Analysis of Divergence in Loglinear Models When Expected Frequencies are Subject to Linear Constraints

Abstract

Consider the loglinear model for categorical data under the assumption of multinomial sampling. We are interested in testing between various hypotheses on the parameter space when we have some hypotheses relating to the parameters of the models that can be written in terms of constraints on the frequencies. The usual likelihood ratio test, with maximum likelihood estimator for the unspecified parameters, is generalized to tests based on -divergence statistics, using minimum -divergence estimator. These tests yield the classical likelihood ratio test as a special case. Asymptotic distributions for the new -divergence test statistics are derived under the null hypothesis.

This is a preview of subscription content, access via your institution.

References

  1. Agresti A (2002). Categorical data analysis. Wiley, London

    MATH  Google Scholar 

  2. Aitchison J, Silvey SD (1958). Maximum likelihood estimation of parameters subject to constraints. Ann Math Stat 29:813–828

    Article  MathSciNet  MATH  Google Scholar 

  3. Bhapkar VP (1979). On tests of symmetry when higher order interactions are absent. J Indian Stat Assoc 17:17–26

    Google Scholar 

  4. Bonett DG (1989). Pearson chi-square estimator and test for log-linear models with expected frequencies subject to linear constraints. Stat Probab Lett 8:175–177

    Article  MathSciNet  Google Scholar 

  5. Christensen R (1997). Log-linear models and logistic regression, 2nd edn. Springer, Berlin Heidelberg New York

    MATH  Google Scholar 

  6. Collet D (1994). Modelling survival data in medical research. Chapman and Hall, London

    Google Scholar 

  7. Cressie N, Pardo L (2000). Minimum -divergence estimator and hierarchical testing in loglinear models. Stat Sinica 10:867–884

    MATH  MathSciNet  Google Scholar 

  8. Cressie N, Pardo L (2002). Model checking in loglinear models using -divergences and MLEs. J Stat Plan Infer 103:437–453

    Article  MATH  MathSciNet  Google Scholar 

  9. Cressie N, Read TRC (1984). Multinomial goodness-of-fit tests. J R Stat Soc B 46:440–464

    MATH  MathSciNet  Google Scholar 

  10. Cressie N, Pardo L, Pardo MC (2003). Size and power considerations for testing loglinear models using -divergence test statistics. Stat Sinica 13:550–570

    MathSciNet  Google Scholar 

  11. Gokhale DV (1973). Iterative maximum likelihood estimation for discrete distributions. Sankhya B 35:293–298

    MathSciNet  Google Scholar 

  12. Haber M (1985). Maximum likelihood methods for linear and log-linear models in categorical data. Comput Stat Data Anal 3:1–10

    Article  MATH  MathSciNet  Google Scholar 

  13. Haber M, Brown MB (1986). Maximum likelihood methods for log-linear models when expected frequencies are subject to linear constraints. J Am Stat Assoc 81:477–482

    MATH  Article  MathSciNet  Google Scholar 

  14. Haberman SJ (1974). The analysis of frequency data. The University of Chicago Press, Chicago

    MATH  Google Scholar 

  15. Imrey PB, Koch GG, Stokes ME (1982). Categorical data analysis: some reflections on the log-linear model and logistic regression. Part II. Data analysis. Int Stat Rev 50:35–63

    MATH  MathSciNet  Article  Google Scholar 

  16. Lang JB (1996). Maximum likelihood methods for a generalized class of loglinear models. Ann Stat 24:726–752

    Article  MATH  Google Scholar 

  17. Lang JB, McDonald JW, Smith WF (1999). Association marginal modeling of multivariate categorical responses: a maximum likelihood approach. J Am Stat Assoc 94:1161–1171

    MATH  Article  MathSciNet  Google Scholar 

  18. Lindsay BG (1994). Efficiency versus robutness: the case for minimum Hellinger distance and other methods. Ann Stat 22:1081–1114

    MATH  Article  MathSciNet  Google Scholar 

  19. Matusita K (1954). On the estimation by the minimum distance method. Ann Inst Stat Math 5:59–65

    MATH  MathSciNet  Article  Google Scholar 

  20. Neyman J (1949). Contribution to the theory of the χ2-test. Proceedings of the first symposium on mathematical statistics and probability. Univ. of Berkeley Press, Berkeley, pp. 239–273

  21. Pardo L (2006). Statistical inference based on divergence measures. Statistics: textbooks and monographs. Chapman Hall/CRC, New York

    MATH  Google Scholar 

  22. Pardo L, Pardo MC (2005). Nonadditivity in loglinear models using -divergences and MLEs. J Stat Plan Infer 127:237–252

    Article  MATH  MathSciNet  Google Scholar 

  23. Pardo JA, Pardo L, Zografos K (2002). Minimum -divergence estimators with constraints in multinomial populations. J Stat Plan Infer 104(1).:221–237

    Article  MATH  MathSciNet  Google Scholar 

  24. Parr WC (1981). Minimum distance estimation: a bibliography. Commun Stat (Theory Methods). 10:1205–1224

    Article  MathSciNet  Google Scholar 

  25. Read TRC, Cressie N (1988). Goodness of fit Statistics for discrete multivariate data. Springer, Berlin Heidelberg New York

    MATH  Google Scholar 

  26. Searle SR (1971). Linear models. Wiley, New York

    MATH  Google Scholar 

  27. Silvey SD (1959). The Lagrangian multiplier test. Ann Math Stat 30:389–407

    Article  MathSciNet  MATH  Google Scholar 

  28. Vajda I (1989). Theory of statistical inference and information. Kluwer, Dordrecht

    MATH  Google Scholar 

  29. Wedderburn RWM (1974). Generalized linear models specified in terms of constraints. J R Stat Soc B 36:449–454

    MATH  MathSciNet  Google Scholar 

Download references

Author information

Affiliations

Authors

Corresponding author

Correspondence to L. Pardo.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Pardo, L., Menéndez, M.L. Analysis of Divergence in Loglinear Models When Expected Frequencies are Subject to Linear Constraints. Metrika 64, 63 (2006). https://doi.org/10.1007/s00184-006-0034-2

Download citation

Keywords

  • Asymptotic distribution
  • Multinomial sampling
  • -divergence test statistics
  • Loglinear models