Abstract
This paper presents and extends the concept of recursive residuals and their estimation to an important class of statistical models, Linear Mixed Models (LMM). Recurrence formulae are developed and recursive residuals are defined. Recursive computable expressions are also developed for the model’s likelihood, together with its derivative and information matrix. The theoretical framework for developing recursive residuals and their estimation for LMM varies with the estimation method used, such as the fitting-of-constants or the Best Linear Unbiased Predictor method. These methods are illustrated through application to an LMM example drawn from a published study. Model fit is assessed through a graphical display of the developed recursive residuals and their Cumulative Sums.
Similar content being viewed by others
References
Agresti, A.: Categorical Data Analysis, 2nd edn. Wiley, New York (2002)
Ahmed, B.-M.: Recursive Residuals for Linear Mixed Models. Ph.D. thesis. University of Western Sydney, Australia (2004)
Bartlett, M.: An inverse matrix adjustment arising in discriminant analysis. Ann. Math. Stat. 22, 107–111 (1951)
Bates, D., DebRoy, S.: Linear mixed models and penalized least squares. J. Multivar. Anal. 91(1), 1–17 (2004)
Brown, R.L., Durbin, J., Evans, J.M.: Techniques for testing the constancy of regression relationships over time. J. R. Stat. Soc. Ser. B (Method.) 37(2), 149–192 (1975)
de Luna, X., Johansson, P.: Testing exogeneity under distributional misspecification. Working Paper 2001; 9. Institute for Labour Market Policy Evaluation (IFAU), Uppsala
Does, R., Koning, A.: CUSUM charts for preliminary analysis of individual observations. J. Qual. Technol. 32, 122–132 (2000)
Galpin, J., Hawkins, D.: The use of recursive residuals in checking model fit in linear regression. Am. Stat. 38, 94–105 (1984)
Godolphin, J.: New formulations for recursive residuals as a diagnostic tool in the fixed-effects linear model with design matrices of arbitrary rank. Comput. Stat. Data Anal. 53(6), 2119–2128 (2009)
Haslett, J., Haslett, S.: The three basic types of residuals for a linear model. Int. Stat. Rev. 75(1), 1–24 (2007)
Hawkins, D.: Diagnostics for use with regression recursive residuals. Technometrics 33(2), 221–234 (1991)
Hawkins, D., Olwell, D.: Cumulative Sum Charts and Charting for Quality Improvement. Springer, New York (1998)
Henderson, C.: Estimation of variance and covariance components. Biometrics 9(2), 226–252 (1953)
Henderson, C.R.: Estimation of variances and covariances under multiple trait models. J. Dairy Sci. 67, 1581–1589 (1984)
Hilden-Minton, J.A.: Multilevel diagnostics for mixed and hierarchical linear models [dissertation]. University of California, Los Angeles (2001)
Hodges, J.: Some algebra and geometry for hierarchical models, applied to diagnostics. J. R. Stat. Soc. Ser. B Stat. Methodol. 60, 497–536 (1998)
Houseman, E., Ryan, L., Coull, B.: Cholesky residuals for assessing normal errors in a linear model with correlated outcomes. J. Am. Stat. Assoc. 99(466), 383–394 (2004)
Jacqmin-Gadda, H., Sibillot, S., Proust, C., Molina, J., Thiebaut, R.: Robustness of the linear mixed model to misspecified error distribution. Comput. Stat. Data Anal. 51(10), 5142–5154 (2007)
Jiang, J.: Goodness-of-fit tests for mixed model diagnostics. Ann. Stat. 29(4), 1137–1164 (2001)
Khamis, F.: Crime and divorce. Can one lead to the other? Using multilevel mixed models. Electron. J. Appl. Stat. Anal. 10(2), 328–348 (2017)
Kianifard, F., Swallow, W.: A review of the development and application of recursive residuals in linear models. J. Am. Stat. Assoc. 91, 391–400 (1996)
Lee, Y., Nelder, J.: Hierarchical generalized linear models. J. R. Stat. Soc. Ser. B (Methodol.) 58(4), 619–678 (1996)
Lin, D., Wei, L., Ying, Z.: Model-checking techniques based on cumulative residuals. Biometrics 58, 1–12 (2002)
McGilchrist, C., Cullis, B.: REML estimation for repeated measures analysis. J. Stat. Comput. Simul. 38(1), 151–163 (1991)
McGilchrist, C., Matawie, K.: Recursive residuals in generalised linear models. J. Stat. Plan. Inference 70(2), 335–344 (1998)
McGilchrist, C., Sandland, R.: Recursive estimation of the general linear model with dependent errors. J. R. Stat. Soc. Ser. B (Methodol.) 41, 65–68 (1979)
McGilchrist, C., Sandland, R., Hennessy, J.: Generalized inverses used in recursive residuals estimation of the general linear model. Aust. N. Z. J. Stat. 25(2), 321–328 (1983)
McGilchrist, C., Lianto, S., Byron, D.: Signed residuals. Biometrics 45, 237–246 (1989)
Nobre, J., Singer, J.: Residual analysis for linear mixed models. Biom. J. Biom. Z. 49(6), 863–875 (2007)
Pickford, M., Haslett, S.: A statistical test of single firm market power. N. Z. Econ. Papers 33, 39–58 (1999)
Pinheiro, J., Bates, D.: Mixed-Effects Models in S and S-PLUS. Springer, New York (2000)
Plackett, R.: Some theorems in least squares. Biometrika 37(1–2), 149–157 (1950)
Schützenmeister, A., Piepho, H.: Residual analysis of linear mixed models using a simulation approach. Comput. Stat. Data Anal. 56, 1405–1416 (2012)
Tobing, H., McGilchrist, C.: Recursive residuals for multivariate regression models. Aust. N. Z. J. Stat. 34(2), 217–232 (1992)
Verbeke, G., Molenberghs, G.: Linear Mixed Models for Longitudinal Data. Springer, New York (2000)
Acknowledgements
This research was supported in part by an Early Career Researcher Grant (ECR); National Health and Medical Research Council (NHMRC) Principle Research Fellowship, Federation University Australia, Ballarat, Australia.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Bani-Mustafa, A., Matawie, K.M., Finch, C.F. et al. Recursive residuals for linear mixed models. Qual Quant 53, 1263–1274 (2019). https://doi.org/10.1007/s11135-018-0814-6
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11135-018-0814-6