Various varying variances: The challenge of nuisance parameters to the practising biostatistician

Abstract

The 1997 Biometrics paper by Mike Kenward and James Roger has become a citation classic (more than 1260 citations by End June 2013 according to Google Scholar) and the solution that they proposed to deal with the problem of significance tests of fixed effects in REML models is now incorporated in many software packages and accepted by all biostatisticians as the method of choice. Nevertheless, it does not solve all problems, since there is more to analysis than just significance and since the problems that models with more than one variance pose arise in many contexts. In this paper, I discuss some problems and applications and make some tentative suggestions as to how they may be tackled. My excuse for raising problems I do not solve is that it may inspire James and Mike to complete what they started.

Keywords

REML Components of Variance Mixed models Analysis of covariance Gauss-Markov theorem History of statistics

Get full access to this article

View all access options for this article.

References

Airy

. On the algebraical and numerical theory of errors of observations and the combination of observations, London: Macmillan, 1875.

Kenward

Roger

. The use of baseline covariates in crossover studies. Biostatistics 2010; 11(1): 1–17.

Senn

. Cross-over trials in clinical research, 2nd ed. Chichester: Wiley, 2002.

Kenward

Roger

. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics 1997; 53(3): 983–997.

Patterson

Thompson

. Recovery of inter-block information when block sizes are unequal. Biometrika 1971; 58(3): 545–554.

Cox

Solomon

. Components of variance, Boca Raton: Chapman and Hall, 2003.

David

. First (questionable) occurrence of common terms in mathematical-statistics. Am Stat 1995; 49(2): 121–133.

Pearson

. Contributions to the mathematical theory of evolution. Phil Trans Roy Soc Lond Ser A 1894; 185: 71–110.

Yule

. On the theory of correlation. J Roy Stat Soc 1897; 60: 812–854.

10.

Fisher

. The correlation between relatives on the supposition of Medelian inheritance. Trans Roy Soc Edinburgh 1918; 52: 339–433.

11.

Fisher

. On the mathematical foundation of theoretical statistics. Phil Trans Roy Soc Lond Ser A 1922; 222: 309–368.

12.

Student. The probable error of a mean. Biometrika 1908; 6: 1–25.

13.

Senn

Richardson

. The first t-test. Stat Med 1994; 13(8): 785–803.

14.

Stigler

. The history of statistics: The measurement of uncertainty before 1900, Cambridge, MA: Belknap Press, 1986.

15.

Hald

. A history of mathematical statistics from 1750 to 1930, 1st ed. New York: John Wiley & Sons Ltd, 1998.

16.

Pearson ES. Student, A statistical biography of William Sealy Gosset (Edited and augmented by RL Plackett with the assistance of GA Barnard). Oxford: Clarendon Press, 1990.

17.

Pfanzagl

Sheynin

. Studies in the history of probability and statistics .44. A forerunner of the t-distribution. Biometrika 1996; 83(4): 891–898.

18.

Fisher

. Statistical methods for research workers, Edinburgh: Oliver and Boyd, 1925.

19.

Fisher

. Applications of ‘Student's' distribution. Metron 1925; 5: 90–104.

20.

Behrens

. Ein beitrag zur Fehlerberechnung bei wenigen Beobachtungen. Landwirtschaftliche Jahrbücher 1929; 62: 807–837.

21.

Yates

. The recovery of inter-block information in variety trials arranged in three-dimensional lattices. Ann Eugen 1939; 9: 136–156.

22.

Yates

. The recovery of inter-block information in balanced incomplete block designs. Ann Hum Genet 1940; 10(1): 317–325.

23.

Yates

Cochran

. The analysis of groups of experiments. J Agric Sci 1938; 28(4): 556–580.

24.

Aitken

. On least-squares and linear combinations of observations. Proc Roy Soc Edinburgh 1934; 55: 42–48.

25.

Neyman

Scott

. Consistent estimates based on partially consistent observations. Econometrica 1948; 16(1): 1–32.

26.

Lancaster

. The incidental paramter problem since 1948. J Econom 2000; 95: 391–413.

27.

Hoerl

Kennard

. Ridge regression: Biased estimation for nonorthogonal problems. Technometrics 1970; 12(1): 56–67.

28.

Peto

. Asymptotically efficient rank invariant test procedures (with discussion). J Roy Stat Soc A 1972; 135(2): 185–207.

29.

Nelder

Wedderburn

RWM

. Generalized linear models. J Roy Stat Soc A 1972; 132: 107–120.

30.

Cox

. Regression models and life-tables (with discussion). J Roy Stat Soc Ser B 1972; 34: 187–220.

31.

Lindley

Smith

AFM

. Bayes estimates for the linear model (with discussion). J Roy Stat Soc B 1972; 55: 1–41.

32.

Nelder

. The combination of information in generally balanced designs. J Roy Stat Soc Ser B 1968; 30(2): 303–311.

33.

Harville

. Maximum likelihood approaches to variance component estimation and to related problems. J Am Stat Assoc 1977; 72(358): 320–338.

34.

Kenward

Roger

. An improved approximation to the precision of fixed effects from restricted maximum likelihood. Comput Stat Data Anal 2009; 53(7): 2583–2595.

35.

Senn

. An unreasonable prejudice against modelling? Pharm Stat 2005; 4: 87–89.

36.

Lesaffre

Senn

. A note on non-parametric ANCOVA for covariate adjustment in randomized clinical trials. Stat Med 2003; 22(23): 3583–3596.

37.

Senn

. Statistical issues in drug development, Hoboken: Wiley, 2007.

38.

Senn

Modelling in drug development. In: Christie

Cliffe

Dawid

Senn

(eds). Simplicity, complexity and modelling, Chichester: Wiley, 2011, pp. 35–49.

39.

Dawid

Fang

. Conjugate Bayes discrimination with infinitely many variables. J Multivar Anal 1992; 41(1): 27–42.

40.

Gessa D. Simple strategies for variance uncertainty in meta-analysis. MSc Thesis, University of Glasgow, UK, 2008.

41.

Popper Shaffer

. The Gauss-Markov theorem and random regressors. Am Stat 1991; 45: 269–272.