Woska Pires da Costa, Cristhian Chagas Ribeiro, Marcos de Moraes Sousa, Cleiton Mateus Sousa, Priscilla Rayanne e Silva, Kodhandaraman Chinnathambi, Manuel Monfort-Pañego, Matias Noll
Decisions to enter funding competitions and subsequent funding outcomes are connected components of a broader process. However, evidence on how researcher-profile characteristics relate to these stages remains limited in public research institutions outside the research-intensive university sector, particularly in emerging-economy research systems. This exploratory case study analyzed the associations between researcher-profile characteristics and two respondent-level outcomes among researchers at a Brazilian Federal Institute: proposal submission and, among submitters, self-reported approval of at least one project. An online cross-sectional survey was completed by 83 of 327 eligible researchers (response rate 25.4%) in 2023. Data from this self-selected, single-institution sample were analyzed using bivariate contingency-table tests, exact tests for sparse tables, and Cramér's V as a descriptive effect-size measure. Findings were treated as hypothesis-generating. Proposal submission was associated with education level (p = 0.020; V = 0.285) and teaching in a master's or doctoral degree program (p < 0.001; V = 0.409). Among respondents who submitted proposals, reporting at least one approved project was associated with age group (p = 0.011; V = 0.403) and CNPq productivity fellowship (PQ) status (p = 0.047; V = 0.284). The approval outcome was measured at the respondent level and did not represent a proposal-level success rate. By distinguishing submission from approval, this study contributes to research evaluation by indicating that proposal submission and reporting at least one approved project among submitters were associated with different researcher-profile characteristics. These exploratory findings do not establish barriers or causal mechanisms. Multicenter, proposal-level, and longitudinal studies using explicit, context-sensitive definitions of career stage should test whether these patterns vary across institutions, fields, and career trajectories.
David Arratea-Pillco, José Gabriel Seijas-Díaz, Enrique Alejandro Barbachan-Ruales, Julio Iván Torres-Manrique, Italo W. Alejos-Patiño, Marco Arratea-Pillco
Aggregate patent counts provide information on filing volume but do not reveal whether a national filing corpus is technologically concentrated, diversified, or structurally reordered over time. This study examines the International Patent Classification (IPC) composition of Peru's patent and utility model filing corpus and distinguishes distributional diversification from structural reordering. We analyzed 32,966 IPC-classified patent and utility model applications filed before INDECOPI between 2000 and 2024, distributed across 131 primary IPC classes. An integrated descriptive patent analytics framework combined concentration, diversity, dominance, rank persistence, set overlap, distributional distance, cumulative hierarchy, legal-instrument robustness, and applicant-origin contextual indicators. The filing distribution became less concentrated and more even over time. Across predefined 5-year windows, mean annual HHI decreased from 0.220 in 2000–2004 to 0.072 in 2020–2024, while mean Shannon entropy increased from 3.670 to 5.022 bits. Mean Top-1, Top-3, and Top-10 shares also declined. Nevertheless, A61, C07, and A01 together accounted for 49.5% of the 2000–2024 panel. Spearman correlations across consecutive periods ranged from 0.887 to 0.941 in the fixed 131-class universe, while robustness checks using alternative class universes supported the same general interpretation, although with lower coefficients. Instrument-stratified analyses indicated that the diversification pattern was not solely attributable to combining patents and utility models. The results indicate diversification without full structural reordering. The study characterizes an office-based jurisdictional filing corpus before INDECOPI and should not be interpreted as direct evidence of domestic technological capability, patent quality, or innovation performance. The framework provides a reproducible approach for distinguishing broader IPC-classified filing participation from deeper changes in technological hierarchy.
Jannat, Charlie Dil, Tom Arodz, Bridget T. McInnes
The rapid expansion of biomedical literature requires automated methods for accurate and efficient information extraction. This study addresses relation classification: given a pair of annotated biomedical entities in a research article title and abstract, assigning the relation that holds between them from a pre-defined set. We evaluate a retrieval-augmented generation (RAG) framework in which semantically similar annotated examples are retrieved from a training pool and supplied to a large language model as in-context demonstrations. Four open-source LLMs are compared against a supervised DeBERTa+CNN baseline on GutBrainIE, and four further corpora (ChemProt, BioRED, its BioCreative VIII release BC8, and DDI-2013) are used to test whether the findings generalize. We report macro- and micro-averaged precision, recall and F1-score, and the rate of predictions that fall outside the defined relation set. On GutBrainIE, Qwen3:14B with 25 retrieved examples reaches 0.877 macro and 0.866 micro F1, against 0.440 and 0.562 for the supervised baseline, while the smaller Llama-3.1-8B and Gemma3:12B models do not exceed the baseline on macro F1. Macro F1 rises from 0.105 without demonstrations to 0.877 with 25, with the steepest gains over the first ten retrieved examples and subsequent variation smaller than the run-to-run standard deviation. Predictions falling outside the defined relation set drop from 39 at zero shots to at most one between 20 and 35, so retrieval improves output validity as well as accuracy. Across the four further corpora these gains do not generalize uniformly: the best configuration exceeds the supervised baseline on ChemProt, improves macro but not micro F1 on BioRED and BC8, and falls below it on both measures on DDI-2013. Since DDI-2013 is the most balanced corpus in the set and, like ChemProt, is annotated within single sentences, that reversal is not explained by class imbalance or annotation scope, but is consistent with what the relation labels encode: whether they name relationships recoverable from biomedical knowledge, or boundaries set by an annotation guideline.
Piotr Religa, Michel-Edwar Mickael, Marzena Łazarczyk, Norwin Kubick, Jarosław Olav Horbańczuk, Biniyam Tsegaye, Mariusz Sacharczuk, Atanas G. Atanasov
Several social media platforms provide a route for communication about scientific topics. X, formerly known as Twitter, has been shown to offer researchers an opportunity to exchange scientific information with audiences extending well beyond the research community. In recent years another platform, Bluesky, has emerged and appears to encourage scientific discourse, and a wave of migration of researchers from X to Bluesky has been documented in survey data, poll responses and altmetric records. Whether scientific claims are described differently on the two platforms has not been fully studied. We collected 8,697 Bluesky posts and 3,48,900 X posts mentioning the term science over a matched observation window, restricted in both cases to English-language text written by the posting account. A combination of data analysis and natural language processing tools was then used to compare the discourse on the two platforms, quantifying disciplinary composition, field concentration, cross-disciplinary reference and linguistic register, with effect sizes reported alongside every test and correction applied for multiple comparisons. The two platforms carried posts across a similar range of science-related topics, including health and medicine, biology and life sciences, and engineering and technology. However, posts on X were easier to read (Flesch–Kincaid grade 9.07 against 10.02), addressed the reader more directly (second-person pronouns 0.0153 against 0.0093) and were less complex (complex word density 0.092 against 0.108). Posts on Bluesky were longer (30.6 against 28.7 words; 15.98 against 14.54 words per sentence), more concentrated within single disciplines (mean field isolation 0.6882 against 0.6200), higher in first-person pronoun use (0.0255 against 0.0208) and more often classified as academic (16.6% against 12.7%). These patterns are consistent with greater peer-oriented scientific discourse on Bluesky and greater audience-oriented dissemination on X, although audience composition was not measured directly.
Measurement-model specification remains a persistent source of conceptual and statistical error in social science research. Existing guidance distinguishes reflective, causal-indicator/formative, hybrid, higher-order, and composite models, yet applied studies often select among them through disciplinary convention, software defaults, or post hoc empirical performance. This conceptual article reframes the problem as construct–measurement alignment and develops an explanatory and diagnostic framework rather than another binary decision rule. A structured conceptual synthesis was used to compare foundational and contemporary literature across five domains: construct ontology, indicator role, formal specification, validation logic, and contextual interpretation. The analysis integrates common-factor CFA and SEM, MIMIC models, higher-order structures, and composite-based SEM, with explicit attention to identification, measurement error, and empirically equivalent models. From inconsistencies across these five domains, the framework derives a taxonomy of seven recurrent forms of misalignment: ontological–structural, indicator–construct, epistemological–statistical, conceptual–identification, error–specification, level-of-abstraction, and context–interpretation misalignment. Three worked applications demonstrate that the same construct label can warrant different model specifications when its definition, level, purpose, or indicator role changes. The framework therefore explains why measurement decisions fail, identifies where misalignment occurs, and specifies the evidence required for diagnosis and correction. It offers researchers, reviewers, and editors a transparent basis for model justification while limiting claims that statistical fit alone can establish construct meaning or causal direction.
