Cuando la variabilidad varía: Heterocedasticidad y funciones de varianza

Autores/as

  • Facundo J. Oddi Universidad Nacional de Río Negro. Instituto de Investigaciones en Recursos Naturales, Agroecología y Desarrollo Rural. San Carlos de Bariloche, Río Negro, Argentina. Consejo Nacional de investigaciones Científicas y Técnicas (CONICET). Instituto de Investigaciones en Recursos Naturales, Agroecología y Desarrollo Rural. San Carlos de Bariloche, Río Negro, Argentina.
  • Fernando E. Miguez Department of Agronomy, Iowa State University, Ames, IA, USA.
  • Guido G. Benedetti Universidad Nacional de Río Negro. Instituto de Investigaciones en Recursos Naturales, Agroecología y Desarrollo Rural. San Carlos de Bariloche, Río Negro, Argentina.
  • Lucas A. Garibaldi Universidad Nacional de Río Negro. Instituto de Investigaciones en Recursos Naturales, Agroecología y Desarrollo Rural. San Carlos de Bariloche, Río Negro, Argentina. Consejo Nacional de investigaciones Científicas y Técnicas (CONICET). Instituto de Investigaciones en Recursos Naturales, Agroecología y Desarrollo Rural. San Carlos de Bariloche, Río Negro, Argentina.

DOI:

https://doi.org/10.25260/EA.20.30.3.0.1131

Palabras clave:

modelos lineales generales, mínimos cuadrados generalizados, selección de modelos, modelos anidados, criterios de información, AIC, función gls, R

Resumen

La variabilidad es una característica inherente al mundo que nos rodea. Cuantificarla es clave para comprender muchos de los procesos de interés para las ciencias ambientales y sociales como, por ejemplo, la adaptación de las especies al cambio climático o la desigualdad social. Para cuantificar la variabilidad se suele usar la varianza, uno de los parámetros de la distribución normal. Sin embargo, los modelos lineales clásicos asumen que la varianza es constante (supuesto de homocedasticidad) y se preocupan sólo por los cambios en las tendencias promedio. Es posible extender los modelos clásicos y relajar el supuesto de homocedasticidad mediante funciones de varianza, muy poco difundidas y abordadas por los textos en español. En esta ayuda didáctica nos proponemos introducir las funciones de varianza en modelos lineales desde un enfoque teóricoaplicado. Comenzamos introduciendo un problema real en el que se espera que la varianza no sea constante, y lo acompañamos con un ejemplo simulado. Posteriormente, planteamos el modelo lineal clásico y discutimos cómo se lo puede extender para modelar la heterocedasticidad. A continuación, explicamos algunas de las funciones de varianza y las aplicamos al caso real y a los datos simulados. Para ello hacemos uso de la función gls() del paquete nlme de R y proveemos el código para la reproducción del análisis. También exponemos otras opciones disponibles en R para tratar con datos heterocedásticos. Esperamos que este artículo brinde las bases para que profesionales y científicos con conocimientos estadísticos básicos comiencen a utilizar funciones de varianza y amplíen el conjunto de herramientas para analizar sus datos.

Citas

Bolker, B. 2008. Ecological models and data. Princeton University Press, Princeton.

Brazzale, A. R. 2005. hoa: An R package bundle for higher order likelihood inference. Rnews, 5/1 May 2005, 20-27. ISSN 609-3631. URL: www.r-project.org/doc/Rnews/Rnews_2005-1.pdf.

Dehn, J. 2000. The effects on growth of commodity price uncertainty and shocks. Policy Research Working Paper No. 2455. World Bank, Washington, DC. https://doi.org/10.1596/1813-9450-2455.

Fanelli, J. 2008. Macroeconomic volatility, institutions and financial architectures: the developing world experience. Palgrave Macmillan, London. https://doi.org/10.1057/9780230590182.

Garibaldi, L. A., F. Aristimuño, F. Oddi, and F. Tiribelli. 2017. Inferencia multimodelo en ciencias sociales y ambientales. Ecología Austral 27:348-363. https://doi.org/10.25260/EA.17.27.3.0.513.

Garibaldi, L. A., F. J. Oddi, F. Aristimuño, and A. N. Behnisch. 2019. Modelos estadísticos en lenguaje R. Editorial UNRN, Viedma.

Gałecki, A., and T. Burzykowski. 2013. Linear Mixed-Effects Models Using R. A Step-by-Step Approach. Springer Texts in Statistics, Springer Science+Business Media, New York. https://doi.org/10.1007/978-1-4614-3900-4.

Hallgrímsson, B., and B. Hall. 2005. Variation: a Central Concept in Biology. Elsevier Academic Press, Boston.

Howell, R. T., and C. J. Howell. 2008. The relation of economic status to subjective well-being in developing countries: A meta-analysis. Psychological Bulletin 134:536-560. https://doi.org/10.1037/0033-2909.134.4.536.

Kose, M. A., E. S. Prasad, and M. E. Terrones. 2003. Financial Integration and Macroeconomic Volatility. IMF Staff Papers 50(1). https://doi.org/10.5089/9781451846997.001.

Li, H., L. Squire, and H. F. Zou. 1998. Explaining international and intertemporal variations in income inequality. The Economic Journal 108:26-43. https://doi.org/10.1111/1468-0297.00271.

Loayza, N., and V. V. Hnatkovska. 2004. Volatility and Growth. Policy Research Working Paper No. 3184. World Bank, Washington, DC. https://doi.org/10.1596/1813-9450-3184.

Møller, A. P., and M. D. Jennions. 2002. How much variance can be explained by ecologists and evolutionary biologists? Oecologia 132:492-500. https://doi.org/10.1007/s00442-002-0952-2.

Nelder, J. A., and Wedderburn R. W. M. 1972. Generalized Linear Models. Journal of the Royal Statistical Society, Series A 135:370-384. https://doi.org/10.2307/2344614.

Neyman, J., and E. L. Scott. 1960. Correction for bias introduced by a transformation of variables. The Annals of Mathematical Statistics 31:643-655. https://doi.org/10.1214/aoms/1177705791.

Oddi, F. J., F. J. Aristimuño, C. Coulin, and L. A. Garibaldi. 2018. Ambigüedades en términos científicos: sobre el uso del “error” y el “sesgo” en estadística. Ecología Austral 28:525-536. https://doi.org/10.25260/EA.18.28.3.0.680.

Oddi, F. J., F. Miguez, L. Ghermandi, L. O. Bianchi, and L. A. Garibaldi. 2019. A nonlinear mixed-effects modelling approach for ecological data: Using temporal dynamics of vegetation moisture as an example. Ecology and Evolution 9:10225-10240. https://doi.org/10.1002/ece3.5543.

O’Hara, R. B., and D. J. Kotze. 2010. Do not log-transform count data. Methods in Ecology and Evolution 1:118-122. https://doi.org/10.1111/j.2041-210X.2010.00021.x.

Pinheiro, J. C., and D. M. Bates. 2000. Mixed‐effects models in S and SPLUS. Springer‐Verlag, New York. https://doi.org/10.1007/978-1-4419-0318-1.

Pinheiro, J. C., D. M. Bates, S. DebRoy, D. Sarkar, and R Core Team. 2016. _nlme: Linear and Nonlinear Mixed Effects Models_. R package version 3.1‐126. URL: CRAN.R-project.org/package=nlme.

Poorter, H., Ü. Niinemets, L. Poorter, I. J. Wright, and R. Villar. 2009. Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytologist 182:565-588. https://doi.org/10.1111/j.1469-8137.2009.02830.x.

Posthuma Partners. 2019. lmvar: Linear Regression with Non-Constant Variances. R package version 1.5.2. URL: CRAN.R-project.org/package=lmvar.

Quinn, G. P., and M. J. Keough. 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, New York. https://doi.org/10.1017/CBO9780511806384.

R Core Team. 2018. R: A language and environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. URL: www.R-proje ct.org.

Ramey, G., and V. A. Ramey. 1995. Cross-country evidence on the link between volatility and growth (No. w4959). National bureau of economic research. The American Economic Review 85:1138-1151. https://doi.org/10.3386/w4959.

Schielzeth, H., and S. Nakagawa. 2013. Nested by design: Model fitting and interpretation in a mixed model era. Methods in Ecology and Evolution 4:14-24. https://doi.org/10.1111/j.2041-210x.2012.00251.x.

Shiller, R. J. 1989. Market volatility. MIT press, Cambridge.

Smyth, G. K. 2002. An efficient algorithm for REML in heteroscedastic regression (remlscore, randomizedBlock, and mixedModel2 functions). Journal of Computational and Graphical Statistics 11:836-847. https://doi.org/10.1198/106186002871.

Tagliamonte, S. A. 2006. Analysing Sociolinguistic Variation (Key Topics in Sociolinguistics). Cambridge University Press, Cambridge. https://doi.org/10.1017/CBO9780511801624.

Viechtbauer, W. 2010. Conducting meta-analyses in R with the metafor package. Journal of Statistical Software 36:1-48. URL: www.jstatsoft.org/v36/i03. https://doi.org/10.18637/jss.v036.i03.

Warton, D., and F. Hui. 2011. The arcsine is asinine: the analysis of proportions in ecology. Ecology 92:3-10. https://doi.org/10.1890/10-0340.1.

West, B. T., K. B. Welch, and A. Gałecki. 2014. Linear Mixed Models: A Practical Guide Using Statistical Software. Chapman and Hall, Boca Raton. https://doi.org/10.1201/b17198.

Whitham T. G., J. K. Bailey, J. A. Schweitzer, S. M. Shuster, R. K. Bangert, C. J. LeRoy, E. V. Lonsdorf, G. J. Allan, S. P. DiFazio, B. M. Potts, D. G. Fischer, C. A. Gehring, R. L. Lindroth, J. C. Marks, S. C. Hart, G. M. Wimp, and S. C. Wooley. 2006. A framework for community and ecosystem genetics: from genes to ecosystems. Nature Reviews Genetics 7:510-523. https://doi.org/10.1038/nrg1877.

Wolf, H. 2004. Accounting for consumption volatility differences. IMF Staff Papers 51:109-125.

Wolf, H. 2005. Volatility: Definitions and Consequences. Pp. 45-64 en J. Aizenman and B. Pinto (eds.). Managing Economic Volatility and Crises: A Practitioner's Guide. Cambridge University Press, Cambridge. https://doi.org/10.1017/CBO9780511510755.004.

Xiao, X., E. P. White, M. B. Hooten, and S. L. Durham. 2011. On the use of log‐transformation vs. nonlinear regression for analyzing biological power laws. Ecology 92:1887-1894. https://doi.org/10.1890/11-0538.1.

Zuur, A. F., E. N. Ieno, N. Walker, A. A. Saveliev, and G. M. Smith. 2009. Mixed effects models and extensions in ecology with R. Springer, New York. https://doi.org/10.1007/978-0-387-87458-6.

Cuando la variabilidad varía: Heterocedasticidad y funciones de varianza

Descargas

Archivos adicionales

Publicado

2020-10-23

Cómo citar

Oddi, F. J., Miguez, F. E., Benedetti, G. G., & Garibaldi, L. A. (2020). Cuando la variabilidad varía: Heterocedasticidad y funciones de varianza. Ecología Austral, 30(3), 438–453. https://doi.org/10.25260/EA.20.30.3.0.1131