Is regression adjustment supported by the Neyman model for causal inference?

This paper examines both theoretically and empirically whether the common practice of using OLS multivariate regression models to estimate average treatment effects (ATEs) under experimental designs is justified by the Neyman model for causal inference. Using data from eight large U.S. social policy experiments, the paper finds that estimated standard errors and significance levels for ATE estimators are similar under the OLS and Neyman models when baseline covariates are included in the models, even though theory suggests that this may not have been the case. This occurs primarily because treatment effects do not appear to vary substantially across study subjects.     

Ecology of urban lawns under three common management programs

Turfgrass lawns are a central part of urban and suburban landscapes throughout North America and are often managed using repeated applications of chemical fertilizers and pesticides. These inputs are expensive and may negatively affect ecological processes in lawns. Therefore, we evaluated the influence of three most common lawn care programs on ecological characteristics of turfgrass lawns. Twenty-eight home lawns, separated into 3 groups based on the lawn care program (professional, do-it-yourself [DIY], and no-input), were studied. Data on lawn quality, weed and insect infestation, disease incidence, soil nematode community, soil nitrogen pools, microbial biomass (MBN), and soil organic matter (SOM) were collected. Results indicated that professional lawn care resulted in the highest aesthetic lawn quality mainly due to better weed control, compared to DIY and no-input programs. However, professional and DIY programs negatively affected MBN and SOM pools and enhanced disease (rust) severity. No significant differences in soil nematode population and nematode community indices across the three programs were found, indicating no differences in net ecosystem productivity among the three programs. Overall, soil nematode food web in turfgrass lawns represented a disturbed food web compared to natural grasslands and forest ecosystems, irrespective of the lawn care program used.     

Simultaneous Confidence Intervals for the Difference of Two Survival Functions

In comparing two treatments with failure time observations, confidence bands for the "difference" of two survival curves provide useful information about a global picture of the treatment difference over time. In this note, we propose a rather simple procedure for constructing such simultaneous confidence intervals. Our technique can also be used in the one-sample case, which has been extensively studied in the literature.     

Computing Optimal Beams in Two and Three Dimensions

The problem of computing an optimal beam among weighted regions (called the optimal beam problem) arises in several applied areas such as radiation therapy, stereotactic brain surgery, medical surgery, geological exploration, manufacturing, and environmental engineering. In this paper, we present computational geometry techniques that enable us to develop efficient algorithms for solving various optimal beam problems among weighted regions in two and three dimensional spaces. In particular, we consider two types of problems: the covering problems (seeking an optimal beam to contain a specified target region), and the piercing problems (seeking an optimal beam of a fixed shape to pierce the target region). We investigate several versions of these problems, with a variety of beam shapes and target region shapes in 2-D and 3-D. Our algorithms are based on interesting combinations of computational geometry techniques and optimization methods, and transform the optimal beam problems to solving a collection of instances of certain special non-linear optimization problems. Our approach makes use of interesting geometric observations, such as utilizing some new features of Minkowski sums.     

ASYMPTOTIC EXPANSIONS APPLIED TO F-TEST CRITERIA

Using asymptotic expansions of the Kummer hypergeometric function, the sequential. F-test criterion is evaluated asymptotically in terms of the sample size. The continuation region inequalities for the test are inverted and expressed in terms of known test criteria. A rapidly converging algorithm for carrying out the sequential procedure is provided. This makes the F-test easier for the practitioner to use. Almost sure finite termination of the sequential. F-test is asserted by appealing to the continuation inequalities and a heuristic asymptotic expansion of the test criterion. Average stopping times of the sequential procedure for a variety of population means and population number configurations are tabulated. The computer symbolic manipulation program MAPLE was used to derive some formulae.     

When Infants Lose Exclusive Maternal Attention: Is It Jealousy?

To characterize infant reactions to jealousy evocation, 94 6‐month‐olds and their mothers were videotaped in an episode where the mothers directed positive attention toward a lifelike doll, and in 2 contrasting interactions: face‐to‐face play and a still‐face perturbation. Cross‐context comparisons of affects and behaviors revealed that jealousy evocation responses were distinguished by diminished joy and heightened anger and intensity of negative emotionality, comparable to levels displayed during the still‐face episode; heightened sadness, with durations exceeding those displayed during still‐face exposure; and an approach response consisting of interest, looks at mother, and diminished distancing, which was more pronounced than that demonstrated during play. Infants' heightened anger and sadness during jealousy evocation correlated with heightened maternal sensitivity and dyadic vocal turn‐taking, respectively, during play; and infants' diminished joy and interest during jealousy evocation were associated with heightened maternal withdrawal and intrusiveness, respectively, during play. Both fear and mother‐directed gaze were greater in girls. The discussion argues for interpreting the infant's mixed and agitated reaction to jealousy evocation as evidence of jealousy.     

Efficient Estimation in Marginal Partially Linear Models for Longitudinal/Clustered Data Using Splines

Abstract.  We consider marginal semiparametric partially linear models for longitudinal/clustered data and propose an estimation procedure based on a spline approximation of the non-parametric part of the model and an extension of the parametric marginal generalized estimating equations (GEE). Our estimates of both parametric part and non-parametric part of the model have properties parallel to those of parametric GEE, that is, the estimates are efficient if the covariance structure is correctly specified and they are still consistent and asymptotically normal even if the covariance structure is misspecified. By showing that our estimate achieves the semiparametric information bound, we actually establish the efficiency of estimating the parametric part of the model in a stronger sense than what is typically considered for GEE. The semiparametric efficiency of our estimate is obtained by assuming only conditional moment restrictions instead of the strict multivariate Gaussian error assumption.     

Genetic algorithms for numerical optimization

Genetic algorithms (GAs) are stochastic adaptive algorithms whose search method is based on simulation of natural genetic inheritance and Darwinian striving for survival. They can be used to find approximate solutions to numerical optimization problems in cases where finding the exact optimum is prohibitively expensive, or where no algorithm is known. However, such applications can encounter problems that sometimes delay, if not prevent, finding the optimal solutions with desired precision. In this paper we describe applications of GAs to numerical optimization, present three novel ways to handle such problems, and give some experimental results.     

Bayesian prediction for progressively censored data from the Burr model

Received: August 14, 2000; revised version: April 23, 2001