Journal of the American Statistical Association

Journal covers statistical science and its applications, theory, and methods in economic, social, physical, engineering and health sciences.

Articles from Vol. 90, No. 431, September

A Bayesian Method for Combining Results from Several Binomial Experiments
1. INTRODUCTION Consider a collection of I independent binomial experiments, with experiment i having size [n.sub.i] and success probability [[Theta].sub.i], i = 1, . . . , I. A typical Bayesian hierarchical approach would assume the [[Theta].sub.i]'s...
An Alternative Definition of Finite-Sample Breakdown Point with Applications to Regression Model Estimators
1. INTRODUCTION Often in statistical practice, data possibly contaminated by recording errors are the only data available. Even in situations in which the data are themselves clean, observations may occasionally be present that arise from a process...
Approximations to Multivariate Normal Rectangle Probabilities Based on Conditional Expectations
1. INTRODUCTION Rectangle and orthant probabilities from the multivariate normal distribution have many applications in statistics. These include the multivariate probit model (e.g., Ochi and Prentice 1984), the multivariate ordinal response model (e.g.,...
Comparison of Regression Curves Using Quasi-Residuals
1. INTRODUCTION In scientific and social studies, it is very common to compare groups of individuals (or items) in terms of some parameter. For example, a medical researcher may wish to compare the mean reaction time for a drug in male and female patients....
Diagnostics and Robust Estimation in Multivariate Data Transformations
1. INTRODUCTION Consider a p-variate random vector X - ([X.sub.1], . . . , [X.sub.p])[prime] such that all its components take positive values. If X is not multivariate normal, Andrews, Gnanadesikan, and Warner (1971) proposed the following transformation...
Diagnostics for Linearization Confidence Intervals in Nonlinear Regression
1. INTRODUCTION Linear approximation (LA) confidence intervals are very popular in statistical applications. They appear most often as an estimate and standard error obtained by linear approximation methods. Their use is in fact many times more extensive...
Diagnostics in Linear Discriminant Analysis
1. INTRODUCTION Many articles have been published on the identification of regression outliers and influential observations in the last decade. But the corresponding research in multivariate analysis, such as the discriminant analysis, is much less....
Estimators of Odds Ratio Regression Parameters in Matched Case-Control Studies with Covariate Measurement Error
1. INTRODUCTION Case-control studies are often conducted to investigate the association between occurrence of disease and a specific binary risk factor ("exposure") of primary interest. Attention then focuses on the odds ratio between exposure and disease,...
Inference for Likelihood Ratio Ordering in the Two-Sample Problem
1. INTRODUCTION Stochastic ordering of distributions is an important concept in the theory of statistical inference. Many different types of stochastic ordering have been defined in the literature, and in fact a comprehensive volume from Academic Press...
Minimax Stimation of Proportions under Random Sample Size
1. INTRODUCTION Consider a finite population of known size M divided into k strata of unknown sizes [M.sub.i] (i = 1, ..., k). For the simultaneous estimation of the k proportions [M.sub.i]/M, a sample of size n is drawn by simple random sampling without...
Modeling Lifetime Data with Application to Fatigue Models
1. INTRODUCTION Consider, for example, a piece of chain, wire, or cable (used, for example, in building suspension bridges) which is subjected to a stress test until it breaks. The variable of interest here is the lifetime X of the piece when the stress...
Modeling Satellite Ozone Data
1. INTRODUCTION In recent years there has been an increasing concern over the decline in stratospheric ozone related to human activity Substantial ozone decreases in the spring season over Antarctica are now well documented (Farman, Gardiner, and Shanklin...
Modeling the Drop-Out Mechanism in Repeated-Measures Studies
Subjects often drop out of longitudinal studies prematurely, yielding unbalanced data with unequal numbers of measures for each subject. Modern software programs for handling unbalanced longitudinal data improve on methods that discard the incomplete...
Nonparametric Likelihood Ratio Estimation of Probabilities for Truncated Data
1. INTRODUCTION Truncated data is common in actuarial, demographic, epidemiologic, and other studies. Let (X, Y) be a pair of positive random variables with cumulative distribution functions [F.sub.0] and [G.sub.0]. We say that X is left truncated by...
Nonparametric Methods for Stratified Two-Sample Designs with Application to Multiclinic Trials
1. INTRODUCTION The design and analysis of multiclinic trials has been studied extensively in the literature. The case where the observed data can be modeled by the normal distribution has been considered in detail by Fleiss (1986). In many situations,...
Power Robustification of Approximately Linear Tests
1. INTRODUCTION Classical models for treatment effect often assume that the treatment produces a simple translation of the null distribution ([equivalent] control distribution without treatment). This idea has brought statisticians to develop and use...
Prediction and Creation of Smooth Curves for Temporally Correlated Longitudinal Data
1. INTRODUCTION Many studies involve collection of data at multiple time points for several individuals with the objective of producing a separate curve for each individual. These curves may be used as a visual descriptor, to obtain balanced data, to...
Selection of the K Largest Order Statistics for the Domain of Attraction of the Gumbel Distribution
1. INTRODUCTION The theory of extreme values and its applications have been extensively studied (e.g., Galambos 1987). Techniques for drawing inferences about the tail behavior of a distribution are well developed, and most are based on the extreme...
Simultaneous Confidence Bands for Linear Regression with Heteroscedastic Errors
1. INTRODUCTION Consider the linear regression model [Y.sub.i] = [[Beta].sub.0] + [summation of] [[Beta].sub.j][x.sub.ij] + [[Epsilon].sub.i] where i = 1 to p i = 1, . . ., n = f(x.sub.ij], . . ., [x.sub.ip]) + [[Epsilon].sub.i], where for simplicity...
Some Applications of Covariance Identities and Inequalities to Functions of Multivariate Normal Variables
1. INTRODUCTION The main purpose of this article is to point out some consequences and particular cases of the covariance identities and inequalities Obtained by Houdre and Perez-Abreu (1995) and Houdre and Kagan (1995). The framework of these papers...
Testing Homogeneity of Uniform Scale Distributions against Two-Sided and One-Sided Alternatives
1. INTRODUCTION AND SUMMARY Consider the model where [X.sub.ij], i = 1, 2, ..., k; j = 1, 2, ..., [n.sub.i] are independent uniform random variables with scale parameter [[Theta].sub.i] [greater than] 0; that is, the density of [X.sub.ij] is f[x.sub.ij]([x.sub.ij];...
Testing Ordered Alternatives in the Presence of Incomplete Data
1. INTRODUCTION Consider a randomized block experiment, [x.sub.ij] = [b.sub.i] + [[Tau].sub.j] + [e.sub.ij] i = 1, . . . , n j = 1, . . . , k, where k is the number of treatments, n is the number of blocks, and {[e.sub.ij]} are independent with identical...
Tests for Cointegration Based on Canonical Correlation Analysis
1. INTRODUCTION For I(1) processes, the Box-Tiao procedure (Box and Tiao 1977) for estimating cointegrating vectors produces canonical variates that asymptotically are nonstationary for variates corresponding to unit canonical correlations and are stationary...