ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478728

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

摘要:

Three detailed case studies illustrating the seasonal analysis of economic time series are presented using component models for seasonal and nonseasonal behavior. Analyses are performed within a semi-Bayesian framework where inferences for target quantities of interest, such as seasonally adjusted values, are obtained as posterior distributions conditional on observed data and fitted parameter values. Such an approach is similar to previous model-based methods of seasonal analysis, but new models and algorithms are used and, more important, a sensitivity analysis is performed to determine the extent to which conclusions vary across a range of plausible fitted models. It is found that sensitivity to variation across plausible models is not unusual in practice. The logical conclusion of the investigation is that a fully Bayesian analysis is required that averages conditional posteriors over a posterior distribution for the model parameters. Such an analysis is necessarily sensitive to the choice of prior distribution. We believe that such dependencies on assumptions external to the data are inevitable in complex problems such as seasonal adjustment. The stochastic components of our nonseasonal and seasonal models are based on modified fractional Gaussian noise, as described in Carlin, Dempster, and Jonas (1985). The models allow for joint estimation of both fixed and random effects, a capability that is used to estimate the initial values of nonstationary components and is further illustrated in the data analysis by the use of trading-day adjustments. Our model structure is described in detail, but technical details of algorithms used to perform likelihood and conditional posterior calculations for fitted models are omitted in favor of the empirical case studies. The examples include some comparisons to the nonparametric X-11 method and to an autoregressive moving average modeling approach. One of the examples exhibits particularly striking differences between the model-based seasonal adjustments and those of X-11.

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478729

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478730

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478731

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478732

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478733

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478734

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478735

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

摘要:

This article reports results from a longitudinal study investigating the effects of low-to-moderate prenatal and postnatal lead exposure on the cognitive development of children during the first 18 months of life. Study hypotheses are expressed as a sequence of linear models for the outcome variable adjusted score on the Bayley Scales of Mental Development (MDIA), as a function of cord blood-lead concentration, infant blood-lead concentration at semiannual examinations, and other characteristics of study participants. These models are fitted to MDIA measurements on three occasions for as many as 214 infants, first assuming an arbitrary multivariate covariance structure for the repeated measurements and then with covariance structure arising from a random-effects model for errors. Estimates of the effects of lead exposure are not sensitive to the assumed covariance structure. The article describes several approaches to residual analysis and outlier detection in the longitudinal setting. In particular, it shows how empirical Bayes residuals can be used to estimate the partial regression coefficient of covariates not included in the linear model. The major findings concerning the effect of lead exposure on cognitive development are (a) a clear association between cord blood-lead concentration and Bayley scores in the first 18 months of life, (b) no clear evidence of an effect of cumulative postnatal exposure, and (c) a tendency for children who had higher lead concentrations at 6 months of age to exhibit poorer performance at 18 months than children with low lead concentrations. These findings have implications for acceptable blood-lead concentrations in children and pregnant women.

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478736

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor

摘要:

Log-linear models provide a flexible means of extending life table techniques for the analysis of censored survival data with categorical covariates, as discussed by Holford (1980) and Laird and Olivier (1981). We extend this methodology to incorporate cases in which one or more of the categorical covariates are sometimes missing. Maximum likelihood estimates of the parameters are calculated using data from all cases. This can result in large gains in efficiency over standard methods that require the exclusion of cases with incomplete data. With this approach, we assume that the hazard function, conditional on the covariates, is a stepwise function over disjoint intervals of time. The model has two parts: a log-linear model describing the hazard parameters, and a multinomial model describing the probabilities in the contingency table defined by the covariates. The main interest is in the model for the hazard parameters. We show how to calculate maximum likelihood estimates of parameters of the model either by an application of the EM algorithm in conjunction with one cycle of iterative proportional fitting in the M step or by using the Newton—Raphson algorithm. Estimates of standard errors are computed from the empirical information matrix. When using our proposed maximum likelihood approach, two additional assumptions are needed in addition to the usual assumptions of noninformative censoring. First, the mechanism causing missing covariates must be ignorable (Rubin 1976) in that the probability that a covariate is missing cannot depend on the covariate itself or on other covariates that are missing. The second assumption is that the distribution of the random censoring variable does not depend on any covariate that is missing. The first example, investigating the influence of several covariates on time to diagnosis of high blood pressure in a large cohort of men, shows clear gains in efficiency of our approach over analysis of complete cases and illustrates the flexibility of the log-linear approach. A second example of survival times of symptomatic and asymptomatic lymphoma patients shows interesting differences between the complete-case analysis and the maximum likelihood approach, which could be due to a nonrandom missing-value mechanism.

ISSN:0162-1459    

DOI:10.1080/01621459.1989.10478737

出版商:Taylor & Francis Group    

年代:1989

数据来源: Taylor