The Role of Gender in Borderline Personality Disorder (BPD-G)

Scientific Background Studies have shown that people with BPD experience increased levels of stigma and that this has harmful effects on their mental health (e.g., Rüsch et al., 2006; Quenneville et al., 2020; Ring & Lawn, 2025; van Schie et al., 2024). The role of sex (biological or sex assigned at birth) and gender (socially and culturally constructed gender) aspects in the stigmatization experienced by individuals with BPD has been scarcely investigated to date. It has been hypothesized, however, that the heightened experience of stigmatization is partly related to the fact that the group of individuals with BPD studied predominantly belongs to the female sex assigned at birth (Quenneville et al., 2020). BPD is a severe mental illness characterized by, among other things, unstable emotions, relationships, and identity (APA, 2013). The prevalence ranges from 0.7% to 2.7% in the general population and from 6% to 22% in clinical samples (Leichsenring et al., 2024). In the latter, women are significantly overrepresented at over 70%, while interview-based epidemiological studies find comparable rates among men and women (Grant et al., 2008; Tomko et al., 2014). Feminist-oriented studies criticize psychiatric classification systems for reproducing patriarchal gender stereotypes: Thus, traits considered "too masculine for a woman" (e.g., the criterion of "inappropriate anger") as well as traits with "too feminine" connotations (e.g., the criterion of "mood instability") would be pathologized and thereby stigmatized (Cohen & Hartmann, 2021; Fellows, 2023).

The few studies that have examined the role of gender in people with BPD have focused on samples of people with sexual and gender diversity (SGD)-that is, samples comprising individuals with a sexual orientation other than heterosexuality or a gender identity other than cisgender (where gender identity aligns with the sex assigned at birth). These studies revealed associations between gender and homosexual or bisexual orientation and BPD symptoms; however, the authors emphasize the insufficient data and the need for further research (Shu et al., 2024; Denning et al., 2022). Recent experimental studies also point to gender-dependent differences in social stigmatization: For instance, male-coded BPD vignettes were rated with more anxiety and perceived danger, while female-coded ones were rated with more compassion by the general population (Masland & Null, 2022; Brown, 2024). Together with the knowledge that invalidating relationship experiences are considered central etiological factors in the development of BPD (e.g., Linehan, 1993; Porter et al., 2020; Bateman & Fonagy, 2010), these findings underscore the importance of conducting a differentiated examination of gender-specific relationship and stigmatization experiences in BPD.

Studies that focused on gender-related stigmatization and its impact on mental health also primarily examined SGD samples (Swann et al., 2024; Liu et al., 2025; Scandurra et al., 2020). Experiences of gender-related rejection among transgender and gender nonconforming (TGNC) adults were associated with poorer mental health via internalized trans-negativity-that is, internalized negative attitudes toward trans* individuals-but this effect was reduced with higher mentalization ability (Scandurra et al., 2020), suggesting that mentalization may serve a protective function in this context.

Mentalization describes the ability to explain one's own and others' behavior based on internal states, while maintaining an appropriately curious, open, and "correctable" attitude. Mentalization, as a kind of "psychological buffer" against the effects of aversive experiences on stress and mental health, has frequently been discussed as a resilience-promoting ability (Fonagy et al., 2019). Mentalization develops in early attachment relationships through "social biofeedback" (Gergely & Watson, 1996) as early as childhood, as the child receives feedback about their internal states (affective mirroring) from their immediate environment. Similarly, Koivisto et al. (2022) hypothesized that gender stereotypes might be relevant to the experience of stigmatization among patients with BPD: Rigid self-invalidations observed in a mentalization-based therapy (MBT) group could be associated with female gender stereotypes, which children exhibiting "feminine" behavior internalized as an "Alien Self" through feedback from their environment. In mentalization theory, the "Alien Self" refers to the internalization of a self-image that corresponds, for example, to destructive, incongruent "mirrorings" of one's own experience by the environment (Fonagy et al., 2018).

For example, a belief such as "I am manipulative" may become part of a person's self-concept due to frequent attribution by those around them, yet it may (unconsciously) feel foreign to them in their own experience and remain difficult to regulate effectively. Based on findings regarding the increased stigmatization of BPD vignettes coded as male (Masland & Null, 2022), it could be analogously assumed for individuals perceived as male that "male gender stereotypes" contribute to the formation of an Alien Self (e.g., "I am dangerous"). While MBT was developed specifically for the treatment of BPD and its core goal is to improve mentalization ability, including support in dealing with the Alien Self, it must also be noted here that gender has been neglected thus far.

To address the long-standing neglect of the significance of sex and, in particular, gender in theories and practices regarding the treatment of borderline personality disorder, this research project aims to 1) examine gender-related relationship and self-experiences in individuals with BPD and 2) conduct a pilot study on the use of the Single-Category Implicit Association Test (adapted from von Hippel et al., 2018) to investigate implicit thought processes related to sex and BPD.

Project Objectives This project aims to investigate, for the first time, the significance of gender in attachment relationships and self-perception among 15 individuals with BPD using the Critical-Constructivist Grounded Theory (CCGT; Levitt, 2025) method. Furthermore, the research toolkit in this area is to be expanded. Therefore, the study will also examine the extent to which Single-Category Implicit Association Tests (SC-IATs, by Hippel et al., 2018) can be validly used in the target population to investigate implicit gender- and BPD-related thought processes. Since the aim of the study is not limited to the exclusive examination of gender but also takes "sex" into account, the term "gender" is used in the following for summary purposes. Correlations with self-report measures of relationship and self-experience will be explored.

Endpoints and Hypotheses Research Question 1: What gender-related relationship and self-experiences can be identified in individuals with BPD? How are these experiences related to mentalization ability (as measured by coded data and self-reports), as well as to well-being, quality of life, patriarchal attitudes within the family, and psychological symptoms (as reported by the participants)? Research Question 2: Is there preliminary evidence that Single-Category Implicit Association Tests (SC-IATs) are suitable and reliable instruments for measuring implicit gender- and BPD-related thought processes in people with BPD? Primary Endpoints Research Question 1: A conceptual description of the significance of gender in relationships with oneself and others among individuals with BPD, based on a hierarchical categorical structure and taking structural power dynamics into account.

Research Question 2: D-scores (ratio of reaction times between experimental blocks), stimulus accuracy (proportion of correct trials per stimulus/proportion of all trials per stimulus), internal consistency of the SC-IATs, indications of multicollinearity (intercorrelation of D-scores), and variance

Secondary Endpoints Research Question 1: Superordinate and subordinate categories, identification of relevant contextual factors, differential typologies, alignment of the summarized main findings with patients' experiences, descriptive correlations with quantitatively measured psychosocial constructs Research Question 2: Duration of administration, error rates, indications of convergent validity (exploratory, correlation between D-scores and explicit measures), participant acceptance (qualitative feedback), exploratory correlations with the quantitatively measured psychosocial constructs

Study population Participants The study aims to recruit 15 individuals diagnosed with BPD (based on psychological, psychiatric, or medical evaluations, with the most recent diagnosis made no more than 2 years ago) and representing a range of gender identities (target: N=5 cis female, N=5 nonbinary/trans/gender-nonconforming/other self-defined identities, N=5 cis male).

The primary focus of the study is to investigate the role of gender in individuals with BPD and to evaluate a research instrument in this area; therefore, it is necessary to conduct the study among a group of individuals with BPD and, in some cases, members of a vulnerable group based on a non-binary or trans* gender identity.

Inclusion and Exclusion Criteria Inclusion and exclusion criteria are assessed by psychologists holding at least a bachelor's degree through a self-report questionnaire. The presence of a borderline personality disorder diagnosis is verified by obtaining existing psychological and psychiatric diagnostic results that were conducted within the last 2 years as part of psychological or psychiatric care and/or study participation. This retrieval of diagnostic results is based on a waiver of confidentiality provided by the participants.

Inclusion Criteria:

1. Signed informed consent form
2. Individuals of all genders (target: n=5 with female gender identity, n=5 with nonbinary/trans*/gender-non-conforming (TGNC) gender identity, n=5 with male gender identity) between the ages of 18 and 65 at the time of consent
3. Individuals with a BPD diagnosis made within the last 2 years (ICD-10: F60.3; DSM-5: 301.83, ICD-11: Borderline qualifier)
4. Understanding of the study procedure, ability and willingness to participate
5. The participant is able to read and understand the informed consent form and can provide a written, personally signed, and dated informed consent form.

Exclusion Criteria

1. Acute suicidal ideation or acute psychotic symptoms reported by the patient or suspected
2. Insufficient language skills

Recruitment measures and materials The majority of recruitment takes place among participants in the study "Mentalization-based treatment versus bona-fide treatment for patients with Borderline Personality Disorder in Germany (MAGNET): a prospective, multi-center randomized controlled trial" (S-493/2023). Only those participants who indicated in the MAGNET study consent form that they consent to being contacted for further studies and have not withdrawn this consent will be contacted.

For the purposes of Research Question 1, theoretical sampling (Glaser and Strauss, 1967) is applied during recruitment. Here, data collection takes place in several rounds (Levitt, 2021). Initially, participants with different characteristics that appear relevant to the topic (here: various gender identities) are recruited. Based on the ongoing data analysis for Research Question 1, subsequent rounds will specifically invite individuals with experiences that have been underrepresented thus far. This allows for the consideration of variations in characteristics whose relevance only becomes apparent during the analysis.

In addition, individuals with diverse gender identities may be recruited through the Institute for Psychosocial Prevention and Psychotherapy at Heidelberg University Hospital using flyers (posted at the institute, on the website, and distributed during initial consultations).

Informed consent procedure At the study center, participants will receive written and verbal information about the study from study staff members who hold at least a bachelor's degree in psychology. Informed consent will be obtained in writing.

Methodology and Implementation Study Design This study is a mixed-methods, single-arm cross-sectional study comprising 1) a qualitative, exploratory, critical-constructivist content analysis examining the significance of gender in relationship experiences and self-perception among individuals with borderline personality disorder and various gender identities, and 2) a pilot study on the use of the Single-Category Implicit Association Test (SC-IATs, adapted from von Hippel et al., 2018) as preparation for a larger-scale experimental psychological study using the SC-IAT to assess implicit gender- and BPD-related thought processes in people with borderline personality disorder and various gender identities.

Description of Study Procedure The study consists of two sessions. The first session lasts approximately 4 hours, including breaks, and takes place on-site. The second session is an online survey conducted approximately 6-12 months after the first session, which takes about 20 minutes (T2).

T1 - On-site Assessment: First, participants are asked to create a pseudonymized code, which is used to store their data and link it to their identity. Participants then complete computer tasks, take part in an interview about the significance of gender in their lived experience, and finally fill out questionnaires regarding their well-being.

Computer tasks:

Participants are first asked to complete three SC-IATs (adapted from von Hippel et al., 2018) on the computer to examine 1) the implicit evaluation (positive/negative) of BPD, 2) the implicit identification (self/others) with BPD, and 3) the implicit association of gender (in binary terms: female/male) with BPD. Each SC-IAT consists of 2 blocks, each with 15 practice trials and 30 test trials. Completing an SC-IAT takes between seven and 10 minutes. Including breaks, completing the 3 SC-IATs takes approximately 45 minutes.

Task description using the example of the 1) SC-IAT for the implicit evaluation of BPD: In each trial, participants are asked to assign a term from the positive/negative category (e.g., "pleasant") or the BPD category (e.g., "emotionally unstable")-presented in the center of the screen-as quickly and accurately as possible to one of two categories displayed at the top of the screen. To examine the implicit evaluation of BPD, the word "BPD" is presented in each trial alongside one of the category labels at the top of the screen.

For example, "positive+BPD" appears in the upper-left corner of the screen; the other option is not associated with BPD, so "negative" appears in the upper-right corner of the screen. This assignment ("positive+BPD" versus "negative") is used throughout the entire first block across all trials. Participants are informed that by pressing the "E" key in this block, they indicate their assignments to the "positive+BPD" category, and by pressing the "I" key, they assign a term to the "negative" category. For example, it would be correct to press "E" for "pleasant," press "I" for "bad," and press "E" for "emotionally unstable." In Block 2, the association between "BPD" and "positive" is then reversed. Now "negative+BPD" represents one category and "positive" the other. Before Block 2 begins, participants are informed of the switch on an instruction page and asked to adjust to the new categorization. During the 15 practice trials per block, participants receive feedback in the form of a red cross if they have incorrectly assigned a term.

The other two SC-IATs are structured similarly: In SC-IAT 2, which examines implicit identification with BPD, a distinction is made between the categorization of "I+BPD" versus "others" in Block I, and "I" versus "others+BPD" in Block II. In SC-IAT 3, which examines the implicit association of gender with BPD, a distinction is made between the assignment of "female+BPD" versus "male" in Block I, and "female" versus "male+BPD" in Block II.

The order in which the three SC-IATs are administered is randomized among participants. Short breaks are taken between each of the three tasks. After completing all three tasks, participants are asked in a final questionnaire how they found the tasks and the terms used (quantitative and qualitative assessment of comprehensibility and perceived stress; qualitative suggestions for improvement are also asked for).

Interview:

Participants will then take part in a semi-structured interview about the significance of gender in their lived experience. The interview will last between 60 and 90 minutes and will be recorded on tape using a voice recorder. At the beginning of the interview, participants will be informed of their right not to answer questions, to take breaks, or to stop the interview at any time. It will be emphasized that, as participants, they are experts on their own experiences, and that it is precisely these experiences that are relevant to the interview. To create a relaxed atmosphere, the interviewer may also share information about their own identity, such as their gender identity. After brief questions regarding demographic data, participants are asked about their experiences with BPD and gender in various areas of life, with a focus on the interaction between these two aspects. The interview protocol was developed to best capture the significance of gender in lived experience; its structure was based on the Brief Reflective Functioning Interview for assessing mentalization ability in attachment relationships (Andreas et al., 2022). Finally, participants are given the opportunity to add further aspects they consider relevant. Following the principle of theoretical sampling, the focus of the interview may vary from person to person within the same set of topics and may shift slightly as the research process progresses and new insights emerge.

After the interview, the audio file is transcribed. The text is analyzed using MAXQDA 2025 software (VERBI Software, 2025). To do this, the text is divided into units of meaning, which are then labeled. These units of meaning are continuously compared both within and across interviews, and overlapping units are grouped into initial categories. These are then compared, and larger categories with subcategories are formed from them. In the course of the analysis, these categories are grouped into clusters until a core category emerges at the top of the hierarchy. The final hierarchy should allow for as much differentiation as possible while remaining as streamlined as possible. The interview analysis is also conducted quantitatively through the parameterization of the level of mentalization using the Reflective Functioning Scale (Fonagy et al., 1998), which allows for the standardized coding of participants' levels of reflection based on content markers in their interviews.

After completing the computer tasks and the interview, participants are also asked to complete questionnaires online via the SoSciSurvey platform regarding their well-being, their ways of reflecting, thinking, and feeling, their experiences of stigmatization due to mental illness, their borderline personality disorder symptoms, as well as their mental well-being and quality of life. For this purpose, the questionnaires listed under Outcome Measures are used.

T2 - Assessment online:

After the research team completes a preliminary analysis of the interviews approximately 6 to 12 months following the on-site session, participants will be contacted again via email and asked to provide a brief online assessment of a summary of the preliminary results (T2). They will be asked about the consistency and inconsistency of the results with their own experiences (a self-developed questionnaire based on Levitt et al., 2024) and will be asked to provide their pseudonymized code again.

Timeline First Participant anticipated 29th May 2026. End of T1-Assessments 31.10.2026. End of Analyses 30.11.2026. End of T2 Assessments 31.12.2026. Anticipated End 31.03.2027

(Statistical) Analysis: Research Question 1: The qualitative semi-structured interviews were developed to specifically explore the significance of gender in the relational and self-experiences of individuals with BPD; to this end, the Brief Reflective Functioning Interview (Andreas et al., 2022) was adapted and expanded to examine the role of gender. The focus of the interview is on the participants' experiences with borderline personality disorder, their gender identity, and the interplay between these topics. The focus within these thematic areas may vary among participants depending on their experiences. To this end, a sample of N=15 individuals with BPD and various gender identities (target: 5 female, 5 nonbinary/trans*/gender-nonconforming/other self-identified identities, 5 male) will be interviewed. For analysis, the interviews will be transcribed and divided into meaning units (Meaning Units) using the software MAXQDA 2025 (VERBI Software, 2025), which will then be named/labeled. Through constant comparison, each meaning unit is systematically compared with all others to consolidate overlapping meanings first into initial categories and then into categories with subcategories. Meaning units can be assigned to multiple categories. Redundant categories are consolidated. In an iterative process, categories are grouped into clusters, and the core category is ultimately developed from their shared concept. The final hierarchy should be as lean as possible while remaining comprehensive enough to capture all relevant distinctions.

For Research Question 2, the same 15 participants will be studied. The sample size is not based on a power calculation, as no inferential statistical hypotheses are being tested. Given the objective of conducting a methodological pilot study on the use of the SC-IAT in the target population, the sample size can be considered sufficient. Prior to analysis, the data are cleaned (trial level: reaction times under 300 ms and over 10,000 ms are removed; participant level: error rates exceeding 50% in an SC-IAT block, dropouts, self-reported issues such as dishonest responses or processing).

For the main analyses,

1. SC-IAT D-scores are calculated (D = (mean RT (Block 2) - mean RT (Block 1)) / pooled SD), where higher values indicate stronger associations (with the "positive" category or the "female" category) or stronger identification (with one's own person, "self" category).
2. The internal consistency of each SC-IAT is determined using the split-half method.
3. Intercorrelations between the three D-scores are calculated to preliminarily examine evidence of multicollinearity (confidence intervals are also reported).
4. For each stimulus, categorization accuracy is calculated (number of correct trials with this stimulus / total number of trials with this stimulus).

Exploratory secondary analyses include correlations between D-scores and participants' explicit self-reports. Descriptive data are reported grouped by gender identity. Quantitative feedback on the tasks is evaluated descriptively. Qualitative feedback on the tasks is evaluated for content.

Based on the results, a decision will be made as to whether 1) the SC-IATs will be implemented unchanged in a larger study, 2) implemented with modifications in a larger study, or 3) a fundamental revision of the design is necessary.

The study complies with the guidelines of the GDPR. Personal data is collected and processed in a pseudonymized form online via the SoSciSurvey platform and at the Institute for Psychosocial Prevention. SoSciSurvey is a secure platform that protects data from unauthorized access and stores it exclusively on servers in Germany. Participants are also asked about their age, gender, sexual orientation, and ethnic background. In addition, the email address and name are requested, and a personal identifier is used for pseudonymization. The email address and name are then stored separately from the data but together with the personal identifier on a list. Both the email address and name can only be linked to the data record via the personal identifier. The collection of the email address is necessary so that participants can be contacted for the T2 surveys.

The collected data is stored electronically using the personal code and secured against unauthorized access. The list linking the participants' names and email addresses to the personal code is accessible only to the study director and the project manager; that is, only these individuals can link the collected data to the name and email address.

Anonymized aggregated data from the computer tasks and questionnaire surveys will be uploaded to OSF. Pseudonymized data will be provided to other researchers exclusively upon plausible, verified request for the purpose of ensuring transparency in scientific research. Data from the interviews are not shared with third parties.

Image and Audio Recordings:

The interview will be recorded on tape; they will be transcribed as soon as possible and deleted immediately after transcription, or by December 31, 2027, at the latest. Absolute anonymization cannot be guaranteed within the scope of the study; however, re-identification is practically impossible due to the measures taken. All persons involved in the analysis are subject to strict confidentiality and may under no circumstances disclose confidential information to third parties.

Withdrawal of Consent:

Study participants may withdraw their consent at any time, without providing a reason and without suffering any adverse consequences. If a participant withdraws from the study, they will be asked whether they consent to the continued processing of data already collected. If they do not consent, any data already collected will be destroyed, provided it has not yet been included in an analysis and/or anonymized.

Data Deletion or Anonymization:

The list containing names, email addresses, and personal identification numbers will be destroyed upon completion of the final surveys, but no later than December 31, 2027. The data will then be pseudonymized, including sociodemographic characteristics such as gender, age, and sexual orientation, and stored for 10 years after the completion of the final surveys, and subsequently deleted. Separately, the consent forms, which contain the participants' names but not their personal identification codes, will be stored in a manner that protects them from unauthorized access. Since the link between direct identifiers such as name and email address and the personal code will be destroyed along with the list upon completion of the final surveys, but no later than December 31, 2027, it will only be possible to link the data to the name after the list's deletion date if the participants provide their code to the study staff for this purpose. It should be noted, however, that data can only be deleted as long as it has not yet been incorporated into an analysis. For interview data, which is continuously analyzed in parallel with data collection, this means that deletion of this data may no longer be possible within just a few days of collection. For questionnaire data and data from the computer tasks, the period during which data can be deleted may be longer due to the later analysis Since participants may be identifiable in the audio recordings they have made, they have the right to have these recordings deleted at any time without suffering any adverse consequences as a result.