# A Shaky Voice Is Not Always a Dodge: Benchmarking Textual and Vocal Evasion Detection in Earnings Calls

arXiv: https://arxiv.org/abs/2608.28040 · HTML: https://mlsanity.com/paper/2608.28040/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 26 references that are in the map.
- First posted on arXiv 28 Aug 2026.

## Builds on (most cited first, 7 of 7)

- [WhisperX: Time-Accurate Speech Transcription of Long-Form Audio](https://mlsanity.com/paper/2303.00747/) (2023)
- [NumHTML: Numeric-Oriented Hierarchical Transformer Model for Multi-task Financial Forecasting](https://mlsanity.com/paper/2201.01770/) (2022)
- [ECC Analyzer: Extract Trading Signal from Earnings Conference Calls using Large Language Model for Stock Performance Prediction](https://mlsanity.com/paper/2404.18470/) (2024)
- ["I Never Said That": A dataset, taxonomy and baselines on response clarity classification](https://mlsanity.com/paper/2409.13879/) (2024)
- [SubjECTive-QA: Measuring Subjectivity in Earnings Call Transcripts' QA Through Six-Dimensional Feature Analysis](https://mlsanity.com/paper/2410.20651/) (2024)
- [EvasionBench: A Large-Scale Benchmark for Detecting Managerial Evasion in Earnings Call Q&A](https://mlsanity.com/paper/2601.09142/) (2026)
- [KCLarity at SemEval-2026 Task 6: Encoder and Zero-Shot Approaches to Political Evasion Detection](https://mlsanity.com/paper/2603.06552/) (2026)

## Abstract (arXiv)

Existing approaches to evasion detection in earnings calls focus on textual transcripts, treating evasion as a single-dimensional phenomenon. We argue that evasion in spoken communication is inherently multidimensional: beyond what executives say, how they say it carries independent and complementary information. To study these dimensions jointly, we introduce DualEvasion, a benchmark for evasion detection across text and audio in earnings call Q&A. The benchmark contains 505 annotated question-answer pairs from 60 earnings calls, each with two independent labels: textual evasion (direct vs. evasive) and vocal cues operationalized as speaker confidence (confident vs. unconfident). Our experiments show that state-of-the-art multimodal models struggle to detect vocal confidence, particularly on unconfident responses. Our analysis suggests these models interpret acoustic cues in isolation rather than relative to each speaker's baseline. Providing speaker-level references yields modest improvements, but a substantial gap with human performance remains.