This study investigates the use of topic modeling techniques to analyze the industrial priorities embedded within Italian budget laws from 2020 to 2023. By employing four methods -Latent Semantic Analysis (LSA), Fuzzy Latent Semantic Analysis (fLSA), Latent Dirichlet Allocation (LDA), and Correlated Topic Model (CTM)- the analysis identifies coherent expenditure topics and assesses their relevance to fiscal and industrial strategies. Results indicate that, within the corpus examined here, fLSA and CTM provide more informative representations than LSA and LDA in terms of stability and complexity preservation, with fLSA offering more parsimonious solutions and CTM capturing more nuanced topic correlations. The findings illustrate how a comparative use of topic-modeling approaches can support the interpretation of policy priorities in legal and budgetary texts.

This study examines Italian-language content on the social media platform Mastodon, employing some typical tools for textual analysis. A set of posts (known as 'toots' on Mastodon) related to the hashtag #intelligenzaartificiale from the past two years were collected. Co-occurrence networks between hashtags and cross-author referrals within the same toots were analyzed. Next, structural topic modeling was utilized to identify four topics and their most important keywords. A comparison was then made between the predictive ability of a set of words and that of a transformer using a classifier, resulting in similar findings. Furthermore, a SHAP analysis was conducted to demonstrate the impact of individual words on the classification model, providing an explanation of the contribution of individual features.

This study examines public discourses on prenatal testing and newborn screening on Facebook among Italian and German users from 2013-2023. Using a computational content analysis technique - specifically Structural Topic Modelling - we identify distinct discursive healthscapes that illustrate how digital biocitizenship is articulated in the context of reproductive health issues. The analysis reveals both shared concerns and culturally specific engagements across the two contexts. Key similarities include framing reproductive health as individual responsibility and ambivalence toward biomedical authority. Notable differences emerge in the politicisation of healthcare debates, approaches to de-medicalization, and the role of popular culture in shaping perceptions. The Italian corpus emphasises policy advocacy and critiques of healthcare commodification, while the German data focuses more on affective solidarity and awareness campaigns. Both contexts reflect tensions between demands for expanded screening and resistance to over-medicalization. Digital biocitizenship manifests differently when driven by organised advocacy versus individualised engagement. Overall, the findings illustrate how social media discourses on prenatal and newborn screening reflect broader sociopolitical tensions, while highlighting divergent pathways through which individuals navigate biomedical uncertainties by blending expert and experiential knowledge. This study contributes to understanding how reproductive health issues are constructed and contested in digital public spheres.

This article addresses the issue of science to contribute to the ongoing theoretical-methodological debate in the Social Representations Theory (SRT). It investigates the idea that attitudes are integral to social representations (SRs) and proposes solutions for empirically examining this nested relationship, expanding the technique of free associations in SRT. It examines the SR of science, considering diverse patterns of relationship with it based on literacy, interest, trust, and attitude. The study, utilising a purpose-built questionnaire with a sample of Italians, reveals a somewhat stereotypical SR of science rooted in a traditional view. While predominantly positive, distinct positions emerge in relation to the different patterns of relationship with science. Four patterns emerge, aligning along the dimensions of acceptance (affective-oriented) and commitment (cognitive/behavioural-oriented) and contributing to constructing a multifaceted and nuanced SR of science. The study corroborates the idea that attitude is part of SRs, and methodological developments, particularly assessing associations’ valence, prove effective. The comparison between freely evoked and ranked associations reveals reordering dynamics influenced by patterns of relationships with science. These methodological proposals offer a comprehensive understanding of SRs, allowing for longitudinal reflections on their trajectories and co-construction mechanisms.

The article aims to understand the process through which scientific experts gain and maintain remarkable media visibility. It has been analysed a corpus of 213,875 articles published by the eight most important Italian newspapers across the Covid-19 pandemic in 2020 and 2021. By exploring this process along the different phases of the management of the emergency in Italy, it was observed that some scientific experts achieve high media visibility-and sometimes notwithstanding their low academic reputation-thus becoming a sort of "media star". Scientific literature about the relationship between experts and media is considerable, nonetheless we found a lack of theoretical models able to analyse under which conditions experts are able to enter and to remain prominent in the media sphere. A Media Experts Evolutionary Model (MEEM) is proposed in order to analyze the main conditions under which experts can acquire visibility and how they can "survive" in media arena. We proceeded by analysing visibility of experts during SARS-CoV-2 pandemic and considering both their individual credentials previously acquired and the media environment processes of selection; MEEM acts hence as a combination of these two levels. Regarding the credentials, we accounted for i) institutional role/position, ii) previous media visibility, and iii) matches between scientific credentials and media competence. In our analysis, we collected evidence that high visibility in newspapers can be seen as evolutionary in the sense that some profiles-i.e. a particular configuration of credentials-are more adapt to specific media environments.

This paper focuses on the relationship between financial analysts' recommendations and press sentiment from the perspective of the attention-grabbing theory. Specifically, attention-grabbing should not be enough to explain the effect that media coverage has on investment decisions, since investors are wary of making a mistake and anticipate the regret of a future loss. Our case study pertains to a column reporting on secondhand information and analysts' recommendations. Once the column did not report the analysts' advice anymore, we hypothesized investors also assess the sentiment of the column to make sure they are not making a costly mistake. Event studies on abnormal returns and multivariate analyses show that for columns with explicit analysts' recommendations the attention-grabbing mechanism directs buying decisions while has no influences on selling decision. In the absence of explicit recommendations, investors transform the columns' content into implicit recommendations leading their buying decisions when the sentiment is highly positive.

This dataset includes metadata of the newspaper articles used for the paper "When scientific experts come to be media stars: an evolutionary model tested by analysing coronavirus media coverage across Italian newspapers". The dataset is in JSON format. The metadata includes: "uuid" (unique identifier we associated to an article), "URLs" (the URLs where the article was published), "sources" (newspaper and feed/section where the article was published), "datesPublished" (dates when the article was published/updated). License: Attribution-ShareAlike 4.0 International (https://creativecommons.org/licenses/by-sa/4.0/legalcode)

The aim of this paper is to study the role of citation network measures in the assessment of scientific maturity. Referring to the case of the Italian national scientific qualification (ASN), we investigate if there is a relationship between citation network indices and the results of the researchers’ evaluation procedures. In particular, we want to understand if network measures can enhance the prediction accuracy of the results of the evaluation procedures beyond basic performance indices. Moreover, we want to highlight which citation network indices prove to be more relevant in explaining the ASN results, and if quantitative indices used in the citation-based disciplines assessment can replace the citation network measures in non-citation-based disciplines. Data concerning Statistics and Computer Science disciplines are collected from different sources (ASN, Italian Ministry of University and Research, and Scopus) and processed in order to calculate the citation-based measures used in this study. Then, we apply logistic regression models to estimate the effects of network variables. We find that network measures are strongly related to the results of the ASN and significantly improve the explanatory power of the models, especially for the research fields of Statistics. Additionally, citation networks in the specific sub-disciplines are far more relevant than those in the general disciplines. Finally, results show that the citation network measures are not a substitute of the citation-based bibliometric indices.

This study examined the social representation (SR) of nanotechnologies and its relationships with those of science and technology. It aimed to understand the role of pre-existing and neighbouring forms of shared knowledge in orienting the way laypeople autonomously develop ideas about an unfamiliar issue, and related implications for its perceived risks and public acceptance or rejection. The study involved 489 Italian participants, stratified according to gender, age and education level. They completed an ad-hoc questionnaire with multiple free association tasks to the word-stimuli nanotechnologies, science and technology, and close-ended questions to gauge 'familiarity' levels with nanotechnologies and 'engagement' in nanotechnologies through media. The results suggested the presence of a rather shared and organised SR of nanotechnologies whose content is cautiously enthusiastic. The biomedical domain seems to be the most easily accepted field of application for nanotechnologies, indicating a 'preferential channel' through which they could be more welcome and trusted. Concerning inter-representation relationships, comparisons among SRs identified a strong connection among these three objects, indicating the existence of a coherent representation system where the SR of nanotechnologies is nested in those of science and technology, which, in turn, are in a reciprocal relationship. Fine-grained comparisons allowed for grasping further insights. The results showed that the SR of nanotechnologies is the least enthusiastic, highlighting critical voices among laypeople, although peripherally. Specifically, it presents elements of perceived risks and rejection on one hand, and elements indicating a detached approach in which individuals struggle to detect potentialities and advantages on the other, contrary to what has been found for science and technology at large.

Purpose This paper focuses on the influence of social, cultural and religious factors on investors' attention. In particular, the authors examined if the attention-grabbing mechanism works on Sundays, that is, if the Italians' Sunday activities and habits lead to a lower attention to second-hand financial news, compared to Saturdays. Design/methodology/approach The authors analyzed the market reaction to equivalent stale events published on the Saturday and Sunday editions of an Italian financial newspaper and conducted a standard event study on abnormal returns and abnormal volumes for Saturday and Sunday columns and a multivariate analysis on abnormal returns for columns reporting positive recommendations. As a robustness check, the authors performed a sentiment analysis of the columns and included this variable in the regression analysis, but sentiment proved to be not significant in the final model. Findings The study’s results confirmed that the attention-grabbing mechanism directed buying decisions, while had no influence on selling decisions. Furthermore, event study and multivariate analysis showed a significant lower market reaction to Sunday columns, supporting the study hypothesis of a Sunday investors' inattention which can be traced to cultural and/or religious factors since Sunday in Italy is a day devoted to family, entertainment and religious rituals. Practical implications The lower investors' attention on Sundays and the related influence of social, cultural and religious factors have implications for the timing of both corporate communications and financial advertising. Originality/value The authors’ paper provides an original contribution, on the empirical ground, to the attention-grabbing theory and to the growing theoretical literature in microeconomics that models attention.

"SARS-CoV-2 general corpus" includes Italian newspaper articles published in the timespan between January 1, 2020 and June 15, 2020, containing at least one of the following terms: [covid, corona virus, OR coronavirus]. Sources: Corriere della Sera, La Repubblica, Il Sole – 24 Ore, La Stampa, Avvenire, Il Giornale, Il Mattino di Napoli, Il Messaggero.

. In this paper we describe the participation of the WordUp! team in the VaxxStance shared task at IberLEF 2021. The goal of the competition is to determine the author’s stance from tweets written both in Spanish and Basque on the topic of the Antivaxxers movement. Our approach, in the four di(cid:27)erent tracks proposed, combines the Logistic Regression classi(cid:28)er with diverse groups of features: stylistic, tweet-based, user-based, lexicon-based, dependency-based, and network-based. The outcomes of our experiments are in line with state-of-the-art results on other languages, proving the e(cid:30)cacy of combining methods derived from NLP and Network Science for detecting stance in Spanish and Basque.

The SARS-CoV-2 pandemic has emerged as one of the most dramatic health crises of recent decades. This paper treats mainstream news about the current pandemic as a valuable entry point for analyzing the relationship between science and politics in the public sphere, where the outbreak must be both understood and confronted through appropriate public-health policy decisions. In doing so, the paper aims to examine which actors, institutions, and experts dominate the SARS-CoV-2 media narratives, with particular attention to the roles of political, medical, and scientific actors and institutions within the pandemic crisis. The study relies on a large dataset consisting of all SARS-CoV-2 articles published by eight major Italian national newspapers between January 1, 2020 and June 15, 2020. These articles underwent a quantitative analysis based on a topic modeling technique. The topic modeling outputs were further analyzed by innovatively combining ad-hoc metrics and a classifier based on the stacking ensemble method (combining regularized logistic regression and linear stochastic gradient descent) for quantifying scientific salience. This enabled the identification of relevant topics and the analysis of the roles that different actors and institutions engaged in making sense of the pandemic. The results show how the health emergency has been addressed primarily in terms of political regulation and concerns and only marginally as a scientific matter. Hence, science has been overwhelmed by politics, which, in media narratives, exerts a moral as well as regulatory authority. Media narratives exclude neither scientific issues nor scientific experts; rather, they configure them as a subsidiary body of knowledge and expertise to be mobilized as an ancillary, impersonal institution useful for legitimizing the expansion of political jurisdiction over the governance of the emergency.

In this paper we aim to analyze the Italian social media communication about COVID-19 through a Twitter dataset collected in two months. The text corpus had been studied in terms of sensitivity to the social changes that are affecting people's lives in this crisis. In addition, the results of a sentiment analysis performed by two lexicons were compared and word embedding vectors were created from the available plain texts. Following we tested the informative effectiveness of word embeddings and compared them to a bag-of-words approach in terms of text classification accuracy. First results showed a certain potential of these textual data in the description of the different phases of the outbreak. However, a different strategy is needed for a more reliable sentiment labeling, as the results proposed by the two lexicons were discordant. Finally, although presenting interesting results in terms of semantic similarity, word embeddings did not show a predictive ability higher than the frequency vectors of the terms.

In this contribution we describe the system (i.e. a statistical model) used to participate in Evalita conference 2020, SardiStance (Tasks A and B) and Haspeede2 (Tasks A and B). We first developed a classifier by extracting features from the texts and the social network of users. Then, we fit the data through an extreme gradient boosting, with cross-validation tuning of the hyper-parameters. A key factor for a good performance in SardiStance Task B was the features extraction by using Multidimensional Scaling of the distance matrix (minimum path, undirected graph) applied on each network. The second system exploits the same features above, but it trains and performs predictions in twosteps. The performances proved to be lower than those of the single-step model.

In 2015, the United Nation General Assembly adopted the 2030 Agenda for Sustainable Development and its 17 Sustainable Development Goals aiming at ending all forms of poverty, fighting inequalities, and tackling climate change. We collected Twitter data about the 2030 Agenda from May 9th to November 9th, 2018. The aim of this work is to obtain a classification of each tweet in the corpus according to the "Information"-"Action" categories, in order to detect whether a tweet refers to an event or it has only an informative-disclosure purpose. It seems particularly interesting to understand how and to what extent people and organizations are playing a more active role in shaping the process of responding locally and internationally to climate change. Explicit intention to act or inform had been captured by hand coding of a randomly selected sample of tweets and then the classification had been extended to the whole corpus through a supervised machine learning method. Overall, our classification supervised model has produced satisfactory results.

Recent studies indicate that Internet skills have a positive impact on academic achievement. This article presents a national study that seeks to validate an Internet skills scale that was already tested in other EU countries (the Netherlands and the United Kingdom) to understand the competence level of the population as a whole as well as across population sectors. The scale questionnaire was completed by a sample of the Italian population stratified by gender, age and geographical area. The result is globally consistent at the empirical level as well as at the cross-national level. All the five scales showed excellent internal consistency.