ESRA 2025 Preliminary Program
All time references are in CEST
Advances in in Survey Methodology: Accounting for Spatial and Temporal Information |
| Session Organisers |
Dr Veronica Ballerini (University of Florence)
Ms Lisa Braito (University of Florence)
|
| Time | Tuesday 15 July, 11:00 - 12:15 |
| Room |
Ruppert 042 |
This session aims to explore cutting-edge advancements in survey methodology, focusing on the integration of spatial and temporal dimensions to enhance data analysis. The discussion will cover a range of innovative approaches tailored to address the complexities of spatiotemporal dynamics within survey data, from longitudinal investigations to sequential designs and spatial sampling methodologies. The spotlight will be on advancements in spatially balanced sampling techniques, showcasing their wide-ranging applications in fields such as environmental studies, geology, biology, and agriculture. Particularly noteworthy is their efficacy in surveys where population characteristics are inherently geo-referenced, presenting a novel perspective on sampling strategies in these specialized domains. The session will also address the challenges associated with studying the longitudinal dimensions of socioeconomic phenomena. In scenarios where data stem from rotational panel designs and sample sizes may be insufficient for detailed analysis, navigating issues, such as poverty transitions, within sub-populations poses a significant hurdle. The discussion will shed light on strategies to overcome these obstacles and extract meaningful insights from limited sample sizes, ensuring a comprehensive understanding of dynamic socioeconomic trends.
The session is thought for hosting contributions of early career statisticians.
Keywords: Spatially balanced sampling, longitudinal studies, panel data, georeferenced data, sequential designs
Papers
An Investigation of the Compound Effect of Sampling Design and Estimator in Environmental Surveys
Dr Francesco Pantalone (Department of Social Statistics and Demography, University of Southampton) - Presenting Author
We present an investigation of the impact of the sampling strategy, defined as the couple sampling design and estimator, in environmental surveys. Literature has focused more on the gain in efficiency (in terms of, for example, mean squared errors) coming from either the sampling design or the estimator, but not on the compound effect (if any) of the two. This investigation provides a starting point on the analysis of the compound effect, which can be extended to the following research question: given the choice, would it be better to focus on the (spatial) sampling design, or the (spatial) estimator, or both? Results from Monte Carlo simulations are presented, where spatially balanced sampling designs and model-assisted estimators were employed, and generated and real data were used.
A Bayesian approach to uncover local and temporal determinants of heterogeneity in repeated cross-sectional health surveys
Dr Mattia Stival (Department of Economics, Ca' Foscari University of Venice) - Presenting Author
Dr Lorenzo Schiavon (Department of Economics, Ca' Foscari University of Venice)
Professor Stefano Campostrini (Department of Economics, Ca' Foscari University of Venice)
In several countries, including Italy, a prominent approach to population health surveillance involves conducting repeated cross-sectional surveys at short intervals of time. These surveys gather information on the health status of individual respondents, including details on their behaviours, risk factors, and relevant socio-demographic information. While the collected data undoubtedly provides valuable information, modelling such data presents several challenges. For instance, in health risk models, it is essential to consider behavioural information, local and temporal dynamics, and disease co-occurrence. In response to these challenges, our work proposes a multivariate temporal logistic model for chronic disease diagnoses at local level. Linear predictors are modelled using individual risk factor covariates and a latent individual propensity to diseases. Leveraging a state space formulation of the model, we construct a framework in which temporal heterogeneity in regression coefficients is informed by exogenous information at local level, corresponding to different contextual risk factors that may affect the occurrence of chronic diseases in different ways. To explore the utility and the effectiveness of our method, we analyse behavioural and risk factor surveillance data collected in Italy (PASSI), which is well-known as a country characterised by high peculiar administrative, social and territorial diversities reflected on high variability in morbidity among population subgroups.
A Monte Carlo Procedure for the Estimation of Species Coverage in Dunes
Dr Rosa Maria Di Biase (University of Siena) - Presenting Author
Dr Agnese Marcelli (University of Siena)
Assessing and monitoring natural resources and biodiversity is a pressing issue in environmental and ecological surveys. In this context, species coverage is a key element for evaluating ecosystem health, but it is often unfeasible to conduct complete censuses to acquire this information. Therefore, reliable statistical techniques for estimating species coverage are essential.
In a design-based inference framework, i.e., considering the population under study as a fixed set of locations with fixed values of the survey variable attached to each location, the sampling scheme adopted for placing sample sites across a continuum is fundamental for reliable inference. Moreover, due to the presence of positive spatial autocorrelation and of spatial heterogeneity, sampling schemes ensuring that the selected points are “well spread” across the study region should be adopted. Well spread or spatially balanced samples can be easily obtained by tessellation sampling schemes, such as tessellation stratified sampling and systematic grid sampling, or through more complex, ad hoc, sampling schemes.
The aim of this work is to demonstrate that species coverage in dunes, usually conducted by strip sampling, can be appropriately expressed as an integral and, thus, unbiasedly estimated by means of a Monte Carlo estimator, using different spatially balanced sampling schemes for placing the strips. Furthermore, the availability of remote sensed information retrieved from satellite images is exploited. Finally, a simulation study and an application to a real-world scenario are presented to confirm the theoretical findings.
Estimating Poverty Transitions from Repeated Cross-Sections: A Statistical Perspective
Dr Aldo Gardini (University of Bologna) - Presenting Author
Dr Riccardo D'Alberto (University of Verona)
Dr Silvia De Nicolò (University of Bologna)
Studying poverty transitions is challenging due to the limited availability of longitudinal survey data. To overcome this limitation, the literature proposes methodological approaches for estimating transition probabilities using cross-sectional data. These approaches can be grouped into two main strands: a parametric approach, which assumes a distributional model for income and leverages econometric techniques on pseudo-panels, and a semi-parametric approach, which employs matching procedures to construct synthetic panels. A critical step in both methodologies is estimating income correlation across time that is not directly available from cross-sectional data. Although these methods are popular in the economics literature, to the best of our knowledge, they have not been systematically reviewed or assessed from a statistical perspective. In this work, we first examine these methods and their limitations before introducing a novel scenario-based framework. This framework aligns with two alternative methodological proposals, each corresponding to one of the two strands of literature. Specifically, the parametric proposal incorporates Bayesian models that use scenarios as prior information for income autocorrelation, while the semi-parametric approach introduces a matching-based procedure. The latter includes a tuning parameter that adjusts the number of neighbors to regulate autocorrelation. Finally, we evaluate and compare the discussed methods, including our newly proposed ones, through a Monte Carlo simulation study based on Italian EU-SILC data.
