Structural Stress and Learned Helplessness in Afghanistan: A Multi-Layer Analysis of the AFSTRESS Dari Corpus

Jawid Ahmad Baktash; Mursal Dawodi; Nadira Ahmadi

cs.CLMar 28, 2026

Structural Stress and Learned Helplessness in Afghanistan: A Multi-Layer Analysis of the AFSTRESS Dari Corpus

Authors: Jawid Ahmad Baktash, Mursal Dawodi, Nadira Ahmadi

Organizations: Technical University of Munich (TUM), Germany · Technical University of Applied Sciences Würzburg-Schweinfurt (THWS), Germany

Abstract

We introduce AFSTRESS, the first multi-label corpus of self-reported stress narratives in Dari (Eastern Persian), comprising 737 responses collected from Afghan individuals during an ongoing humanitarian crisis. Participants describe experienced stress and select emotion and stressor labels via Dari checklists. The dataset enables analysis at three levels: computational (multi-label classification), social (structural drivers and gender disparities), and psychological (learned helplessness, chronic stress, and emotional cascade patterns). It includes 12 binary labels (5 emotions, 7 stressors), with high label cardinality (5.54) and density (0.462), reflecting complex, multi-dimensional stress. Structural stressors dominate: uncertain future (62.6 percent) and education closure (60.0 percent) exceed emotional states, indicating stress is primarily structurally driven. The strongest co-occurrence is between hopelessness and uncertain future (J = 0.388). Baseline experiments show that character TF-IDF with Linear SVM achieves Micro-F1 = 0.663 and Macro-F1 = 0.651, outperforming ParsBERT and XLM-RoBERTa, while threshold tuning improves Micro-F1 by 10.3 points. AFSTRESS provides the first Dari resource for computational analysis of stress and well-being in a crisis-affected population.

Figures & tables

ID	Label	Dari	Description	$n$
Emotion labels (5)
E1	Stress	\arabicfont استرس	Primary stress affect [ 26 ]	428
E2	Anxiety	\arabicfont اضطراب	Anxious arousal [ 32 ]	301
E3	Sadness	\arabicfont غمگینی	Low-activation sadness [ 19 ]	323
E4	Burnout	\arabicfont خستگی شدید	Exhaustion/burnout [ 30 ]	286
E5	Hopeless	\arabicfont ناامیدی	Hopelessness [ 2 ]	412

Table 1: Final AfStress annotation schema (12 labels: 5 emotions + 7 stressors).

\dagger

= merged label.

N=737

Figure 1: Demographic distribution ( $N=737$ ): Female 660 (90%), Male 77 (10%). 65% of respondents aged 18–25 (476 individuals). The 90% female rate reflects differential crisis impact, particularly the girls’ education ban, not a sampling error.

Figure 2: Label prevalence (% of 737 respondents, all 12 labels). Structural stressors, Uncertain future (62.6%) and Education closure (60.0%), exceed the prevalence of primary emotional states (Stress 58.1%).

Emotion	$n$	%	Stressor	$n$	%
Stress	428	58.1	Uncertain	461	62.6
Hopeless	412	55.9	EduClose	442	60.0
Sadness	323	43.8	Economic	382	51.8
Anxiety	301	40.8	Displace	279	37.9
Burnout	286	38.8	Social	275	37.3
			Family	250	33.9

Table 2: Label frequency table (

N=737

Figure 3: Label cardinality: responses by number of active labels. Mean = 5.54 (dashed line). 65% of responses carry $\geq$ 4 labels. High cardinality indicates compound, chronic stress rather than isolated conditions.

Figure 4: Emotion–Stressor co-occurrence matrix: joint positive counts ( $N=737$ ). Hopelessness–Uncertain future (329) and Stress–Uncertain future (300) are the strongest pairs. All cells are positive: every stressor co-occurs with every emotion.

Figure 5: Pearson correlation heatmap (12 $\times$ 12, merged labels). All off-diagonal correlations are positive. Strongest pair: Hopelessness–Uncertain future ( $r=0.40$ ).

Figure 6: Emotional cascade network from Jaccard similarity of emotion label co-occurrences (edges shown for $J\geq 0.15$ ). The sequence Stress $\to$ Anxiety $\to$ Sadness $\to$ Burnout $\to$ Hopelessness matches emotional cascade theory [ 40 ] . Strongest edge: Stress–Anxiety ( $J=0.49$ ).

System	Hamming $\downarrow$	Micro-F 1	Macro-F 1	Samp.-F 1
TF-IDF baselines
(A) Char TF-IDF + LR	0.491	0.657	0.647	0.615
(B) Char TF-IDF + SVM	0.480	0.663	0.651	0.620
(C) Char TF-IDF + NB	0.428	0.638	0.578	0.600
(D) Char+SVM+Metadata	0.488	0.662	0.653	0.618
Transformer models

Table 3: Baseline results on held-out test set (

N=111

, 12 labels, tuned thresholds). Best per-metric in bold .

Figure 7: Model comparison: Micro-F1 and Macro-F1 for all systems (tuned thresholds). Classical TF-IDF models outperform both transformers on this small corpus.

Affective Computing

Affective computing develops systems that recognize, model, and respond to human emotions and affective states from behavior and physiological signals.

See all 141 papers

An AI-Powered Culturally Aware Chatbot for Stress Detection and Wellness Support among Pakistani University Students Using NLP and Machine Learning
Sep 10, 2026Muhammad Fahad Bashir, Muhammad AfzalMental Health Mental Health Conversational Agents
Department of Computer Science & Information Technology, Ghazi University Dera Ghazi Khan, Pakistan
Empath: Tracing Multi-Level Emotion Dynamics in Crisis Counseling Dialogues
Sep 24, 2026Ziwei Gong, Yuchen Huang, Wen Liang +4Mental Health Counseling Emotional Support Conversation
Columbia University, USA · Barnard College, USA
Automatic Detection of Stress from Speech in the Trier Social Stress Test
Jul 1, 2026Hanna Drimalla, Wieland R. Cremer, Christine Kraus +1Speech Emotion Recognition Affective Computing
Human-Centered Artificial Intelligence Group, Faculty of Technology, Bielefeld University, Bielefeld, Germany. · Department of Cognitive Psychology, Faculty of Psychology, Ruhr University Bochum, Bochum, Germany.

Sep 10, 2026J/K move · Enter open · S save

An AI-Powered Culturally Aware Chatbot for Stress Detection and Wellness Support among Pakistani University Students Using NLP and Machine Learning

Muhammad Fahad Bashir, Muhammad Afzal

Department of Computer Science & Information Technology, Ghazi University Dera Ghazi Khan, Pakistan

Mental Health Mental Health Conversational Agents Conversational Agents

Open paper AskPDF arXiv

With the existing digital mental health tools specifically developed for Western settings, Pakistani students are exposed to a uniquely compounded stress situation in their university that includes academic, financial, familial, and relational stressors, which have become a serious concern for academic and psychological development of students in Pakistani universities. This paper introduces a new, AI-driven and culturally sensitive stress detection and wellness support system that is tailored to the context of Pakistani university students. The system is based on a machine learning model called Random Forest which is trained using a validated student stress data set of 1100 responses on 20 features from psychological, physiological, academic, environmental and social aspects, with an accuracy of 89.09% and a macro F1-score of 0.89, in three stress severity levels. The classification outputs are passed on to an open-source large language model through OpenRouter API, where an appropriately crafted system prompt, culturally aware, gives the model a conversation about wellness, in English, Urdu and Roman Urdu. The second most predictive stress factor in this population identified by feature importance analysis was teacher-student relationship, which is a culturally important stress factor highlighting the need for region-aware mental health systems. Future research will involve primary data collection from students at various academic levels of Pakistani Universities with the validated DASS-21 instrument focusing on the students who are moving from FSc to undergraduate studies, which is a time of being psychologically vulnerable which is under-researched.

Abstract

Affective Computing

Explore similar work

An AI-Powered Culturally Aware Chatbot for Stress Detection and Wellness Support among Pakistani University Students Using NLP and Machine Learning

Empath: Tracing Multi-Level Emotion Dynamics in Crisis Counseling Dialogues

Automatic Detection of Stress from Speech in the Trier Social Stress Test