@inproceedings{hassan-2026-divergence,
    title = "The Divergence Hypothesis: Unmasking Lexical Interference and Label Bias in Mental Health {NLP}",
    author = "Hassan, Moustafa Yehia",
    editor = "Demner-Fushman, Dina  and
      Ananiadou, Sophia  and
      Roberts, Kirk  and
      Tsujii, Junichi",
    booktitle = "{B}io{NLP} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-naloma/2026.bionlp-1.1/",
    doi = "10.18653/v1/2026.bionlp-1.1",
    pages = "1--14",
    ISBN = "979-8-89176-434-7",
    abstract = "Computational mental health (CMH) classifiers often degrade under distribution shift because human annotators and distant-supervision pipelines reward different linguistic signals. We introduce TSS (Triple-Stream Stress probe), a multi-channel diagnostic framework that decomposes text into (A) lexical character $n$-grams, (B) a small, mostly content-free morpho-syntactic channel, and (C) a 154-feature psycholinguistic style channel. Across four English datasets ($N = 12,906$), TSS reveals a \textit{lexical interference effect}: adding lexical features to the style channel reduces Macro-F1 on human-labeled data (mean drop 0.072, $p < 10^{−4}$) but not on auto-labeled data. We propose Degree of Divergence (DoD), a difference-in-differences statistic adapted from econometrics for label-source auditing, with instance-level bootstrap inference; the headline estimate is DoD$_{BC–A}$ = 0.0374, 95{\%} CI [0.0097,0.0651], $p = 0.0032$. A platform-stratified Twitter-only DoD (which removes the Reddit vs. Twitter contrast) reproduces the pattern with bootstrap inference: DoD$^{Tw}_{BC–A}$ = +0.096 ($p < 0.001$) and DoD$^{Tw}_{AC–A}$ = {\ensuremath{-}}0.089 ($p < 0.001$). Interventional masking (\texttt{pos{\_}only}) retains {\ensuremath{\sim}}95{--}99{\%} of Channel C{'}s performance after destroying content words on human datasets, indicating that the style channel does not rely primarily on lexical surface form. TSS is positioned as a diagnostic audit framework, not a clinical screening tool: it flags label-source-specific shortcut learning before generalization claims are made."
}