distinction and so the waters should not be muddied otherwise. 

It would be helpful to have a bit more context regarding the participants, since they frequently refer to their school and home situations. 

Author Response

Thank you overall for your very helpful and concise feedback on our manuscript. Please see below for our responses to your specific suggestions.

Point 1: The first sentence is an overgeneralization, particularly since the authors note later that stigmatization is context-dependent. 

Response 1:  Thank you for drawing our attention to this. On reflection we’ve decided to remove the sentence entirely.

Point 2: I would recommend adding some discussion of the role of structural stress and violence into the introduction, since these may also play a role in the development of stigmatized identities. 

Response 2: This is an excellent point. We’re mindful of the length of the introduction already, but have added the following sentence in the first paragraph to try to include some discussion of this early on:

Stigma is theorised to affect people with diverse sexual orientations in a variety of ways, including direct experiences of prejudice, structural discrimination, and abuse, but also more indirect problems such as leading individuals with diverse sexual orientations to expect rejection from their peers and leading them to conceal their sexual orientation from others, resulting in stress and distress (Meyer, 2003).

We received feedback from another reviewer to shorten the introduction, and so made this addition quite brief. We are open to expanding on this topic if necessary.

Point 3: The term "sexuality-diverse" is awkward and not commonly used--suggest revising to avoid it. Sexuality and sexual orientation are not the same thing and should not be conflated this way. The interview questions suggest that the participants understood the distinction and so the waters should not be muddied otherwise. 

Response 3:  Thank you for alerting us to this. This may be a cultural difference, as in our country of origin these terms are largely interchangeable, and “sexuality-diverse” is the recommended term by our most prominent youth mental health organisation. The term was agreed on based on feedback from young LGBTQ+ people as well. However, to make the research appropriate for an international audience, we’ve shifted the terminology to “diverse sexual orientations” instead which we hope is an appropriate compromise. We would prefer to avoid the term “sexual minority” if possible, as in our experience this term itself can be felt as stigmatising and othering. We are open to further discussion on this if needed.

Point 4: It would be helpful to have a bit more context regarding the participants, since they frequently refer to their school and home situations. 

Response 4: In hindsight we agree that this could have been helpful, however, we did not systematically collect demographic information other than the young person’s age, gender, and sexual orientation in the interests of helping participants remain as anonymous and safe to participate as possible. We have added a sentence stating that all but one participant was enrolled in high school, as this was previously not included. Other contextual information did arise during the interviews where participants chose to share it anyway, however, we are unable to provide additional demographic information in the form of descriptive statistics other than what is already provided.

Reviewer 3 Report

Thank you so much for the opportunity to review your manuscript! It was a pleasure to learn about self-stigma from the perspectives of sexual orientation minority youth. I hope this review gives some direction and feedback towards this paper becoming a meaningful contribution in the literature an in advocacy of youth wellness. 

This paper addresses self-stigma, a construct that is not commonly represented in the literature. While researchers provide insight to this construct via qualitative thematic analysis, I have general concerns that I would recommend addressing.

- Introduction is long, difficult to follow, and lacks a persuasive argument warranting this study. 

- Thematic analysis is not described as to why it is the enacted methodology, and there are several pitfalls that limit the effective implementation of this methodology (e.g., the inclusion of deductive insight, use of text messaging apps to inform analysis, and lack of explanation of methods used within this technique).

Semi-structured interviews were leading, and codes aligned with the way questions were asked... limiting the inductive insight generated from this paper. 

Themes also contradict themselves (having a code that discusses acceptance followed by one that discusses the emotional experience of self-invalidation and shame, that does not discuss or acknowledge resilience). 

Discussion lacks insight gained from qualitative findings and has limited interaction with sources from the introduction. 

While I recommend these broad items, please see my more in depth review attached here. Thank you for the opportunity to review this manuscript!

Comments for author File: Comments.pdf

Author Response

Point 1: Thank you so much for the opportunity to review your manuscript! It was a pleasure to learn about self-stigma from the perspectives of sexual orientation minority youth. I hope this review gives some direction and feedback towards this paper becoming a meaningful contribution in the literature an in advocacy of youth wellness. 

Response 1: Thank you for the time taken with your very comprehensive and thoughtful feedback. We greatly appreciate the opportunity to improve the paper, and have taken as many opportunities to do so as we can within the scope of what we can change. We have implemented changes based on almost all of the feedback from the attachment provided, but as most of it fell within the scope of the summary points, we have primarily responded to these in our responses below. However, we have also responded to some specific comments from the attachment that we either have not implemented, or on which we would appreciate greater clarity.

Point 2:  Introduction is long, difficult to follow, and lacks a persuasive argument warranting this study.

Response 2: Thank you for your specific suggestions about how to address this. We agree that the provided introduction could be long and difficult to follow, and we hope that having addressed this, the argument for the study is clearer as well. For example, some of the changes we have made seeking to address this include:

• We have shortened the sections at lines 53-68, and 69-70, as suggested.
• We have reordered the flow of information to introduce internalised homophobia first, and the need to focus on self-stigma specifically second. We have also introduced some headings per this suggestion.

Point 3: Thematic analysis is not described as to why it is the enacted methodology, and there are several pitfalls that limit the effective implementation of this methodology (e.g., the inclusion of deductive insight, use of text messaging apps to inform analysis, and lack of explanation of methods used within this technique).

Response 3: Thank you for raising these concerns. As there’s a lot to cover here, we’ll address each specifically (with some attention to the more specific points in your attachment as well):

• Thematic analysis

Thank you for pointing out that we did not include reasoning behind our selection of thematic analysis. We have now added a brief rationale of our choice of thematic analysis in the Introduction and Method sections.

• Inclusion of deductive insight

We understand your concern here around the inductive use of thematic analysis. In this study, we were guided by thematic analysis deductively as supported and described by Braun & Clarke (2006).  We’ve provided the specific relevant part of this paper below:

Themes or patterns within data can be identified in one of two primary ways in thematic analysis: in an inductive or ‘bottom up’ way (eg, Frith and Gleeson, 2004), or in a theoretical or deductive or ‘top down’ way (eg, Boyatzis, 1998; Hayes, 1997). An inductive approach means the themes identified are strongly linked to the data themselves (Patton, 1990) (as such, this form of thematic analysis bears some similarity to grounded theory). In this approach, if the data have been collected specifically for the research (eg, via interview or focus group), the themes identified may bear little relation to the specific questions that were asked of the participants. They would also not be driven by the researcher’s theoretical interest in the area or topic. Inductive analysis is therefore a process of coding the data without trying to fit it into a pre-existing coding frame, or the researcher’s analytic preconceptions. In this sense, this form of thematic analysis is data-driven. However, it is important to note, as we discussed earlier, that researchers cannot free themselves of their theoretical and epistemological commitments, and data are not coded in an epistemological vacuum.

In contrast, a ‘theoretical’ thematic analysis would tend to be driven by the researcher’s theoretical or analytic interest in the area, and is thus more explicitly analyst driven. This form of thematic analysis tends to provide less a rich description of the data overall, and more a detailed analysis of some aspect of the data. Additionally, the choice between inductive and theoretical maps onto how and why you are coding the data. You can either code for a quite specific research question (which maps onto the more theoretical approach) or the specific research question can evolve through the coding process (which maps onto the inductive approach).

For example, if a researcher was interested in talk about heterosex, and had collected interview data, with an inductive approach they would read and re-read the data for any themes related to heterosex, and code diversely, without paying attention to the themes that previous research on the topic might have identified. For example, the researcher would not look to the influential research of Hollway (1989), identifying discourses of heterosex, and code just for male sexual drive, have/hold or permissive discourse themes. In contrast, with a theoretical approach, the researcher may well be interested in the way permissiveness plays out across the data, and focus on that particular feature in coding the data. This would then result in a number of themes around permissiveness, which may include, speak to, or expand on something approximating Hollway’s original theme.

We have bolded the final sentence specifically as we feel that it speaks especially well to what we have done in our study, in seeking to understand how a previously understood construct plays out in a different population (with a result that speaks to the original definition in some ways and expands on it in others).

Essentially, though, taking the prior definition of self-stigma to the data and understanding it within that lens was always intended to be the case. To try to improve clarity on this, we have termed our method of analysis ‘theoretical’ thematic analysis and expanded on our rationale for this choice as above.

In terms of why we included a deductive component (theoretical thematic analysis) in this study specifically, we have attempted to address this already in the following paragraph:

Further, given both the first and last author’s immersion in the stigma and self-stigma literatures, we inevitably and intentionally approached the analysis with a pre-conceived broad definition of self-stigma, as well as the assumption that both stigma and self-stigma would be experienced in some fashion. We acknowledge that significant aspects of the analytic process were therefore theoretically and deductively informed. In particular, the interview protocol was intrinsically developed with this understanding of these constructs in mind. In keeping with the aim of developing a new understanding of how self-stigma is experienced in this group, however, themes in the data with regard to the experience of self-stigma and its content for this group were generated inductively, without regard for previous research and thought on the topic in this particular population.

In our case, we believe that this was unavoidable, given that self-stigma itself is a well-established construct in other populations and one with which we were already quite familiar. Still, we did create themes that were different from existing research on self-stigma, so our approach was limited to a theoretical thematic analysis in keeping with the above description. We chose to be quite explicit about the lack of a truly inductive component of our analysis here as a result, as we felt that our preconceived understanding would affect the outcome regardless (as we also explain in the manuscript).

• Use of text messaging apps

This is also difficult to answer. We would generally disagree that a conversation can’t be adequately held via text messaging app, particularly as among the target population, doing so would be an extremely common method of communication (if not their primary method; e.g., see Lenhart, 2012; Lenhart, Ling, Campbell, & Purcell, 2010).

It was also not our experience that the text messaging data were less rich than those derived from voice interviews. Instead what we observed was that while WhatsApp interviews were briefer, this tended to be more due to the absence of speech lacking in rich meaning (given that participants could more easily parse out their specific response to the question being asked). This is consistent with what Shapka et al. 2016 describe in their analysis of this topic. We also did not observe a clear difference in the themes provided by the two interview types, as we note in the paper.

Overall, we would therefore find it difficult to justify the use of this method being a problem for data quality with this age group. We do acknowledge that this is relatively unfamiliar territory, however, as you note, we felt that this was justified by the need for participant safety, and potential bias in participation were we not to do so.

Lenhart, A. (2012). Teens, smartphones, and texting. Retrieved March 22, 2012, from http://pewinternet. org/~/media//Files/Reports/2012/PIP_Teens_Smartphones_and_Texting.pdf

Shapka, J. D., Domene, J. F., Khan, S., & Yang, L. M. (2016). Online versus in-person interviews with adolescents: An exploration of data equivalence. Computers in human behavior, 58, 361-367.

• Lack of explanation of methods used

Thank you. We have described our methods of analysis in lines 198-212 in the updated manuscript, and this explanation is in line with the methodology of Braun & Clarke (2006). If this is insufficient, we would greatly appreciate any specific recommendations about what is missing.

Point 4: Semi-structured interviews were leading, and codes aligned with the way questions were asked... limiting the inductive insight generated from this paper. 

Response 4: We have addressed this concern to some extent above; the understanding of the definition of self-stigma as an overall construct was never intended to be understood inductively, and we therefore would also argue that providing our definition of it was not inappropriate. We note as well that the specific question the reviewer highlights leaves the participant the ability to disagree before exploring openly, which some participants did (and which led to our discussion throughout the paper about stigma being buffered against by acceptance, for example). We believe that this contributes insight into not only how self-stigma is experienced by those who experience it, but also how it is that some people did not experience it.

As to why we asked about self-stigma in the way we employed, we note that this interview was developed in consultation with a youth reference group comprised of young people with diverse sexual orientations, who suggested that because self-stigma was a complex topic that could not easily be tapped and was experienced by people at different times (not everyone around realisation, and in some cases not at all), this was what they endorsed. It was their opinion that speaking to them about the topic without providing that scaffolding would be very unlikely to get information within the scope of what we were interested in within the process of the theoretical thematic analysis.

Finally, another purpose to asking the question in this way was to be validating and to help them feel safe to answer, recognising as well from our consultation with the reference group that this topic is a difficult and sensitive one. It is common in clinical psychology to preface a question about such topics with a statement that is normalising and validating. Other published and peer-reviewed qualitative research on the topic of self-stigma has used similar questions (Eaton et al., 2016).

Eaton, K., Ohan, J. L., Stritzke, W. G., & Corrigan, P. W. (2016). Failing to meet the good parent ideal: Self-stigma in parents of children with mental health disorders. Journal of Child and Family Studies, 25(10), 3109-3123.

Point 5: Themes also contradict themselves (having a code that discusses acceptance followed by one that discusses the emotional experience of self-invalidation and shame, that does not discuss or acknowledge resilience). 

Response 5: We aren’t entirely sure what is meant by this, but if we are interpreting this point correctly, we would disagree that the themes contradict themselves, given that we note the following:

Overall, self-stigma occurred on continua, varying between individuals and within individuals over time. There was rarely a simple, binary divide between self-stigma and self-acceptance, and many individuals occupied space in between, expressing both at different times.

There was also a specific comment in the reviewer’s attachment about the inclusion of the following under Theme 2, which may also be what’s being referenced here:

As a contrast to this theme, although acceptance from others seemed to be a crucial factor in developing self-acceptance, some individuals also highlighted their own agency in rejecting stigma:

“At first it really bothered me but over years of just soul searching I realised that I’m not a monster others make me out to be and I had to build a thick skin that is a part of me now… I never had the support of others in fighting the stigma put against me, so I had to do it by myself”.

This was included because Braun & Clarke (2006) suggest highlighting contrasting information when discussing themes, and because this particular concept itself was not substantiated in the data sufficiently to provide its own rich theme. While contrasting to the rest of the information under the theme, we don’t feel that it is contradictory- as what it speaks to is what happens when acceptance is absent. That is, acceptance is a precursor to self-acceptance but clearly not the only one, and to exclude this information would arguably make the theme less rich. Future research might also contribute more to these ideas.

Point 6: Discussion lacks insight gained from qualitative findings and has limited interaction with sources from the introduction. 

Response 6: We’ve made some edits based on this feedback, including moving some of the text around as suggested. The general topics outlined in the introduction (the need for mental health intervention strategies, the definitions of internalized homophobia and self-stigma, and the experience and content of self-stigma) are all mirrored in the discussion section. If our current edits are insufficient, we would greatly appreciated more specific feedback about where else this concern could be improved.

Point 7: Please do a spell check for stigmatised and internalised. I believe these are spelled with Zs not Ss

Response 7: Thank you for indicating this. We use ‘s’ instead of ‘z’ as is typical for written English spelling convention outside of North America. We have now changed our spelling to American English.

Point 8: Lines 157: Below 18 is assent, not consent. Cite source that recommends waived consent practices for sexual minority youth.

Response 8: Thank you for raising this concern. We have added a source supporting waiving consent for youth with diverse sexual orientations. Given that parental permission was not sought, our understanding is that referring to this as consent rather than assent is appropriate, however, we are able to revise if this is incorrect.

Point 9: There is no rationalization of thematic data. Needs to be included.

Response 9: Thank you for this feedback, however, we’re not entirely sure what is meant by ‘rationalization’ here. We may have already addressed this under Response 3, but thought it best to seek more clarity on this in case the reviewer was referring to something else.

Point 10: “self-shame or self-invalidation”. These are two different terms. Should these be two subthemes?

Response 10: Thank you for querying this. Our justification for making these two subthemes rather than separate themes was that although the contents were different, the process was the same. That is, they were ultimately two expressions of self-stigma rather than separate constructs. We are open to revising such that they are presented as separate themes.

Point 11: Why are contexts discussed if they are not one of the themes? Include in themes or remove.

Response 11: Thank you for raising this concern, as we were unaware of convention that spoke against this. We provided the contexts as it is information that arose that was not directly related to self-stigma itself, but that our reference group endorsed was crucial to understanding the themes that follow. We are open to revising this; however, we are unaware of guidelines against this and our perspective is that it ultimately enriches the results and we are unsure of how it would compromise the overall quality.

Point 12: THEME 4: This looks at outcomes (the painful and damaging outcomes as a result of self stigma). The intro argues that we have information on outcomes, but need to learn about the process. Should this code be its own paper? On what outcomes are experienced qualitatively?

Response 12: Thank you for raising this concern as well. Our stated research question was “How do adolescents with diverse sexual orientations experience self-stigma?”, and what we found in the interviews was that the emotional experience of self-stigma was ultimately inextricable from its contents and overall experience. The same argument applies to the inclusion of the self-acceptance theme. We therefore don’t believe that the experience of self-stigma can be fully understood without description of some of the impacts that it has as well, and our perspective is that separating this theme to its own paper would make this one less complete.

Round 2

Reviewer 3 Report

Lines 124-125:  I would also be curious how this community discusses this construct, to learn if it actually is discrete—to address the assumption that self stigma DOES differentiate from internalized homophobia. 

Lines 144-145: Instead of saying to likely elicit and be responsive to respect, I would say “with aims or goals of being respectful”, to show what values you place in the protocol as opposed to 

Lines 151-155: I know these are your interview questions, but starting interviews with “do you…” implies a yes or no response, and does not elicit expanded answers. I know you used probes, but the structure of these questions are not aligned with recommended methods of qualitative research to elicit rich data. Please consider discussing this as a limit in the discussion.

Lines 186: Spell out acronym, and throughout.

Within your data analysis, what methods did you use to ensure trustworthiness (Morrow, 2005) and reliability (interrater reliability)? 

Results: What’s the rationale for including themes with low representation among participants? (e.g., Theme 4: Self stigma can be painful and damaging.

Author Response

Response to Reviewer 3 Comments

Point 1: Lines 124-125:  I would also be curious how this community discusses this construct, to learn if it actually is discrete—to address the assumption that self-stigma DOES differentiate from internalized homophobia. 

Response 1: This is a great point and one we would like to investigate further as well in future research. We did not discuss the term internalised homophobia with participants in this study and so cannot provide any direct comparison of these constructs in the paper from the perspective of the community, however, we can affirm that the overall concept of the study was strongly supported by the reference group and self-stigma was seen as an important construct to investigate. We have added a sentence in the Discussion pointing toward this as an important future line of enquiry.

‘Future research may also seek to investigate how the community discusses these constructs, to more thoroughly understand if separating self-stigma from internalized homophobia is a meaningful distinction for community members as well.’

Point 2:  Lines 144-145: Instead of saying to likely elicit and be responsive to respect, I would say “with aims or goals of being respectful”, to show what values you place in the protocol as opposed to 

Response 2: Thank you for this suggestion. We’ve amended the sentence in question to instead read: “The group helped develop the questions for the interview protocol with the aim of being respectful and sensitive.”

Point 3: Lines 151-155: I know these are your interview questions, but starting interviews with “do you…” implies a yes or no response, and does not elicit expanded answers. I know you used probes, but the structure of these questions are not aligned with recommended methods of qualitative research to elicit rich data. Please consider discussing this as a limit in the discussion.

Response 3: Thank you for this feedback and we can understand the concern. Our reasoning for leading with a closed-ended question in this way was simply to acknowledge (with the prior understanding that we had of this topic) that some participants may experience this and some may not, and we had separate prompts in the interview protocol to follow depending on whether a yes or no response was provided. Our fears with not doing so would be that to assume the experience of self-stigma (or the absence of self-stigma) by not asking about it in this way would be invalidating to participants. It also was not our experience that the richness of our data was limited, as this question was always followed with further prompts that were open-ended themselves (and many participants did expand regardless, as it was a topic that was particularly meaningful to them). Regardless, we have noted this as a limitation in the discussion as you have suggested, as how it affected the data is ultimately difficult to predict:

“We also note the limitation that some of the questions asked in the semi-structured interview protocol were closed-ended (e.g. At times, do you feel any of this stigma about yourself?), rather than open-ended. This decision was made with the knowledge that self-stigma may not be a universal experience, and to avoid invalidating participants we sought to allow them the opportunity to be frank about this. Other studies of self-stigma have structured their interview protocols in a similar way (e.g., Eaton et al., 2016). We attempted to mitigate the effect of this on the richness of the data we elicited by seeking elaboration with open-ended prompts based on their responses; however, we acknowledge that this question structure is unconventional to general recommendations for qualitative research (Morrow, 2005).’

Point 4: Lines 186: Spell out acronym, and throughout.

Response 4: Thank you for highlighting that we had not previously explained that acronym; given it only appears once, we have replaced it with the full terms instead.

Point 5: Within your data analysis, what methods did you use to ensure trustworthiness (Morrow, 2005) and reliability (interrater reliability)? 

Response 5: We sought to establish reliability by comparing independent coding of the data between two authors, as well as verification of the overall synthesis of these codes with the remaining team members and the reference group. Per Morrow’s (2005) paper, we sought to enhance trustworthiness by:

1) seeking to understand participants’ constructions with depth by seeking to understand their context, building rapport, and entering these interviews with credibility by informing them they would be speaking to someone who is themselves queer;

2) embracing the positioning of the researcher as a co-constructor in meaning, and making some of the first author’s context and potential biases overt within the study;

3) self-reflection on these biases and the first author’s interpretation of the data with reflexive journaling conducted throughout the process;

4) repeated consultation with the youth reference group as to the interpretation of the data and construction of themes, as well as consultation with the other research team members, to receive correction and feedback;

5) the first author’s immersion in the data, by conducting the interviews and transcribing them personally;

6) following a systematic process of coding the data and collating these codes into themes, as described by Braun and Clarke (2006)

We have now added some of Morrow’s terminology around this topic in the Method section to make the purpose of these decisions more clearly identifiable.

Point 6: Results: What’s the rationale for including themes with low representation among participants? (e.g., Theme 4: Self stigma can be painful and damaging.

Response 6: Our reasoning for this was that although severe distress associated with self-stigma was only found in a small number of participants, it was found in almost all of the subset of participants who experienced persistently high levels of self-stigma. We therefore don’t believe the small number of participants ultimately discussed in this theme itself fails to substantiate the theme, as it is still the case that self-stigma, when experienced frequently and/or severely, can be very painful and damaging. To make this clearer in the manuscript, we have added the following clarifying statement:

“As previously discussed, most participants did not identify with self-stigmatizing beliefs in a stable and long-term way in the present, and for these individuals the occasional experience of self-stigma was uncomfortable but was overall not a pressing concern for their current mental health. For a small number of participants for whom self-stigma was a persistent experience (n = 4), however, self-stigma contributed to serious mental health difficulties currently or in the near past.”

We also note that although the overall proportion of participants for whom this theme directly applies was low, for those that it did, it was of great personal importance and consistently noted by all in this subset. To remove this theme may diminish these voices and provide a biased picture of the experience of self-stigma.