# DAIC-WOZ: On the Validity of Using the Therapist's prompts in Automatic Depression Detection from Clinical Interviews

arXiv: https://arxiv.org/abs/2404.14463 · HTML: https://mlsanity.com/paper/2404.14463/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 33 references that are in the map.
- First posted on arXiv 22 Apr 2024.

## Builds on (most cited first, 10 of 10)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [AVEC 2016 - Depression, Mood, and Emotion Recognition Workshop and Challenge](https://mlsanity.com/paper/1605.01600/) (2016)
- [An Exploration of Hierarchical Attention Transformers for Efficient Long Document Classification](https://mlsanity.com/paper/2210.05529/) (2022)
- [Automatic Depression Detection: An Emotional Audio-Textual Corpus and a GRU/BiLSTM-based Model](https://mlsanity.com/paper/2202.08210/) (2022)
- [Gender Bias in Depression Detection Using Audio Features](https://mlsanity.com/paper/2010.15120/) (2020)
- [Affective Conditioning on Hierarchical Networks applied to Depression Detection from Transcribed Clinical Interviews](https://mlsanity.com/paper/2006.08336/) (2020)
- [Node-weighted Graph Convolutional Network for Depression Detection in Transcribed Clinical Interviews](https://mlsanity.com/paper/2307.00920/) (2023)

## Built on by (most cited first, 15 of 15)

- [Factual Serialization Enhancement: A Key Innovation for Chest X-ray Report Generation](https://mlsanity.com/paper/2405.09586/) (2024)
- [Leveraging Audio and Text Modalities in Mental Health: A Study of LLMs Performance](https://mlsanity.com/paper/2412.10417/) (2024)
- [Fine-tuning LLMs for Passive Depression Severity Estimation from AI Mental Health Dialogue](https://mlsanity.com/paper/2606.17973/) (2026)
- [Leveraging Embedding Techniques in Multimodal Machine Learning for Mental Illness Assessment](https://mlsanity.com/paper/2504.01767/) (2025)
- [When Consistency Becomes Bias: Interviewer Effects in Semi-Structured Clinical Interviews](https://mlsanity.com/paper/2603.24651/) (2026)
- [Psychologically-Grounded Graph Modeling for Interpretable Depression Detection](https://mlsanity.com/paper/2604.24126/) (2026)
- [ADAPTS: Agentic Decomposition for Automated Protocol-agnostic Tracking of Symptoms](https://mlsanity.com/paper/2605.03212/) (2026)
- [A Multi-Probe Audit of Clinical-Interview Depression Detection Benchmarks](https://mlsanity.com/paper/2605.23977/) (2026)
- [WEE-Therapy: A Mixture of Weak Encoders Framework for Psychological Counseling Dialogue Analysis](https://mlsanity.com/paper/2510.02320/) (2025)
- [A Systematic Evaluation of Large Language Models for PTSD Severity Estimation: The Role of Contextual Knowledge and Modeling Strategies](https://mlsanity.com/paper/2602.06015/) (2026)
- [Can Conversational Temporal Dynamics Improve Depression Detection in Dyads? A Preliminary Investigation in Multi-Modality Perspectives](https://mlsanity.com/paper/2607.03744/) (2026)
- [HiMA-MDD: A Hierarchical Multi-Agent Harness for Interpretable Multimodal Depression Detection in Clinical Interviews](https://mlsanity.com/paper/2608.21868/) (2026)
- [LongCounsel-8: A Benchmark Suite for Longitudinal Depression Tracking from Multi-Session Counseling Dialogues](https://mlsanity.com/paper/2609.03507/) (2026)
- [DiaWhisper-DPO: Role-Attributed Transcription of Clinical Interviews via Failure-Mined Preference Optimization](https://mlsanity.com/paper/2609.16661/) (2026)
- [Word-Aligned Prosody Reflects Speaker Differences More Than Depression or PTSD Severity](https://mlsanity.com/paper/2610.15181/) (2026)

## Abstract (arXiv)

Automatic depression detection from conversational data has gained significant interest in recent years. The DAIC-WOZ dataset, interviews conducted by a human-controlled virtual agent, has been widely used for this task. Recent studies have reported enhanced performance when incorporating interviewer's prompts into the model. In this work, we hypothesize that this improvement might be mainly due to a bias present in these prompts, rather than the proposed architectures and methods. Through ablation experiments and qualitative analysis, we discover that models using interviewer's prompts learn to focus on a specific region of the interviews, where questions about past experiences with mental health issues are asked, and use them as discriminative shortcuts to detect depressed participants. In contrast, models using participant responses gather evidence from across the entire interview. Finally, to highlight the magnitude of this bias, we achieve a 0.90 F1 score by intentionally exploiting it, the highest result reported to date on this dataset using only textual information. Our findings underline the need for caution when incorporating interviewers' prompts into models, as they may inadvertently learn to exploit targeted prompts, rather than learning to characterize the language and behavior that are genuinely indicative of the patient's mental health condition.