Reddit After Roe: A Computational Analysis of Abortion Narratives and Barriers in the Wake of Dobbs

Aria Pessianzadeh; Alex H. Poole; Rezvaneh Rezapour

cs.CLMar 23, 2026

Reddit After Roe: A Computational Analysis of Abortion Narratives and Barriers in the Wake of Dobbs

Authors: Aria Pessianzadeh, Alex H. Poole, Rezvaneh Rezapour

Organizations: Information Science Department, Drexel University, United States

Abstract

The 2022 U.S. Supreme Court decision in Dobbs v. Jackson Women's Health Organization reshaped the reproductive rights landscape, introducing new uncertainty and barriers to abortion access. We present a large-scale computational analysis of abortion discourse on Reddit, examining how barriers to access are articulated across information-seeking and information-sharing behaviors, different stages of abortion (before, during, after), and three phases of the Dobbs decision in 2022. Drawing on more than 17,000 posts from four abortion-related subreddits, we employed a multi-step pipeline to classify posts by information type, abortion stage, barrier category, and expressed emotions. Using a codebook of eight barrier types, including legal, financial, emotional, and social obstacles, we analyzed their associations with emotions and information behaviors. Topic modeling of model-generated barrier rationales further revealed how discourse evolved in response to shifting legal and cultural contexts. Our findings show that emotional and psychological barriers consistently dominate abortion narratives online, with emotions such as nervousness, confusion, fear, and sadness prevalent across discourse. By linking information behaviors, barriers, emotions, and temporal dynamics, this study provides a multi-dimensional account of how abortion is navigated in online communities.

Figures & tables

Barrier Category	$\kappa$	AC1
Legal & Policy	0.89	0.99
Financial & Insurance	0.89	0.99
Logistical & Geographical	0.69	0.98
Provider & Infrastructure	0.28	0.99
Medical & Physical	0.57	0.96
Informational & Navigation	0.30	0.84

Table 1: Inter-annotator agreement comparison (Cohen’s

\kappa

vs. Gwet’s AC1) for the eight abortion barrier categories.

Model	Precision	Recall	F1-Score
BERT	0.86	0.86	0.86
RoBERTa	0.93	0.93	0.93
DistilBERT	0.83	0.83	0.83
DeBERTa	0.88	0.87	0.87
Llama	0.78	0.75	0.76
GPT-4	0.86	0.85	0.85

Table 2: Classification of Information-Seeking vs. Sharing using 400 annotated posts

Figure 1: Number of abortion-related posts in 2022

Figure 2: RoBERTa results for information types across the whole dataset

Table 3: Comparison of abortion-related stages and information-seeking/sharing across phases.

Figure 3: Temporal Changes in Different Types of Abortion Barriers

Panel of Figure 4 — Figure 4: Barriers across (a) information types, (b) stages of abortion, and (c) phases of Dobbs

Model	LP	FI	LG	PI	MP	IS	EP	SI	Average F1
GPT 4	0.88	0.61	0.81	0.67	0.56	0.62	0.85	0.84	0.73
GPT 4.1	0.79	0.72	0.79	0.38	0.58	0.57	0.75	0.73	0.66
GPT 4.1-mini	0.81	0.69	0.81	0.55	0.57	0.64	0.8	0.86	0.72
GPT 4.1-nano	0.76	0.76	0.77	0.36	0.57	0.43	0.85	0.84	0.67
GPT 4o	0.92	0.8	0.82	0.67	0.58	0.7	0.8	0.72	0.75
GPT 4o-mini	0.79	0.59	0.77	0.44	0.55	0.72	0.8	0.82	0.68

Table 4: Performance of different LLMs on 8 Barrier Types. LP: Legal & Policy, FI: Financial & Insurance, LG: Logistical & Geographical, PI: Provider & Infrastructure, MP: Medical & Physical, IS: Informational & System Navigation, EP: Emotional & Psychological, SI: Social & Interpersonal.

Panel of Figure 5 — Figure 5: Emotions statistically stronger in: (a) Stages of abortion, (b) Phases of Dobbs

Figure 6: Emotions expressed with different barrier types

Appendix figures & tables9 assets

Supplementary material from the paper’s appendix.

Appendix

Pair	Normalized No.
EP & SI	0.25
EP & MP	0.11
EP & IS	0.07
EP & LG	0.07
EP & FI	0.06
MP & IS	0.05

Appendix

Table 5: Normalized frequency of barrier types that co-occur in posts. LP: Legal & Policy, FI: Financial & Insurance, LG: Logistical & Geographical, PI: Provider & Infrastructure, MP: Medical & Physical, IS: Informational & System Navigation, EP: Emotional & Psychological, SI: Social & Interpersonal

Figure 8: Emotion in Seeking vs. Sharing. Positive values indicate seeking is more frequent, negative values reflect more frequent sharing.

Figure 9: Most frequent pairs of emotions that co-occur

Model	F1 Score
GPT-4.1	0.91
GPT-4.1-mini	0.87
GPT-5-mini	0.87
GPT-5-nano	0.81

Appendix

Table 6: F1 scores for abortion stage classification across four LLMs evaluated on the same held-out labeled test set. Higher values indicate better agreement with human-annotated stage labels.

Group	Barrier	Statistic	p-value	Cramér’s V	Significant
Information Type	Legal & Policy	0.82	$3.66e^{-1}$	0.010	No
Information Type	Financial & Insurance	14.33	$1.53e^{-4}$	0.041	Yes
Information Type	Logistical & Geographical	10.62	$1.12e^{-3}$	0.036	Yes
Information Type	Provider & Infrastructure	64.65	$8.95e^{-16}$	0.088	Yes
Information Type	Medical & Physical	92.81	$5.75e^{-22}$	0.105	Yes
Information Type	Informational & System Navigation	525.06	$3.35e^{-116}$	0.251	Yes

Appendix

Table 7: Chi-square test results for barrier prevalence across information type, abortion stage, and Dobbs phase.

Group	Emotion	Statistic	p-value	Cramér’s V	Significant
Information Type	nervousness	1136.68	$3.53e^{-249}$	0.368648	Yes
Information Type	confusion	958.01	$2.41e^{-210}$	0.338437	Yes
Information Type	fear	141.04	$1.58e^{-32}$	0.129854	Yes
Information Type	sadness	509.40	$8.57e^{-113}$	0.246787	Yes
Information Type	curiosity	873.89	$4.66e^{-192}$	0.323237	Yes
Information Type	relief	1075.57	$6.73e^{-236}$	0.358602	Yes

Appendix

Table 8: Chi-square test results for the top 10 emotions across information type, abortion stage, and Dobbs phase.

Barrier Type	Phase	Topics
	Phase 1	Early Pregnancy uncertainty- Informal/Underground Pill access- MA Timing and Appointment Constraints
	Phase 2	Pregnancy Discovery- State Restrictions- Navigating Aid resources
Legal & Policy	Phase 3	Legal Risk Anxiety- Cross-Region Pill Access- Legality and Weeks Limits
	Phase 1	Cost Anxiety- Ordering Pills Online: Payment + Legitimacy- Financial barriers
	Phase 2	Financial Help-Online Ordering- Financial hardship decisions
Financial & Insurance	Phase 3	General Financial Stress- Access Inequality- Financial support

Appendix

Table 9: Topics extracted for each Barrier in 3 Phases

Barrier Type	Info Type	Topics
	Seeking	Abortion legality concerns- Abortion legal restrictions- Abortion access restrictions- Accessing abortion pills- Parental consent abortion- Abortion timing laws- Customs and legal anxiety
Legal & Policy	Sharing	Abortion access barriers- Roe v. Wade
	Seeking	Abortion access barriers- Financial hardship abortion- Abortion cost barriers
Financial & Insurance	Sharing	Abortion access barriers
	Seeking	Abortion access challenges- Delays accessing medication- Package customs delays- Abortion pill access- Abortion appointment delays
Logistical & Geographical	Sharing	Abortion access challenges- Shipping delays concerns- Abortion pill access

Appendix

Table 10: Topics extracted for each barrier in 2 Information Types

Healthcare

Healthcare AI develops and evaluates models and systems for clinical care, medical research, public health, and healthcare operations.

See all 616 papers

Jun 11, 2026

A Context-Aware Dataset for Stance Detection in Bioethical Controversies on Reddit

Bioethical debates increasingly unfold on social media, yet stance detection research lacks large-scale, domain-specific resources for modeling such context-dependent discourse. We present BioStance, a context-aware dataset of 39,600 annotated Post-Comment pairs from Reddit bioethical discussions. BioStance covers six controversial targets across three dimensions of bioethical controversy: fundamental value conflicts, individual liberty versus collective responsibility, and technological uncertainty. Each instance preserves hierarchical conversational context and is labeled by three independent annotators using a three-class stance scheme: Favor, Against, and None. The annotations achieve a mean Krippendorff's

α

of 0.82, indicating substantial reliability. By combining thematic diversity, conversational structure, and high-quality human annotation, BioStance supports research on context-aware stance detection, argument mining, and computational analysis of bioethical discourse.

Hu Huang, Genan Dai, Fuqiang Niu +3

School of Cyber Science and Technology, University of Science and Technology of China, Hefei, China · School of Artificial Intelligence, Shenzhen Technology University, Shenzhen, China · School of Urban Planning and Design, Peking University, Shenzhen, China

Social Media Analysis Argument Mining Text Classification

Ask PDF

Mar 10, 2026

Privacy and Safety Experiences and Concerns of U.S. Women Using Generative AI for Seeking Sexual and Reproductive Health Information

The rapid adoption of generative AI (GenAI) chatbots has reshaped access to sexual and reproductive health (SRH) information, particularly following the overturning of Roe v. Wade, as individuals assigned female at birth increasingly turn to online sources. However, existing research remains largely model-centered, paying limited attention to user privacy and safety. We conducted semi-structured interviews with 18 U.S.-based participants from both restrictive and non-restrictive states who had used GenAI chatbots to seek SRH information. Adoption was influenced by perceived utility, usability, credibility, accessibility, and anthropomorphism, and many participants disclosed sensitive personal SRH details. Participants identified multiple privacy risks, including excessive data collection, government surveillance, profiling, model training, and data commodification. While most participants accepted these risks in exchange for perceived utility, abortion-related queries elicited heightened safety concerns. Few participants employed protective strategies beyond minimizing disclosures or deleting data. Based on these findings, we offer design and policy recommendations, such as health-specific features and stronger moderation practices, to enhance privacy and safety in GenAI-supported SRH information seeking.

Ina Kaleva, Xiao Zhan, Ruba Abu-Salma +1

King’s College London London, United Kingdom · VRAIN, Universitat Politècnica de València & University of Cambridge Valencia, Spain, Cambridge, United Kingdom · INGENIO (CSIC-Universitat Politècnica de València) Valencia, Spain

Human-AI Interaction Privacy-Utility Tradeoff

Ask PDF

Dec 15, 2025

Behavioral Coherence: A Method for Sensitive-Domain LLM Evaluation

People use LLMs for reproductive-health questions, including abortion-related support. A response can sound supportive while answers reinforce harmful assumptions: judgment is likely, secrecy is safer, and support is limited. We introduce behavioral coherence evaluation, a design-time method that uses validation evidence from an established instrument to test relations among outputs. Using the Individual Level Abortion Stigma Scale, we prompted five LLMs to complete questionnaires for 627 personas and reviewed flagged patterns with five reproductive-health experts. Models scored personas lower on self-judgment but higher on worries about judgment; most made worries the highest-scoring dimension, although it was lowest in the ILAS reference sample. Four of five models reversed the reference direction by generating significantly higher worries about judgment scores for Black personas. Models defaulted to extreme secrecy after abortion despite varying stigma patterns across personas. Expert review showed that disclosure guidance requires context about relationship safety, legal risk, and trusted support.

Anika Sharma, Malavika Mampally, Chidaksh Ravuru +2

Society-Centered AI Lab UNC Chapel Hill, USA · School of Nursing UNC Chapel Hill, USA

Language Model Bias Evaluation Demographic Bias in Language Models Medical LLMs

Ask PDF

Abstract

Healthcare

Explore similar work

A Context-Aware Dataset for Stance Detection in Bioethical Controversies on Reddit

Privacy and Safety Experiences and Concerns of U.S. Women Using Generative AI for Seeking Sexual and Reproductive Health Information

Behavioral Coherence: A Method for Sensitive-Domain LLM Evaluation