A one-way randomeffectsmodel is consideredfor the log-transformedshift-longpersonal exposure measurements, where the random effect in the model represents an effect due to the worker. Following a previousapproach,we address a hypothesis-testingproblem involving the proportionof workers for whom the mean exposure exceeds the occupational exposure limit. A conŽ dence interval is constructed for t...

This paper explores how observed and unobserved parental investments compensate for low birth weight. Controlling for family fixed effects, which encompass unobserved parental investment, we find birth weight positively correlates with math and reading scores and these estimates are considerably larger in magnitude than estimates derived from models that do not control for family fixed effects....

A simulation study is designed to explore accuracy and precision of parameter estimation of cognitive attributes using the crossed random effects linear logistic test model (CRELLTM) with PROC GLIMMIX. The CRELLTM characterizes the weight of the cognitive attributes for item difficulties as random coefficients. The SAS® GLIMMIX package is applied to obtain parameter estimates. A series of simul...

An important problem in multilevel modeling is what constitutes a sufficient sample size for accurate estimation. In multilevel analysis, the major restriction is often the higher-level sample size. In this paper, a simulation study is used to determine the influence of different sample sizes at the group level on the accuracy of the estimates (regression coefficients and variances) and their s...

Methods Web-based survey, containing discrete choice experiments (DCEs) to elicit preferences of three pre-defined groups of adults with epilepsy (n=750): (i) early epilepsy, (ii) established epilepsy, (iii) women of childbearing age (WOCBA). The DCEs contains five attributes, with two levels, defined using: semi-structured interviews with patients (n=56), a focus group with AED prescribers (n=...

The proportion of variance for service sire effect was estimated for three litter size traits (numbers of piglets born, born alive, and weaned) in Czech Large White (89 231 litters) and Czech Landrace (28 320 litters) pigs. Each trait in the first parity was considered as one trait and that trait in the second and subsequent parities was treated as a repeated trait. Consequently, three two-trai...

a r t i c l e i n f o The existing literature on the relationship between property rights in land and agricultural investment in Africa has given results that are often confusing and contradictory. I make two clarifying contributions to this literature. First, I pull together existing studies and investigate whether the results they find have been affected by research methods or local contexts....