Methods and Models in Statistics: In Hanour of Professor John Nelder, FRS -- CONTENTS -- 1. Preface N. M. Adams, M. J. Crowder, D. J. Hand, and D. A. Stephens -- 2. Foreword J. A. Nelder -- 3. John Nelder: From General Balance to Generalised Models (Both Linear and Hierarchical) S. Senn -- Introduction: some personal remarks -- Childhood -- The Child is Father of the Man -- General balance -- Generalised Linear Models -- Statistical computing -- Collaboration with Youngjo Lee -- In summary: some personal remarks -- References -- 4. Some Remarkes on Model Criticism D. R. Cox -- 1. Introduction -- 2. Formal theory -- 3. Choice of criterion -- 4. Influence and breakdown -- 5. Role in applications -- Reference -- 5. Likelihood Perspectives in the Consensus and Controversies of Statistical Modelling and Inference Y. Pawitan -- 1. Introduction -- 2. Likelihood in modelling -- 3. Extensions of likelihood -- 3.1. Profile, marginal, conditional, modified profile likelihoods -- 3.2. Partial and empirical likelihoods -- 3.3. Quasi- and pseudo-likelihoods and estimating equations -- 3.4. Predictive and hierarchical likelihoods -- 4. Various comments -- 4.1. A case for likelihood-based inference -- 4.2. Computing -- 4.3. Non-likelihood methods -- 5. Consensus -- 6. Controversies in Inference -- 7. Views of probability -- 8. Paradoxes -- 8.1. Exchange paradox -- 8.2. Saint Petersburg Paradox -- 8.3. Prisoner's dilemma -- 9. Ladder of uncertainty -- Other ladders in mathematics -- 9.1. Likelihood versus probability -- 10. Settling the controversies? -- References -- 6. Perspectives of ANOVA, REML and a General Linear Mixed Model B. R. Cullis, A. B. Smith and R. Thompson -- 1. Introduction -- 2. REML and the analysis of Generally Balanced Designs -- 2.1. Preliminaries -- 2.2. Split Plot Design -- 2.3. Balanced Incomplete Block Design.

3. A general linear mixed model -- 3.1. The Model -- 3.2. R- and G-structures -- 3.3. Identifiability of variance models -- 4. Estimation -- 4.1. REML estimation of variance parameters -- 4.2. Prediction and the mixed model equations -- 5. Iterative schemes for REML estimation of variance parameters -- 5.1. The Average Information algorithm -- 5.2. The Expectation-Maximisation algorithm -- 5.2.1. Example: EM updates for an unstructured G-matrix -- 5.3. The Parameter Expanded EM algorithm -- 5.3.1. Example: PXEM updates for an unstructured G-matrix -- 5.4. Improved iterative schemes -- 5.4.1. Local schemes -- 5.5. Analysis of data-sets -- 5.5.1. Lamb weight data -- 5.5.2. Ultrafiltration data -- 5.6. Discussion of Results -- 6. Inference in linear mixed models -- 6.1. Hypothesis tests for variance models -- 6.2. Inference for fixed effects -- 6.3. Computing the scaled F and adjusted variance matrix -- 6.4. Kenward adjustments in ANOVA settings -- 7. Prediction for the general linear mixed model -- 7.1. The Prediction Model -- 7.2. Steps in the prediction process -- 7.3. Prediction process -- 7.4. Computing Strategy -- Acknowledgments -- References -- 7. Algorithms, Data Structures and Languages - the Computational Ingredients for Innovative Analysis R. Payne -- 1. Context and history -- 1.1. Rothamsted -- 1.2. Working Party on Statistical Computing -- 1.3. Imperial College -- 2. Data structures -- 3. Algorithms -- 3.1. Analysis of variance -- 3.2. Generalized linear models -- 3.3. Hierarchical generalized linear models -- 4. Statistical software -- 4.1. Genstat -- 4.2. GLIM -- 4.3. GLIMPSE -- 4.4. The K-, MD- and HG-systems -- 5 . Conclusion -- Acknowledgments -- References -- 8. Non-Linear Regression Modelling and Inference J. C. Wakefield -- 1. Introduction -- 2. Frequentist Inference -- 2.1. Estimating Functions -- 2.2. Likelihood.

2.3. Quasi-Likelihood -- 2.4. Sandwich Estimation -- 3. Bayesian Inference -- 3.1. Summarising the Posterior Distribution -- 3.2. Model Misspecification -- 3.3. The Prior Distribution -- 3.4. Simulation-Based Inference -- 4. Comparison of Frequentist and Bayesian Methods -- 5. Non-Linear Regression Models -- 5.1. Generalized Linear Models -- 5.2. Compartmental Models -- 6 . Pharmacokinetic Data Analysis -- 7. Discussion -- Acknowledgments -- References -- 9. Selecting Amongst Large Classes of Models B. D. Ripley -- 1. Introduction -- 2. Why do we want to select a model? -- 3. A historical perspective -- 4. Cross-validation -- 5. AIC , BIC and all that -- Derivation of AIC -- Crucial assumptions -- 6. Bayesian Approaches -- Crucial assumptions -- 7. Deviance Information Criterion -- 8. Model Averaging -- How do we choose the weights? -- Bagging, boosting, random forests -- 9. Practical model selection in 2004 -- References -- 10. Principles of Designed Experiments in J. A. Nelder's Papers R. A. Bailey -- 1. Experimental protocol -- 2. Plot structure is different from treatment structure -- 3. Crossing and nesting -- 4. Orthogonal structures -- 5. Randomization -- 6. Variance components -- 7. Model fitting, estimation and testing -- 8. Combinatorial design -- 9. Quantitative treatments -- 10. Algorithms for analysis -- References -- 11. Likelihood-based Models Beyond GLMs Y. Lee -- 1. Introduction -- 2. GLMs -- 3. JGLMs -- 3.1. EQL and quasi-models -- 3.2. Fitting algorithm for JGLMs -- 4. HGLMs -- 4.1. H-likelihood -- 4.2. Fitting algorithm for HGLMs -- 5. DHGLMs -- 6. Random effects for temporal and spatial correlations -- 6.1. Random effects described by fixed L matrices -- 6.2. Random effects described by a covariance matrices -- 6.3. Random effects described by a precision matrices -- 6.4. Financial models for dispersion.

6.5. Fitting correlated random effects -- 7. Quasi extended GLM classes of models -- 8. Frailty models via HGLMs -- 9. Application of DHGLMs -- 10. Concluding remarks -- Acknowledgments -- References -- 12. A Statistical Examination of the Hastings Rarities J. A. Nelder -- METHODS -- THE RESULTS -- The distribution of occurrences by season -- The distribution of records by years -- DISCUSSION -- SUMMARY -- ACKNOWLEDGMENTS -- REFERENCES -- Appendix-Rarity classes of species and subspecies analysed -- 13. The Works of John Nelder -- BOOKS -- PAPERS -- INDEX.