Robustness of a Class of Partial Diallel Cross Designs to the Unavailability of a Complete Block of Observations

Complete and partial diallel cross designs are examined as to their construction and robustness against the loss of a block of observations. A simple generalized inverse is found for the information matrix of the line effects, which allows evaluation of expressions for the variances of the line-effect differences with and without the missing block. A-efficiencies, based on average variances of the elementary contrasts of the line-effects, suggest that these designs are fairly robust. The loss of efficiency is generally less than 10%, but it is shown that specific comparisons might suffer a loss of efficiency of as much as 40%.     

A monte carlo comparison of the smoothing,scoring and em algorithms for dispersion matrix estimation with incomplete growth curve data

Incomplete growth curve data often result from missing or mistimed observations in a repeated measures design. Virtually all methods of analysis rely on the dispersion matrix estimates. A Monte Carlo simulation was used to compare three methods of estimation of dispersion matrices for incomplete growth curve data. The three methods were: 1) maximum likelihood estimation with a smoothing algorithm, which finds the closest positive semidefinite estimate of the pairwise estimated dispersion matrix; 2) a mixed effects model using the EM (estimation maximization) algorithm; and 3) a mixed effects model with the scoring algorithm. The simulation included 5 dispersion structures, 20 or 40 subjects with 4 or 8 observations per subject and 10 or 30% missing data. In all the simulations, the smoothing algorithm was the poorest estimator of the dispersion matrix. In most cases, there were no significant differences between the scoring and EM algorithms. The EM algorithm tended to be better than the scoring algorithm when the variances of the random effects were close to zero, especially for the simulations with 4 observations per subject and two random effects.     

Balanced incomplete Latin square designs

Latin squares have been widely used to design an experiment where the blocking factors and treatment factors have the same number of levels. For some experiments, the size of blocks may be less than the number of treatments. Since not all the treatments can be compared within each block, a new class of designs called balanced incomplete Latin squares (BILS) is proposed. A general method for constructing BILS is proposed by an intelligent selection of certain cells from a complete Latin square via orthogonal Latin squares. The optimality of the proposed BILS designs is investigated. It is shown that the proposed transversal BILS designs are asymptotically optimal for all the row, column and treatment effects. The relative efficiencies of a delete-one-transversal BILS design with respect to the optimal designs for both cases are also derived; it is shown to be close to 100%, as the order becomes large.     

Minimum,maximum, and average spherical prediction variances for central composite and box-behnken designs

Single value design optimality criteria are often considered when selecting a response surface design. An alternative to a single value criterion is to evaluate prediction variance properties throughout the experimental region and to graphically display the results in a variance dispersion graph (VDG) (Giovannitti-Jensen and Myers (1989)). Three properties of interest are the spherical average, maximum, and minimum prediction variances. Currently, a computer-intensive optimization algorithm is utilized to evaluate these prediction variance properties. It will be shown that the average, maximum, and minimum spherical prediction variances for central composite designs and Box-Behnken designs can be derived analytically. These three prediction variances can be expressed as functions of the radius and the design parameters. These functions provide exact spherical prediction variance values eliminating the implementation of extensive computing involving algorithms which do not guarantee convergence. This research is concerned with the theoretical development of these analytical forms. Results are presented for hyperspherical and hypercuboidal regions.     

Factorial experimentation in second-order Latin cubes

A Second-order Latin cube of size n x n x n can be used as the design for an experiment in three space dimensions, with the three sets of layers used as three sets of blocks. The n2 treatments are then orthogonal to the main effects X, Y and Z of the blocking systems. Particular interest attaches to second-order Latin cubes whose treatments are an n x n factorial set, with the main effects A and B of the treatment factors orthogonal to the interactions XY, XZ and YZbetween pairs of blocking systems. This note describes such designs where components of the interaction AB are each totally confounded with one of XY, XZ and YZ. Cubes with n = 4 are then described where components of A, B and AB are each partially confounded. Finally, a defective design with n = 4 is described, to illustrate the need for care in composing designs for three dimensions.     

Randomization of neighbour balanced generalized Youden designs

If a crossover design with more than two treatments is carryover balanced, then the usual randomization of experimental units and periods would destroy the neighbour structure of the design. As an alternative, Bailey [1985. Restricted randomization for neighbour-balanced designs. Statist. Decisions Suppl. 2, 237–248] considered randomization of experimental units and of treatment labels, which leaves the neighbour structure intact. She has shown that, if there are no carryover effects, this randomization validates the row–column model, provided the starting design is a generalized Latin square. We extend this result to generalized Youden designs where either the number of experimental units is a multiple of the number of treatments or the number of periods is equal to the number of treatments. For the situation when there are carryover effects we show for so-called totally balanced designs that the variance of the estimates of treatment differences does not change in the presence of carryover effects, while the estimated variance of this estimate becomes conservative.     

Estimation of Missing Cells in Randomized Block and Latin Square Designs

Classical missing cell formulas for the randomized block design and Latin square design are derived using the modified cell means model. The role of the assumptions of no interaction for these designs is made clear in the development. Examples are given for the randomized block and Latin square.     

A note on dealing with missing standard errors in meta-analyses of continuous outcome measures in WinBUGS

A meta-analysis of a continuous outcome measure may involve missing standard errors. This is not a problem depending on assumptions made about the population standard deviation. Multiple imputation can be used to impute missing values while allowing for uncertainty in the imputation. Markov chain Monte Carlo simulation is a multiple imputation technique for generating posterior predictive distributions for missing data. We present an example of imputing missing variances using WinBUGS. The example highlights the importance of checking model assumptions, whether for missing or observed data.     

Defective data in the analysis of variance

Several kinds of defective data are considered, the first being that of data completely missing. General expressions for block designs are given for the variances of treatment contrasts. Sometimes even a moderate loss of data will seriously increase an important variance. A similar defect arises when some plots receive the wrong treatment. Again the consequences can be serious though, if the correct form of analysis used, variances of contrasts may be much better estimated than if defective data are regarded as missing. Finally, the problem of the mixed-up plots is considered, the total value for certain plots being known but not the individual data. Formulae are presented for dealing with such a case. Again the correct form of analysis can sometimes retain useful information that would be lost if affected plots are regarded as missing. Finally some considerations are set out for dealing with defective data in general.     

Some New Estimators for Small-Area Means With Application to the Assessment of Farmland Values

Regression models that account for main state effects and nested county effects are considered for the assessment of farmland values. Empirical predictors obtained by replacing the unknown variances in the formulas of the optimal predictors by maximum likelihood estimates are presented. The computations are carried out by simple iterations between two SAS procedures. Estimators for the prediction variances are derived, and a modification to secure the robustness of the predictors is proposed. The procedure is applied to data on nonirrigated cropland in the Corn Belt states and is shown to yield predictors with considerably lower prediction mean squared errors than the survey estimators and other regression-type estimators.     

Maximum expected entropy transformed Latin hypercube designs

Existing projection designs (e.g. maximum projection designs) attempt to achieve good space-filling properties in all projections. However, when using a Gaussian process (GP), model-based design criteria such as the entropy criterion is more appropriate. We employ the entropy criterion averaged over a set of projections, called expected entropy criterion (EEC), to generate projection designs. We show that maximum EEC designs are invariant to monotonic transformations of the response, i.e. they are optimal for a wide class of stochastic process models. We also demonstrate that transformation of each column of a Latin hypercube design (LHD) based on a monotonic function can substantially improve the EEC. Two types of input transformations are considered: a quantile function of a symmetric Beta distribution chosen to optimize the EEC, and a nonparametric transformation corresponding to the quantile function of a symmetric density chosen to optimize the EEC. Numerical studies show that the proposed transformations of the LHD are efficient and effective for building robust maximum EEC designs. These designs give projections with markedly higher entropies and lower maximum prediction variances (MPV''s) at the cost of small increases in average prediction variances (APV''s) compared to state-of-the-art space-filling designs over wide ranges of covariance parameter values.     

A Bayesian Approach in Estimating Transition Probabilities of a Discrete-time Markov Chain for Ignorable Intermittent Missing Data

This article focuses on data analyses under the scenario of missing at random within discrete-time Markov chain models. The naive method, nonlinear (NL) method, and Expectation-Maximization (EM) algorithm are discussed. We extend the NL method into a Bayesian framework, using an adjusted rejection algorithm to sample the posterior distribution, and estimating the transition probabilities with a Monte Carlo algorithm. We compare the Bayesian nonlinear (BNL) method with the naive method and the EM algorithm with various missing rates, and comprehensively evaluate estimators in terms of biases, variances, mean square errors, and coverage probabilities (CPs). Our simulation results show that the EM algorithm usually offers smallest variances but with poorest CP, while the BNL method has smaller variances and better/similar CP as compared to the naive method. When the missing rate is low (about 9%, MAR), the three methods are comparable. Whereas when the missing rate is high (about 25%, MAR), overall, the BNL method performs slightly but consistently better than the naive method regarding variances and CP. Data from a longitudinal study of stress level among caregivers of individuals with Alzheimer’s disease is used to illustrate these methods.     

The robustness of chain block designs and coat-of-mail designs

Chain block designs are relatively vulnerable to loss of information when missing values or outliers may occur. An alternative class of designs, coat-of-mail designs, are proposed and the relative robustness of the two types of design are compared.     

The Effect of Methods for Handling Missing Values on the Performance of the MEWMA Control Chart

This article is concerned with the effect of the methods for handling missing values in multivariate control charts. We discuss the complete case, mean substitution, regression, stochastic regression, and the expectation–maximization algorithm methods for handling missing values. Estimates of mean vector and variance–covariance matrix from the treated data set are used to build the multivariate exponentially weighted moving average (MEWMA) control chart. Based on a Monte Carlo simulation study, the performance of each of the five methods is investigated in terms of its ability to obtain the nominal in-control and out-of-control average run length (ARL). We consider three sample sizes, five levels of the percentage of missing values, and three types of variable numbers. Our simulation results show that imputation methods produce better performance than case deletion methods. The regression-based imputation methods have the best overall performance among all the competing methods.     

Robustness of designed experiments against missing data

This paper investigates the robustness of designed experiments for estimating linear functions of a subset of parameters in a general linear model against the loss of any t( ≥1) observations. Necessary and sufficient conditions for robustness of a design under a homoscedastic model are derived. It is shown that a design robust under a homoscedastic model is also robust under a general heteroscedastic model with correlated observations. As a particular case, necessary and sufficient conditions are obtained for the robustness of block designs against the loss of data. Simple sufficient conditions are also provided for the binary block designs to be robust against the loss of data. Some classes of designs, robust up to three missing observations, are identified. A-efficiency of the residual design is evaluated for certain block designs for several patterns of two missing observations. The efficiency of the residual design has also been worked out when all the observations in any two blocks, not necessarily disjoint, are lost. The lower bound to A-efficiency has also been obtained for the loss of t observations. Finally, a general expression is obtained for the efficiency of the residual design when all the observations of m ( ≥1) disjoint blocks are lost.     

Crossover designs in the presence of carry-over effects from two factors

Experiments, used in the telecommunications industry and elsewhere, are considered that involve the simultaneous application of levels of two unrelated factors, treatments and stimuli, to each of several subjects in a succession of time periods. The existence is suspected of carry-over effects of treatments and stimuli, in the period immediately following the period of their application. Methods are given for the construction of separate sequences of treatments and of stimuli; these methods are based on the Latin squares of Williams and of Russell. In the resulting designs, the treatments and stimuli are either orthogonal or nearly orthogonal, and the coincidence of the direct and carry-over effects of each factor is either balanced or nearly balanced. The efficiencies of the designs are assessed by comparing the average variances of elementary contrasts in the levels of each factor with appropriate lower bounds.     

Efficient Estimation of Data Combination Models by the Method of Auxiliary-to-Study Tilting (AST)

We propose a locally efficient estimator for a class of semiparametric data combination problems. A leading estimand in this class is the average treatment effect on the treated (ATT). Data combination problems are related to, but distinct from, the class of missing data problems with data missing at random (of which the average treatment effect (ATE) estimand is a special case). Our estimator also possesses a double robustness property. Our procedure may be used to efficiently estimate, among other objects, the ATT, the two-sample instrumental variables model (TSIV), counterfactual distributions, poverty maps, and semiparametric difference-in-differences. In an empirical application, we use our procedure to characterize residual Black–White wage inequality after flexibly controlling for “premarket” differences in measured cognitive achievement. Supplementary materials for this article are available online.     

Properties of maximum likelihood estimators of variance components in the one-way classification,balanced data

We studied properties of maximum likelihood estimators (MLEs) of the variance components obtained from balanced data of the one-way classification. Exact and asymptotic expected values and variances of these MLEs were derived under the usual normality assumptions. Numerical studies illustrate these expected values and variances, and also illustrate the probability of obtaining a negative solution to the maximum likelihood (ML) equation for the between-class variance component. Simulations were used to study the robustness of the ML estimators under non-normal distributions.     

ON THE DESIGN OF FIELD EXPERIMENTS UNDER A ONE-DIMENSIONAL NEIGHBOUR MODEL

This paper examines the effect of randomisation restrictions, either to satisfy conditions for a balanced incomplete block design or to attain a higher level of partial neighbour balance, on the average variance of pair-wise treatment contrasts under a neighbour model discussed by Gleeson & Cullis (1987). Results suggest that smaller average pairwise variances can be obtained by ignoring requirements for incomplete block designs and concentrating on achieving a higher level of partial neighbour balance. Field layout of the design, although often determined by practical constraints, e.g. size, shape of site, minimum plot size and experimental husbandry, may markedly affect average pairwise variance. For the one-dimensional (row-wise) neighbour model considered here, investigation of three different layouts suggests that for a rectangular array of plots, smaller average pairwise variances can generally be obtained from layouts with fewer rows and more plots per row.     

Optimal designs for the development of personalized treatment rules

We study the design of multi-armed parallel group clinical trials to estimate personalized treatment rules that identify the best treatment for a given patient with given covariates. Assuming that the outcomes in each treatment arm are given by a homoscedastic linear model, with possibly different variances between treatment arms, and that the trial subjects form a random sample from an unselected overall population, we optimize the (possibly randomized) treatment allocation allowing the allocation rates to depend on the covariates. We find that, for the case of two treatments, the approximately optimal allocation rule does not depend on the value of the covariates but only on the variances of the responses. In contrast, for the case of three treatments or more, the optimal treatment allocation does depend on the values of the covariates as well as the true regression coefficients. The methods are illustrated with a recently published dietary clinical trial.