Chapter 5: Generalized Linear Models: Structure

  • Peter K. Dunn
  • Gordon K. Smyth
Part of the Springer Texts in Statistics book series (STS)


Chapters  2] and  3 considered linear regression models. These models assume constant variance, which demonstrably is not true for all data, as shown in Chap.  4. Generalized linear models (glms) assume the responses come from a distribution that belongs to a more general family of distributions, and also permit more general systematic components. We first review the two components of a glm (Sect. 5.2) then discuss in greater detail the family of distributions upon which the random component is based (Sect. 5.3), including writing the probability functions in the useful dispersion model form (Sect. 5.4). The systematic component of the glm is then considered in greater detail (Sect. 5.5). Having discussed the two components of the glm, glms are then formally defined (Sect. 5.6), and the important concept of the deviance function is introduced (Sect. 5.7). Finally, using a glm is compared to using a regression model after transforming the response (Sect. 5.8).


  1. [1]
    Data Desk: Data and story library (dasl) (2017). URL
  2. [2]
    Johnson, B., Courtney, D.M.: Tower building. Child Development 2(2), 161–162 (1931)CrossRefGoogle Scholar
  3. [3]
    Jørgensen, B.: The Theory of Dispersion Models. Monographs on Statistics and Applied Probability. Chapman and Hall, London (1997)Google Scholar
  4. [4]
    Keller, D.K.: The Tao of Statistics: A Path to Understanding (with no Math). Sage Publications, Thousand Oaks, CA (2006)CrossRefGoogle Scholar
  5. [5]
    Kokonendji, C.C., Khoudar, M.: On strict arcsine distribution. Communications in Statistics—Theory and Methods 33(5), 993–1006 (2004)MathSciNetCrossRefGoogle Scholar
  6. [6]
    Maron, M.: Threshold effect of eucalypt density on an aggressive avian competitor. Biological Conservation 136, 100–107 (2007)CrossRefGoogle Scholar
  7. [7]
    Nelder, J.A., Pregibon, D.: An extended quasi-likelihood function. Biometrika 74(2), 221–232 (1987)MathSciNetCrossRefGoogle Scholar
  8. [8]
    Shmueli, G., Minka, T.P., Kadane, J.B., Borle, S., Boatwright, P.: A useful distribution for fitting discrete data: Revival of the Conway–Maxwell–Poisson distribution. Journal of the Royal Statistical Society: Series C 54(1), 27–142 (2005)MathSciNetCrossRefGoogle Scholar
  9. [9]
    Singer, J.D., Willett, J.B.: Improving the teaching of applied statistics: Putting the data back into data analysis. The American Statistician 44(3), 223–230 (1990)Google Scholar
  10. [10]
    Smyth, G.K.: Australasian data and story library (Ozdasl) (2011). URL
  11. [11]
    Smyth, G.K., Verbyla, A.P.: Adjusted likelihood methods for modelling dispersion in generalized linear models. Environmetrics 10, 695–709 (1999)CrossRefGoogle Scholar
  12. [12]
    Venables, W.N., Ripley, B.D.: Modern Applied Statistics with S, fourth edn. Springer-Verlag, New York (2002). URL CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  • Peter K. Dunn
    • 1
  • Gordon K. Smyth
    • 2
  1. 1.Faculty of Science, Health, Education and EngineeringSchool of Health of Sport Science, University of the Sunshine CoastQueenslandAustralia
  2. 2.Bioinformatics DivisionWalter and Eliza Hall Institute of Medical ResearchParkvilleAustralia

Personalised recommendations