Advertisement

Letters in Spatial and Resource Sciences

, Volume 10, Issue 2, pp 149–159 | Cite as

What values of Moran’s I and Theil index decomposition really mean under different conditions: on the issue of interpretation

Original Paper
  • 157 Downloads

Abstract

In recent decades, improved methodological apparatuses and increased data availability have enhanced data analyses in social sciences. Moreover, complex analyses using sophisticated methods take just a matter of seconds nowadays thanks to highly powerful software. However, such methods are often poorly understood from a methodological point of view despite the fact that knowledge of their specific properties is crucial to accurately interpreting the results. In this paper we study methods of spatial aspects of variability and examine a specific property of such methods to demonstrate how it can affect the final interpretation. By modelling data in a regular 100 by 100 grid as well as empirical examples from Czechia based on data from the 2011 Czech census, this paper presents possible interpretation-biases and recommendations for how to avoid them. We use the example of spatial autocorrelation (measured by Moran’s I) and variability decomposition (measured by the Theil index); two basic methods which enable us to measure variability in regions and in space.

Keywords

Regional variability Spatial autocorrelation Interpretation Theil index Moran’s I 

JEL Classification

R12 C21 C20 

Notes

Acknowledgments

This work was supported by the Czech Science Foundation (GACR) under Grant No. 15-10493S.

References

  1. Akita, T.: Decomposing regional income inequality using two-stage nested Theil decomposition method. 6th World Congress of the Regional Science Association International. Lugano (2000)Google Scholar
  2. Anselin, L.: Spatial econometrics: methods and models. Kluwer, Dordrecht (1988)CrossRefGoogle Scholar
  3. Anselin, L.: Local indicators of spatial association-LISA. Geogr. Anal. 27, 93–115 (1995)CrossRefGoogle Scholar
  4. Byrne, D.: Interpreting quantitative data. Sage (2002)Google Scholar
  5. Foster, J.E., Ok, E.: Lorenz dominance and the variance of logarithms. Econometrica 67, 901–907 (1999)CrossRefGoogle Scholar
  6. Kritzer, H.M.: The data puzzle: the nature of interpretation in quantitative research. Am. J. Polit. Sci. 1-32 (1996)Google Scholar
  7. Newman, I., Benz, C.R.: Qualitative-quantitative research methodology: exploring the interactive continuum. SIU Press (1998)Google Scholar
  8. Nosek, V., Netrdová, P.: Measuring Spatial Aspects of Variability. Comparing Spatial Autocorrelation with Regional Decomposition in International Unemployment Research. Hist. Soc. Res. 39 (2014)Google Scholar
  9. Novotný, J., Nosek, V.: Comparison of regional inequality in unemployment among four Central European countries: an inferential approach. Lett. Spat. Resour. Sci. 5, 95–101 (2012)CrossRefGoogle Scholar
  10. Onwuegbuzie, A.J., Daniel, L.G.: Typology of analytical and interpretational errors in quantitative and qualitative educational research. Current Issues in Education 6 (2003)Google Scholar
  11. Rey, S.J., Smith, R.J.: A spatial decomposition of the Gini coefficient. Lett. Spat. Resour. Sci. 6, 55–70 (2013)CrossRefGoogle Scholar
  12. Shorrocks, A.: The class of additively decomposable inequality measures. Econometrica, pp. 613–625 (1980)Google Scholar
  13. Shorrocks, A., Wan, G.: Spatial decomposition of inequality. J. Econ. Geogr. 5, 59–81 (2005)CrossRefGoogle Scholar
  14. Silber, J.: Factor components, population subgroups and the computation of the Gini index of inequality. The Review of Economics and Statistics, 107–115 (1989)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2016

Authors and Affiliations

  1. 1.Department of Social Geography and Regional Development, Faculty of ScienceCharles University in PraguePragueCzechia

Personalised recommendations