Newest papers

Which Text Matters: Handling Model Uncertainty in Text-as-Data Regressions, with an Application to Central Bank Communication

Martin Feldkircher, Bernd Prostmaier, Paul Hofmarcher, Pierre Siklos · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.19720 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Os cientistas sociais rotineiramente convertem o texto em covariantes como quotas de tópico, pontos ideais ou pontuações de sentimento, e os inserem em modelos estatísticos de interesse. Este projeto de dois passos é conveniente, mas raramente é guiado pela teoria: pouco diz ao analista quais covariantes baseados em texto para construir ou incluir. Cada abordagem incorpora escolhas de medição questionáveis, e operacionalizações concorrentes da mesma construção podem apoiar conclusões diferentes, de modo que uma única pergunta de pesquisa dá origem a um menu de modelos de regressão concorrentes. Quais covariantes construídos de texto pertencem a um modelo, e em que formalidade funcional, é, portanto, um problema de incerteza do modelo. Nós propomos modelos adicionais generalizados de ponta e fraca (SSGAM

Machine translation (Meta M2M100); the English original follows.

Abstract

Social scientists routinely convert text into covariates like topic shares, ideal points or sentiment scores, and insert them into statistical models of interest. This two-step design is convenient, but it is rarely guided by theory: little tells the analyst which text-based covariates to construct or include. Each approach embodies contestable measurement choices, and competing operationalizations of the same construct can support different conclusions, so a single research question gives rise to a menu of competing regression models. Which text constructed covariates belong in a model, and in what functional form, is thus a problem of model uncertainty. We propose spike-and-slab generalized additive models (SSGAMs), which jointly select covariates and determine whether their effects are linear or nonlinear. We illustrate the approach with the communication of a powerful unelected institution, asking which of 45 competing text indicators from 224 monthly European Central Bank press conferences, spanning late 2002 to mid-2025, track professional inflation expectations. A sparse set is selected; widely used alternatives, and every tested nonlinear effect, carry no incremental information.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Covariate, Construct (python library), Computer science, Set (abstract data type), Econometrics, Ideal (ethics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange