Advertisement

A Heavy-Tailed Multilevel Mixture Model for the Quick Count in the Mexican Elections of 2018

  • Michelle AnzarutEmail author
  • Luis Felipe González
  • María Teresa Ortiz
Conference paper
Part of the Springer Proceedings in Mathematics & Statistics book series (PROMS, volume 301)

Abstract

Quick counts based on probabilistic samples are powerful methods for monitoring election processes. However, the complete designed samples are rarely collected to publish the results in a timely manner. Hence, the results are announced using partial samples, which have biases associated to the arrival pattern of the information. In this paper, we present a Bayesian hierarchical model to produce estimates for the Mexican gubernatorial elections. The model considers the poll stations poststratified by demographic, geographic, and other covariates. As a result, it provides a principled means of controlling for biases associated to such covariates. We compare methods through simulation exercises and apply our proposal in the July 2018 elections for governor in certain states. Our studies find the proposal to be more robust than the classical ratio estimator and other estimators that have been used for this purpose.

Keywords

Bayesian calibration Hierarchical model Model-based inference Multilevel regression Poststratification Zero-inflated model 

References

  1. 1.
    COTECORA: Criterios científicos, logísticos y operativos para la realización de los conteos rápidos y protocolo de selección de las muestras (2018). https://repositoriodocumental.ine.mx/xmlui/bitstream/handle/123456789/96237/CGor201805-28-ap-27-a1.pdf
  2. 2.
    Gabry, J., Simpson, D., Vehtari, A., Betancourt, M., Gelman, A.: Visualization in Bayesian workflow (2017). arXiv preprint arXiv:1709.01449v5
  3. 3.
    Gelman, A., Hill, J.: Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press, Cambridge (2006)Google Scholar
  4. 4.
    Gelman, A., Stern, H.S., Carlin, J.B., Dunson, D.B., Vehtari, A., Rubin, D.B.: Bayesian Data Analysis. Chapman and Hall/CRC, United Kingdom (2013)Google Scholar
  5. 5.
    Ghosh, S.K., Mukhopadhyay, P., Lu, J.C.: Bayesian analysis of zero-inflated regression models. J. Stat. Plan. Inference 136(4), 1360–1375 (2006).  https://doi.org/10.1016/j.jspi.2004.10.008, http://www.sciencedirect.com/science/article/pii/S0378375804004008MathSciNetCrossRefGoogle Scholar
  6. 6.
    Little, R.J.: Calibrated Bayes, an alternative inferential paradigm for official statistics. J. Off. Stat. 28(3), 309 (2012)Google Scholar
  7. 7.
    Mendoza, M., Nieto-Barajas, L.E.: Quick counts in the Mexican presidential elections: a Bayesian approach. Electoral Stud. 43, 124–132 (2016)CrossRefGoogle Scholar
  8. 8.
    Ortiz, T.: quickcountmx: an R package to estimate election results from Mexico. https://github.com/tereom/quickcountmx (2018)
  9. 9.
    Park, D.K., Gelman, A., Bafumi, J.: Bayesian multilevel estimation with poststratification: state-level estimates from national polls. Polit. Anal. 12(4), 375–385 (2004)CrossRefGoogle Scholar
  10. 10.
    Pool, I.d.S., Abelson, R.P., Popkin, S.: Candidates, Issues, and Strategies. Cambridge, MA (1965)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Michelle Anzarut
    • 1
    Email author
  • Luis Felipe González
    • 1
  • María Teresa Ortiz
    • 1
  1. 1.Instituto Tecnológico Autónomo de MéxicoAltavistaMexico

Personalised recommendations