I checked 7 public opinion journals on Wednesday, October 14, 2026 using the Crossref API. For the period October 07 to October 13, I found 14 new paper(s) in 4 journal(s).

Journal of Official Statistics

Applying Machine Learning to Longitudinal Administrative Data: A Case Study in Education
Fabrizio De Fausti, Marco Di Zio, Romina Filippini, Simona Toti
Full text
This paper presents the results of an extensive empirical evaluation study of several machine learning methods used to predict the attained level of education (ALE) in Italy based exclusively on longitudinal administrative data. The assessment compares model predictions with observed ALE values available in administrative sources, evaluating both distributional preservation and unit-level prediction accuracy. Our findings show that machine learning techniques can effectively exploit longitudinal administrative information to generate accurate and coherent ALE estimates, providing valuable support for the production of register-based statistics.
Governing Large Language Model Services for Confidential Microdata: A Decision Framework for National Statistical Offices
Hanan Ather, Mohammed Haddou, Saeid Molladavoudi
Full text
National statistical offices increasingly face requests to provide large language model (LLM) capabilities in secure research environments. The governance problem is to determine which service configurations, if any, can be supported safely on confidential microdata. These services challenge traditional statistical disclosure control because interaction is adaptive, systems may retrieve documents or call analytical tools, and outputs may appear plausible without being grounded in approved data. This paper proposes a governance framework that organizes each request as a sequence of decisions: whether an LLM is needed, which service mode is justified, which threat model applies, which controls are required, and what evidence must be reviewed before release. The framework spans deterministic pipelines, retrieval-augmented generation (RAG), tool-using agents, code-execution agents, linkage configurations, and fine-tuning.
A Framework for Using Metadata to Combine Data Sets in Official Statistics
Yvonne A. P. M. Gootzen, Arnout van Delden, Piet J. H. Daas
Full text
There are increasing numbers of sources available for the production of official statistics. When one is interested in making new output or redesigning the input that is used for existing output, it might not always be clear if and how the intended output can be produced from the variety of input data sets available. The current paper proposes a metadata-based framework with a search algorithm that finds out if a processing path is possible from input data sets to intended output. The framework uses the conceptual metadata of the data set as input, that is, the unit types, the identifier variables, the measurement variables and the reference periods. Our contribution to existing literature is threefold. We provide a set of basic operators, each representing a small, clearly defined step in the process of combining data sets; we can accommodate tailor-made operators (models) to cover a wide range of data configurations, and we provide an implementation in Python code to automatically answer if and how the intended output can be produced from a set of available input data sets. The framework is evaluated on a case study. We provide publicly available code and different examples on Git to facilitate the use and further development of our framework.

Journal of Survey Statistics and Methodology

Sampling with Double Spatial Balance
Blair L Robertson, Chris J Price, Marco Reale, Philip Davies
Full text
Spatial sampling designs determine where response variables are measured in a study area to estimate population parameters precisely. Spatially balanced designs yield precise results for commonly used estimators when positive spatial associations exist. A doubly-balanced design has two balancing properties: spatial balance on spatial variables and balance on auxiliary variables. Doubly-balanced designs tend to be efficient when the response is a linear function of the auxiliary variables with spatially correlated error terms. This article presents a novel sampling objective, where spatial balance is achieved simultaneously in the spaces spanned by the spatial and auxiliary variables. Numerical results show that our double spatial balance method works for both linear and non-linear superpopulation models, making it an attractive alternative to spatially balanced, balanced, or doubly-balanced designs. We provide example applications with equal and unequal probabilities for estimating single and multiple response totals. A design-based variance estimator is also given.
Spatially Balanced Sampling in Real Time
Wilmer Prentius, Fabio Carrer, Anton Grafström, Juha Heikkinen
Full text
Probability sampling in real time or streaming scenarios presents unique challenges, especially when the population size and the spatial locations of units are unknown a priori. Standard methods like Poisson sampling can be applied but often result in spatially clustered samples that are not representative of the population, leading to inefficient estimation. We introduce a novel adaptation of spatially correlated Poisson sampling (SCPS) designed specifically for real-time applications. The proposed algorithm sequentially decides on the inclusion of newly encountered units by dynamically estimating the required spatial neighborhood for weight distribution, using only information from previously observed units and a running estimate of the total population size. Through Monte Carlo simulations on both uniform and clustered populations, we demonstrate that the method greatly improves spatial balance compared to conventional Poisson sampling. Furthermore, it substantially reduces the variability of the achieved sample size, offering more predictable survey costs and effort. This approach provides a practical and efficient tool for obtaining spatially balanced samples from populations where a complete sampling frame is unavailable in advance.

Politics, Groups, and Identities

From coverage to context: how race, sanctuary policies, and media shape immigrant perceptions of police
Diego Esparza, Andrea Silva
Full text
The role of group-based attitudes in policy evaluation: the relationship between sexism, racism, and attitudes toward the affordable care act
Mona S. Kleinberg, Katherine T. McCabe
Full text
Cross-party perspective getting, identity, and support for electoral fairness: who can persuade?
Ryan Strickler
Full text
Wrestling with perceptions: how WWE live events in Saudi Arabia shaped views of Muslims in the United States
Rudy Alamillo
Full text
Evolution by national election? Partisanship, voting behavior, and sentiments toward culture war groups
John V. Kane, Srivardhan Jangili, Joseph Sandor, Vitória Sgorlon
Full text
Pressures of place: urban–rural variation in considerations for political participation
Caroline High Pippert, Jennifer Lin, Kristin Lunz Trujillo
Full text
Ethnocultural identity and candidate self-presentation in Canadian elections
Daniel Westlake, Jacob Robbins-Kanter
Full text

Social Science Computer Review

Who Is Afraid of ChatGPT? A Natural Experiment on the Effect of ChatGPT Release on Labour-Market Concerns
Francesco Nicoli, Gregorio Buzzelli, Brian Burgoon, Stefano Sacchi
Full text
The public release of ChatGPT on November 30th, 2022 revived the public debate on technological change and, particularly, on its impact on the labour market. Despite the sudden growth in AI salience after the launch of the new chatbot, to our knowledge, no empirical investigation has assessed the impact of this technological breakthrough on citizens’ first concerns. By means of a unique survey fielded between November and December 2022 in five European countries, we use an Unexpected Event During Survey (UEDS) design to causally test the effect of ChatGPT’s launch on individual perception of unemployment risk. We rely on statistical matching techniques, combined with regression analyses to explore heterogeneity in responses between socioeconomic groups. We show that the release of the new chatbot mildly but significantly increased unemployment concerns in advanced industrial democracies, particularly among individuals with low and middling levels of income and education, as well as poorly exposed to AI. By capturing reactions before informational framings consolidate and widespread adoption occurs, this study provides a rare window into citizens’ genuine response to technological change, indicating that its socioeconomic consequences are perceived as immediately salient.
Role-Playing as Immigrants for Learning to Trust: Constructing a Theory of Experiential Contact
Brandon C. Bouchillon
Full text
This research examined whether a computer-based role-playing game in which participants created immigrant characters from Mexico and took them through the process of naturalizing to the United States could be used to promote trust. A posttest-only experiment with a control group was conducted in Fall 2024 ( N = 98). Results indicate that participants who played the game had significantly higher levels of trust in immigrants in general, as well as those from Mexico, Central/South America, India, China, and the Middle East specifically, compared to the control condition. Parasocial connections to characters in game-based settings strengthened the effect of role-playing on trusting immigrants from Mexico. Half of the role-playing participants also saw a thank-you video from an outgroup representative, as a form of mediated indirect contact. The video condition appeared to drive the interaction between role-playing and sensing parasocial connections to characters in predicting Mexican immigrant trust. Results demonstrate the potential for role-playing games to foster trust between groups who may struggle to interact in person, along with the additive effect of mediated indirect contact at high levels of parasocial connection.