On blocking rules for the bootstrap with dependent data

We address the issue of optimal block choice in applicationsof the block bootstrap to dependent data. It is shown that optimalblock size depends significantly on context, being equal ton^1/3, n^1/4 and n^1/5 in the cases of variance or bias estimation,estimation of a onesided distribution function, and estimationof a two-sided distribution function, respectively. A clearintuitive explanation of this phenomenon is given, togetherwith outlines of theoretical arguments in specific cases. Itis shown that these orders of magnitude of block sizes can beused to produce a simple, practical rule for selecting blocksize empirically. That technique is explored numerically.     

Conditional Maximum Likelihood Estimation Following a Group Sequential Test

We consider estimation after a group sequential test. An estimator that is unbiased or has small bias may have substantial conditional bias (Troendle and Yu, 1999, Coburger and Wassmer, 2001). In this paper we derive the conditional maximum likelihood estimators of both the primary parameter and a secondary parameter, and investigate their properties within a conditional inference framework. The method applies to both the usual and adaptive group sequential test designs. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Bias of heritability estimates from twin data in presence of errors of zygosity determination

Simple heritability estimators of continuous as well as discrete traits from twin data are known to overestimate the degree of genetic determination of the measured traits for several reasons. Errors of zygosity determination will, however, underestimate the true heritability. The bias due to wrong assignment of dizygous twin pairs into monozygous type is evaluated here, and the results indicate that this negative bias has a compensatory effect on the estimate of the degree of genetic determination when other factors of similarity between twin pairs are taken into account. It is shown that when an estimate of zygosity error is available, the bias due to this factor can be evaluated quantitatively, and hence the adjustment for zygosity error can be incorporated in the estimation of the degree of genetic determination of a trait. Although this theory is explicitly developed here for twin studies, the general principle also applies for other types of errors of determining the degree of biological relationships for estimation of heritability, in which case this type of error may be more important than the simple zygosity error.     

Bias of maximum likelihood estimator of intraclass correlation

Summary A bias correction was derived for the maximum likelihood estimator (MLE) of the intraclass correlation. The bias consisted of two parts: a correction from MLE to the analysis of variance estimator (ANOVA) and the bias of ANOVA. The total possible bias was always negative and depended upon both the degree of correlation and the design size and balance. The first part of the bias was an exact algebraic expression from MLE to ANOVA, and the corrected estimator by this part was ANOVA. It was also shown that the first correction term was equivalent to Fisher's reciprocal bias correction on hisZ scores. The total possible bias of MLE was large for small and moderate samples. Relative biases were larger for small parametric values and vice versa. To ensure a relative bias less than 10% assuming an intraclass correlation of 0.025, which is not unusual in most of the animal genetic studies, the total number of observations (N) should be not less than 500. From a design point of view, minimum bias occurred atn = 2, the minimum family size possible, underN fixed.     

亚氨基酸编码方法及其应用

赋予氨基酸编码方法下除终止子之外的密码子突变为终止子时每一位发生的变化权值,利用矩阵来表示所有的突变方式和难易程度,综合亲水性与疏水性理化性质,提出亚氨基酸编码方法,并给出该编码方法下同义密码子的相对使用度fSubtypesRelativeSynonymousCodonUsage,SRSCU）．然后选取15条H5N1序列,使用MEGA4．0分析它们的同源性,并分别在氨基酸编码、拟氨基酸编码、亚氨基酸编码这三种环境下研究所选序列使用密码子的偏好性,对比结果验证,亚氨基酸编码方法具有相应的优越性．     

Some general points in estimating heterogeneity variance with the DerSimonian-Laird estimator

In this paper we consider estimating heterogeneity variance with the DerSimonian-Laird (DSL) estimator as typically used in meta-analysis. In its general form the DSL estimator requires inverse population-averaged study-specific variances as weights, in which case the estimator is unbiased. It has become common practice, however, to use estimates of the study-specific variances instead of their population-averaged versions. This can lead to considerable bias. Simulations illustrate these findings.     

Small-sample bias and corrections for conditional maximum-likelihood odds-ratio estimators

A number of small-sample corrections have been proposed for the conditional maximum-likelihood estimator of the odds ratio for matched pairs with a dichotomous exposure. I here contrast the rationale and performance of several corrections, specifically those that generalize easily to multiple conditional logistic regression. These corrections or Bayesian analyses with informative priors may serve as diagnostics for small-sample problems. Points are illustrated with a small exact performance comparison and with an example from a study of electrical wiring and childhood leukemia. The former comparison suggests that small-sample bias may be more prevalent than commonly realized.     

Robust estimating functions and bias correction for longitudinal data analysis

Robust methods are useful in making reliable statistical inferences when there are small deviations from the model assumptions. The widely used method of the generalized estimating equations can be "robustified" by replacing the standardized residuals with the M-residuals. If the Pearson residuals are assumed to be unbiased from zero, parameter estimators from the robust approach are asymptotically biased when error distributions are not symmetric. We propose a distribution-free method for correcting this bias. Our extensive numerical studies show that the proposed method can reduce the bias substantially. Examples are given for illustration.