Lexical salience and inference boundaries in Kompas.com COVID-19 reporting during 2021

Abstract Lexical prominence in a news archive does not by itself establish moral panic. This study reassesses the documented analytical record of an earlier corpus analysis of Kompas.com COVID-19 reporting during 2021. It distinguishes reproducible arithmetic from reported lexical measurements and unverified descriptions of software outputs. The accessible empirical materials are the manuscript's tables, narrative and four embedded figures; the raw and cleaned news records, processing scripts and eight separately cited Voyant exports were not available in the revision materials. Monthly entries sum to 22,531 headline records and 649 article records, compared with stated totals of 22,491 and 740. These sums verify internal discrepancies rather than the underlying corpus size. Three embedded count labels rank COVID at 3,160, cases at 1,298 and health at 805, but their counting unit remains unresolved. The documented associations foreground pandemic, prevention and institutional vocabulary without establishing hostility, disproportionate concern or audience anxiety. Descriptions of segmented or non-chronologically ordered Voyant displays cannot support calendar trends, and the underlying displays require recovery before independent inspection. Unresolved selection rules, translation procedures, denominators and co-occurrence settings prevent reproduction of the central corpus findings. The contribution is a case-specific account of evidence requirements connecting lexical description, contextual moral attribution and social consequences, rather than an empirically validated model of moral panic. Record reconciliation and recovery of the original analytical materials remain necessary before corpus-level claims can be verified.