Showing papers in "Statistica in 2000"
TL;DR: In this article, consumption equivalence scales have been obtained from the parameters of a few alternative version of a rank 3 complete demand system with demand shifters to measure the relative levels of spending required by households of different composition to attain given levels of utility.
Abstract: This paper utilises unit record survey data on expenditures and incomes of Italian households of different demographic profiles to estimate consumption equivalence scales which measure the relative levels of spending required by households of different composition to attain given levels of utility. The equivalence scales have been obtained from the parameters of few alternative version of a rank 3 complete demand system with demand shifters.Among the demographic variables included in the models, the type of housing tenure and the age of children have special relevance. We have also computed the expenditure elasticities for the most common types of households and analysed how deflating expenditure with the equivalence scales affects consumption inequality among households. The analysis is developed within the framework of hypothesised utility maximising behaviour of households.
53 citations
TL;DR: In this paper, a modification of the Warner's randomized response procedure is considered and the modified estimator of the proportion of the population belonging to sensitive groups is shown to be unbiased and always more efficient than the original estimator.
Abstract: In the present study, a modification of Warner’s (1965) randomized response procedure is considered. For this case, the modified estimator of ... [Pi Greek], the proportion of the population belonging to sensitive group, is shown to be unbiased and always more efficient than the Warner’s estimator. Under certain conditions, this estimator is also more efficient than the estimator proposed by Mangat et al. (1995)
29 citations
TL;DR: In this paper, three regression type estimators are suggested for the estimation of finite population mean in two situations of random non-response considered by Tracy and Osahan (1994), the results are obtained under the assumption that the number of sampling units on wich information could not be obtained due to random nonresponse follows some distribution.
Abstract: In this paper, three regression type estimators are suggested for the estimation of finite population mean in two situations of random non-response considered by Tracy and Osahan (1994). The results are obtained under the assumption that the number of sampling units on wich information could not be obtained due to random non-response follows some distribution. Some suggestions are given for estimation of the mean square error of the estimators proposed.
23 citations
TL;DR: In this article, a simple method is proposed for obtaining information about attitude from nonresponse that is based on a latent variable model for nominal responses, the analysis treats nonresponse as a separate response category and then fits a model to the set of attitudinal items to provide information about the relationships between attitude and the probability of not responding to an item in the scale.
Abstract: This paper discusses the problem of missing values in attitude scales with categorical items. A simple method is proposed for obtaining information about attitude from nonresponse that is based on a latent variable model for nominal responses, The analysis treats nonresponse as a separate response category and then fits a latent variable model to the set of attitudinal items. The model coefficients provide information about the relationships between attitude and the probability of not responding to an item in the scale. Graphical methods based on response probabilities and posterior probabilities are used to reveal any relationship between attitude and missing values.
22 citations
TL;DR: In this paper, the authors discuss the weighting systems of several linear and non-linear smoothers and evaluate their predictive performances when applied to noisy time series, and illustrate with three Canadian leading indicators which are representative of larger sets of time series characterised by a low, medium and high signal to noise ratio.
Abstract: The purpose of this study is to discuss the weighting systems of several linear and non linear smoothers and to evaluate their predictive performances when applied to noisy time series. On this regard, we illustrate with three Canadian leading indicators which are representative of larger sets of time series characterised by a low, medium and high signal to noise ratio. The smoothers discussed are: (a) loess (a locally weighted regression smoother), (b) Gaussian Kernel smoother, (c) supersmoother, (d) cubic smoothing spline and (e) Dagum’s modified 13-term Henderson filter. Their performances are evaluated on the basis of three following main criteria for current economic analysis: (1) number of unwanted ripples or false turning points in the final estimated trend, (2) time lag in detecting ‘true’ turning points and (3) size of total revision of the concurrent trend estimates.
15 citations
TL;DR: In this article, the authors considered a complex strongly harmonizable stationary symmetric stable process in continuous time and gave an unbiased consistent estimate of the spectral density for this process sampling at periodic instants.
Abstract: We considered a complex strongly harmonizable stationary symmetric stable process in continuous time. In this paper we give an unbiased consistent estimate of the spectral density for this process sampling at periodic instants.To avoid aliasing phenomenon it is assumed that the spectral density has a compact support.
12 citations
TL;DR: In this paper, the single and product moments of order statistics from a generalized exponential distribution have been obtained for all sample size and for any choice of the shape parameter, which can then be used to derive the best linear unbiased estimators (BLUEs) of location and scale parameters from complete as well as Type II censored samples.
Abstract: The single and the product moments of order statistics from a generalized exponential distribution have been obtained for all sample size and for any choice of the shape parameter ... [alpha] . These moments can then be used to derive the best linear unbiased estimators (BLUE’s) of location and scale parameters from complete as well as Type II censored samples. In addition, percentage points of order statistics are also discussed.
9 citations
TL;DR: This article proposed a model to correct gross flows in the labour market from classification errors, when flows are observed by means of longitudinal surveys with retrospective questions, which is a special case of the modified LISREL approach proposed by Hagenaars.
Abstract: The paper proposes a model to correct gross flows in the labour market from classification errors, when flows are observed by means of longitudinal surveys with retrospective questions. The model consists in an application of latent class analysis, specifically it is a special case of the modified LISREL approach proposed by Hagenaars. The loglinear parametrization allows one to specify a very parsimonious model and to introduce into its measurement part a dependence of response errors on the time elapsed between the moment of interview and the event to recall. In this way, it is possible to consider the effect of memory decay on reported conditions in the labour market and , specifically, the fact that the longer the recall period, the higher the probability of making mistakes. The model is applied to correct quarterly gross flows observed by means of the French labour force survey.
8 citations
TL;DR: In this article, the authors consider the problem of studing the properties of the expanded finite Fourier transform for a zero mean, r-dimensional, strictly stationary and continuous time series using data window.
Abstract: We consider the problem of studing the properties of the expanded finite Fourier transform for a zero mean, r-dimensional, strictly stationary and continuous time series using data window. The properties of the kernel ... of a stationary real valued process continuous time series are studied. The covariance, dispersion and cumulants for this transform are determined. Furthermore, the limits of the above quantities as t increases are given
7 citations
TL;DR: In this article, an overview of the most important generalized functions for mathematical statistics is presented, along with several interesting novel results, which may prove useful expecially in harmonizable process analysis.
Abstract: This article sketches out an overview of the most important generalized functions for mathematical statistics. The paper on the one hand, collects and organizes known results in comprehensive setting and, on the other hand, comes up with several somewhat intriguing novel results, which may prove useful expecially in harmonizable process analysis.
6 citations
TL;DR: In this article, a permutation solution to an isotonic inference problem is presented for association case-control studies in genetic epidemiology, where the data may be represented as: X =....... where responses are ordered categorical such as (AA, aA, aa) or the like, where ordering is generally induced by the nature of the problem at hand.
Abstract: In this paper we present a permutation solution to an isotonic inference problem very usual in association case-control studies in genetic epidemiology. In similar studies, the data may be represented as: X = ....... where responses are ordered categorical such as (AA, aA, aa) or the like, where ordering is generally induced by the nature of the problem at hand. The hypotheses we are interested in are: where at least one inequality is strong. This particular system of hypotheses arises when the variables are such that, under the alternative, both X 11 and X 12 stochastically non-increase or both stochastically non-decrease. The permutation solution presented is based on the nonparametric combination of dependent test (Pesarin, 1992, 1999).
TL;DR: In this article, the authors deal with some nonparametric methods for the construction of confidence sets; the empirical likelihood, the empirical saddlepoint and the non-parametric tilting.
Abstract: The paper deals with some nonparametric methods for the construction of confidence sets; the empirical likelihood, the empirical saddlepoint and the nonparametric tilting. After introducing the basic principles of each method, some refined or alternative versions are considered. The paper gives an account of different aspects, advantages and drawbacks of these methods. Furthermore, examples show their applications and provide computational details. Finally, comparisons are carried out and analogous approaches and analytical links are remarked.
TL;DR: In this paper, the authors show how to extend previous models for contingency tables by considering simultaneously both equality constraints and inequality constraints, and also give a self contained discussion of the problem of testing equality and inequality constraint in the context of contingency tables analysis.
Abstract: Currently used methods for the analysis of contingency tables with log-linear models and their extensions are essentially a way to subject the parameters of the saturated model to a set of linear constraints. However, in many contexts, expecially when a subset of the variables are ordinal, it would be very useful to be able to consider also linear inequality constraints. In this paper we show how to extend previous models for contingency tables by considering simultaneously both equality constraints. We also give a self contained discussion of the problem of testing equality and inequality constraints in the context of contingency tables analysis.
TL;DR: In this paper, the relationship between sufficiency, invariance and conditional indipendence is investigated. But the results are restricted to the case of a sufficient [sigma]-field.
Abstract: This paper is concerned with the relationship between sufficiency, invariance and conditional indipendence. Given a sufficient [sigma]-field, two results are obtained about independence and conditional independence of it and the almost invariant [sigma]-field. Moreover, it includes a rigorous proof of a stronger result than a Stein theorem about sufficiency and invariance in its version for statistics and in the discrete case. Somewhat of a strange condition is used to prove this theorem in its original form; several results are given here to clarify it.
TL;DR: In this paper, the authors defined new indices of sample shape ( skewness and kurtosis) by means of letter values and applied them to a set of geographical data, with some comments about the effect of outliers.
Abstract: In this paper some new indices of sample shape ( skewness and kurtosis) are defined by means of letter values (Turkey, 1977). The letter index of skewness (l.c.s.) is built by fitting the standardized sample midsummaries with a least squares straight line, while the letter index of kurtosis (l.c.k.) is obtained by fitting the pseudosigmas (after a simple transformation), in the same way. A robust modification of the above defined indices is proposed. Finally, an application to a set of geographical data is given, with some comments about the effect of outliers.
TL;DR: In this article, the authors considered the problem of recovering a discrete probability distribution taking on a countable value, when only partial information is available, provided by a finite number of available generalized moments, and proved that the necessary and sufficient conditions for the existence of a maximun entropy distribution are identical to the ones of the reduced moment problem.
Abstract: The recovering of a discrete probability distribution taking on a countable values, when only partial information is available, is considered. The partial information is provided by a finite number of available generalized moments, so that the problem of recovering resorts to an underdetermined generalized moment problem. Maximun entropy method is invoked in choosing the analytical form of the approximant distribution having the same moments. It is proved that except the cases in wich M = 2 or M = 3 generalized moments are assigned, the necessary and sufficient conditions for the existence of a maximun entropy distribution are identical to the ones of the reduced moment problem.
TL;DR: In this article, a general model accounting for income, net and total wealth, dept and human capital distribution is presented and the Lorenz curve and Gini ratio can be obtained as function of the parameters of the general model.
Abstract: This paper specifies a general model accounting for income,net and total wealth, dept and human capital distribution and discusses five fondamental properties that models of these variables should fulfill.It is shown that the Lorenz curve and Gini ratio can be obtained as function of the parameters of the general model. Finally the general model and its particular cases are fitted to 1983 U.S. and 1989 Italian data and their resuls are thoroughly analyzed.
TL;DR: In this paper, a critical view on Bayesian conditional analysis and its relationship with the global analysis is presented, and it is shown that "posterior analysis" is nothing else than a global analysis carried out point by point.
Abstract: This paper deals with a critical view on Bayesian conditional analysis, usually known as “posterior analysis”, and its relationship with the “global analysis”. It is shown that “posterior analysis” is nothing else than a global analysis carried out point by point.
TL;DR: Basic statistical aspects which characterize feed forward artificial neural networks with specific reference to biostatistical problems are introduced and the extension of Generalized Linear Models as neural networks for processing censored survival data will be considered.
Abstract: The application of artificial neural networks for statistical modelling has been diffused in several fields This class of models is oriented to inductive inference: the estimate of an unknown functional dependence relationship on the basis of a limited number of experimental observations The largest developments have been achieved for multivariate non-linear regression, by the adoption of feed forward artificial neural networks for the flexible modelling of the effects of the independent variables On the hand, methodological studies have enlighten good statistical properties of these models; on the other hand several eurystic applications, mainly in the biomedical field, have stimulated some criticism This paper introduces basic statistical aspects which characterize feed forward artificial neural networks with specific reference to biostatistical problems The extension of Generalized Linear Models as neural networks for processing censored survival data will be considered Limits of the eurystic use and advantages of the integration with traditional statistical models will be finally discussed
TL;DR: In this paper, a version of casual log-linear models (Drovandi, 1998a), has been applied to demographic data (INF-2) with the introduction of macro indicators, deducted by other fonts (Racioppi and Santagata, 1998), that would be treated as constant covariate, to be considered explanatory variables both in terms of timing and intensity of reproductive process.
Abstract: Many authors are concerned with the analysis of causal relationships among fertility and working careers (Drovandi 1998a end 1998b; Rampichini e Salvini 1998), but until now they have not treated contextual indicators. Otherwise in the real situation the framework is more complex because there are two different dimensions: the micro and macro level that develop causal relationship among variables considered in the social and demographic context. As result of this in this paper, a version of casual log-linear models (Drovandi, 1998a), has been applied to demographic data (INF-2) with the introduction of macro indicators, deducted by other fonts (Racioppi and Santagata, 1998), that would be treated as constant covariate, to be considered explanatory variables both in terms of timing and intensity of reproductive process. According to this some socio-economic macro variables, concerned the residential area (Istat 1994,1995a,1995b), will be considered. These results show that the macro variables can explain, in the same way of individual indicators, same features of reproductive behaviour of women interviewed with the INF/2 Survey. This result can justify the use of Population Census, to develop proxies of women reproductive behaviour, when it is difficult to obtain individual information.
TL;DR: Results from data collected from patients registered in the Congenital Heart Disease Registry at KFSH&RC, Riyadh, Saudi Arabia show an expected number of parents who described a pattern of relationship consistent with a first cousin relationship but fewer numbers of 'other related' than previously reported in the literature.
Abstract: A method has been developed which documents consanguineous relationships in any population. Results from data collected from patients registered in the Congenital Heart Disease Registry at KFSH&RC, Riyadh, Saudi Arabia show an expected number of parents who described a pattern of relationship consistent with a first cousin relationship but fewer numbers of 'other related' than previously reported in the literature. Further studies are in preparation.
TL;DR: The paper takes the point of view of an analytical user wanting to (collect and chiefly) process micro data for statistical/research purposes and focuses on some basic features of the issue and on their policy implications.
Abstract: The issue of privacy versus data access is at the crossroads where two sets of considerations meet: (I) the traditional concern of official statistical agencies for confidentiality, both for ethical reasons and as a guarantee of maintaining respondents’ collaboration; (II) the more recent and broader concern about data protection, that extends to all kinds of personal data and their uses. The paper takes the point of view of an analytical user, i.e. of a person wanting to (collect and chiefly) process micro data for statistical/research purposes. It focuses on some basic features of the issue and on their policy implications. First of all the societal role of statistics and research and their intrinsic needs are considered (section 2). Indeed, they are the basis for motivating a ‘credit of confidentiality ’ when personal data are processed for statistical/research purposes. Some basic principles and guideline for regulating this credit of confidentiality, as they are set out in recent international recommendations and regulations, are then summarised (section 3). Statistical, technological and legal devices, and related practices, for protecting confidentiality while allowing the processing of micro data for statistical/research purposes are briefly reviewed (section 4). The state of the affairs at the level of the EU is finally discussed, as regard both rules and practices followed by Eurostat and draft legislation (Section 5). The concluding remarks stress the need for additional efforts to make micro databases widely available to researchers.
TL;DR: In this article, the proposed selection procedure can be implemented even when the underlying life distributions belong to different families, and some applications of such a selection procedure are discussed by taking gamma, exponential and Lehmann type distributions.
Abstract: ... The proposed selection procedure can be implemented even when the underlying life distributions belong to different families. Some applications of the proposed selection procedure are discussed by taking gamma, exponential and Lehmann type distributions.
TL;DR: In this paper, two transformed ratio estimators have been considered with Sen-Midzuno scheme of sampling and their efficiencies have been compared under a super population model under the Sen-midzuno sampling scheme.
Abstract: In this paper, two transformed ratio estimators have been considered with Sen-Midzuno scheme of sampling and their efficiencies have been compared under a super population model.
TL;DR: In this paper, the estimation of the slope parameter is considered by three popular approaches, viz.,minimization of vertical, horizontal and perpendicular distances between the data points and the line, and asymptotic properties of the three estimators are analyzed when errors are not necessarily normally distributed.
Abstract: In a simple linear regression model with measurement errors, the estimation of the slope parameter is considered by three popular approaches, viz.,minimization of vertical, horizontal and perpendicular distances between the data points and the line. Asymptotic properties of the three estimators are analyzed when errors are not necessarily normally distributed.