Unraveling Spatial Confounding
a comprehensive review and simulation-based evaluation of contemporary methods
Keywords:
Spatial confounding, causal inference, spatial regression models, simulation studyAbstract
Health, environmental, demographic, and other public data sets are typically aggregated (to administrative or geopolitical regions) to facilitate analysis and protect privacy. The use of standard regression models for spatially referenced data can result in spatial dependence in the residuals. For decades, the solution for this problem was to use spatial regression models. The usual setup for a spatial regression defines a set of areal-units with respective observations modeled as a combination of measured and unmeasured covariates. This work reviews current methods in causal inference literature to deal with confounding, using simulations to demonstrate promising results in mitigating unmeasured spatial confounding, paving the way for further exploration and application in real-world scenarios.
Downloads
References
S. Banerjee, B. P. Carlin e A. E. Gelfand. Hierarchical Modeling and Analysis for Spatial Data. 2nd ed. CRC Press, 2015.
E. Dupont, S. N. Wood e N. H. Augustin. “Spatial+: A novel approach to spatial confounding”. Em: Biometrics 78 (2020), pp. 1279–1290. url: https://api.semanticscholar.org/CorpusID:221819113.
B. A. Gilbert, A. Datta e E. L. Ogburn. “Approaches to spatial confounding in geostatistics”. Em: 2022. url: https://api.semanticscholar.org/CorpusID:245634816.
K. Khan e C. Berrett. Re-thinking Spatial Confounding in Spatial Linear Mixed Models. 2023. arXiv: 2301.05743 [stat.ME].
G. Papadogeorgou, C. Choirat e C. M. Zigler. “Adjusting for unmeasured spatial confounding with distance adjusted propensity score matching”. Em: Biostatistics 20.2 (jan. de 2018), pp. 256–272. issn: 1465-4644. doi: 10.1093/biostatistics/kxx074.
B. J. Reich, J. S. Hodges e V. Zadnik. “Effects of residual smoothing on the posterior of the fixed effects in disease-mapping models”. en. Em: Biometrics 62.4 (dez. de 2006), pp. 1197–1206.
B. J. Reich, S. Yang, Y. Guan, A. B. Giffin, M. J. Miller e A. Rappold. “A Review of Spatial Causal Inference Methods for Environmental and Epidemiological Applications”. Em: International Statistical Review 89.3 (2021), pp. 605–634. doi: https://doi.org/10.1111/insr.12452. url: https://onlinelibrary.wiley.com/doi/abs/10.1111/insr.12452.
