Showing papers in "Statistica in 2001"
TL;DR: This paper proposed an elicitation design and a sample selection model that allow correction of the possible bias due to protest votes in contingent valuation surveys, which can sensibly affect the estimate of the value of the public good.
Abstract: Observations with protest votes in contingent valuation surveys could result in biased estimates of the willingness to pay measure, if treated as expressions of genuine indifference to public good or simply removed from the analysis. We propose an elicitation design and a sample selection model that allow correction of the possible bias due to protest votes. An application shows that selection bias can sensibly affect the estimate of the value of the public good. Inference in sample selection models may sometimes be difficult due the flatness of the likelihood function that causes uncertainty about the presence of selection bias. Even so, we recommend use of the sample selection model, since it takes into account this uncertainty.
39 citations
TL;DR: This article used the Hansen and Hurwitz (1946) procedure for treating non-response and found it to be more efficient than the Hansen-Hurwitz estimator, which was used for sampling on two successive occasions when there is nonresponse in the sample.
Abstract: Estimators of the population mean have been presented for sampling on two successive occasions when there is non-response in the sample. These estimators have been derived using the Hansen and Hurwitz (1946) procedure for treating non-response. The estimators were found to be more efficient than the Hansen-Hurwitz estimator.
22 citations
TL;DR: In this paper, a class of mixture models is developed, in order to allow for the coexistence of different distributions of unemployment in the population, making possible an interpretation of long-term unemployment allowing for: (a) the estimation of the duration model referred only to individuals with some opportunity of transition to employment, thus eliminating the bias due to the joint estimation of two different populations; pointing out the distinctive features (as age, gender, education) related to the disadvantaged side of the population.
Abstract: In recent years, specifically in western European countries, it has been observed a notable increase in the number of people showing long duration of job search spells, and this aspect seems to concern mostly some specific groups of the population. Significantly, it has become very common to report besides the unemployment rate the level of long-term unemployment as an important index on the situation of labour markets. In this work a class of mixture models is developed, in order to allow for the coexistence of different distributions of unemployment in the population. These models make possible an interpretation of long-term unemployment allowing for: (a) the estimation of the duration model referred only to individuals with some opportunity of transition to employment, thus eliminating the bias due to the joint estimation of the mixture of two different populations; (b) pointing out the distinctive features (as age, gender, education) related to the disadvantaged side of the population. An empirical application of mixture models on a longitudinal sample coming from the rotating sample structure of the Italian labour force survey is presented, also considering the correction to take into account problems related to length bias.
17 citations
TL;DR: In this article, the effect of random non-response on the study as well as auxiliary variables was measured by considering a class of estimators for population mean, and it has been shown that the estimators defined under situation 1 are more efficient and more nonresponse robust than the class of estimation under situation 2.
Abstract: This paper measures the effect of random non-response (I) on the study as well as the auxiliary variables [Situation 1], and (II) on the study variable only [Situation 2] by considering a class of estimators for population mean. it has been shown that the class of estimators defined under situation 1 is more efficient as well as more non-response robust than the class of estimators defined under situation 2. This study also generalized the work of Tracy and Osahan (1994).
13 citations
TL;DR: In this paper, a hybrid of calibration and imputation has been considered in the presence of a random non-response in survey sampling, and a simulation study is carried out to show the performance of the proposed methods compared to the ratio method of imputation.
Abstract: In the present investigation, a hybrid of calibration and imputation has been considered in the presence of a random non-response in survey sampling. Estimators of population mean associated with the proposed hybrid remain unbiased under design based approach. A simulation study is carried out to show the performance of the proposed methods compared to the ratio method of imputation.
8 citations
TL;DR: In this article, a non-parametric approach to the estimation of the Gini inequality index is introduced, which consists in taking first a preliminary estimation of population density function, and then in computing the corresponding inequality index.
Abstract: In this paper a non-parametric approach to the estimation of the Gini inequality index is introduced. The basic idea consists in taking first a preliminary estimation of the population density function, and then in computing the corresponding inequality index. Statistical properties of the estimator introduced are studied, with particular reference to the case of large samples (usually available from statistical agencies). As a by-product, approximated confidence intervals are obtained.
5 citations
TL;DR: In this paper, the authors developed a Bayesian reliability model using a new loss function that has been introduced by Higgins and Tsokos and showed that the reliability estimates are in general sensitive to the choice of the prior probability distribution.
Abstract: The aim the present study is to develop a Bayesian reliability model using a new loss function that has been introduced by Higgins and Tsokos. The underlyng failure model is the exponential and four priors are being used: the inverted gamma , the general uniform, the exponential and the lognormal. The concept of efficienncy along with the notion of integrated mean square are being used as a criterion to analytically and numerically compare our findings with respect to the natural conjugate prior. It is shown that the Bayesian reliability estimates are in general sensitive to the choice of the prior probability distribution.
5 citations
TL;DR: This paper investigates the effects of using bilateral as opposed to unilateral landmarks, and of missing landmarks, on the ultimate biological interpretation of analyses and examines the degree of morphological integration amongst diverse anatomical regions in the face.
Abstract: This paper reviews some recent advances in the analysis of form variability amongst objects described in terms of landmark configurations. Emphasis is given to the methods of geometric morphometrics that utilise the geometry of Kendall's shape space. This space has well understood and desirable statistical properties. Geometric morphometric methods preserve the geometry of the landmark configurations at every step of analysis and so enable ready visualisation of statistical findings. The methods are illustrated by reference to two example analyses of biological shape variability within growth series of facial skeletons. The first is a study of growth and sexual dimorphism in monkeys and the second is a study of human facial growth. Since the material is biological it is to be expected that its morphogenesis generates correlations amongst landmarks. Thus, the face is bilateral symmetrical, with bilateral as opposed to unilateral landmarks likely contributing little new information about underlying patterns of growth. Further, any one landmark is unlikely to be sited on a region that shows a completely uncorrelated pattern of a variation with the regions on which the other landmarks are sited. This paper investigates the effects of using bilateral as opposed to unilateral landmarks, and of missing landmarks, on the ultimate biological interpretation of analyses. Our empirical findings in these studies to indeed indicate considerable stability of biological findings when diverse landmark sets are chosen. These findings reflect on the degree of morphological integration amongst diverse anatomical regions in the face.
5 citations
TL;DR: In this article, a new test of typicality in linear discriminant analysis was proposed, where the mean and the variance have to be evaluated from the data, and simulation were performed to check the level and the power of the test when the means and the variances are evaluated.
Abstract: Let us consider k populations having p-variate normal distributions with a cammon variance-convariance matrix where the parameters are known. Let X denote the vector of observations on an axperimental unit, of which the origin is uncertain. The problem of allocating X to one of the k given populations is fundamental in multivariate analysis. In many applications , it is also interesting to determine whether X is related to a new population or not. This paper proposes a new test of typicality in linear discriminant analysis. simulation were performed to check the level and the power of the test when the mean and the variance have to be evaluated from the data.
5 citations
TL;DR: In this article, the authors analyzed the relationship between reproductive behavior and working career of the woman on the basis on data deriving from the Fertility Survey carried out in Italy in 1995/96.
Abstract: The paper analyses the relationship between reproductive behaviour and working career of the woman on the basis on data deriving from the Fertility Survey carried out in Italy in 1995/96. In the Italian context, characterised both by the fertility in the world and a low women participation to the labour force, the study of the dynamic relationship between fertility and work participation is very relevant, considering the complex development of the Italian society, both for the women and labour market condition. By means of a logistic multilevel multiple-spell duration model, our analysis shows that the educational/work process influences the timing of the transition to motherhood: women tend to bear their first child after the conclusion of the educational process. Moreover, having a job after this process represents a cause of a greater delay in the moment of child bearing, but the final intensity of the process is quite similar whatever is the educational /working pattern. The working path changes both the timing and the final intensity of the second child, and these effects are quite different in the North , Centre and South of Italy. Our model enables us to use the casual approach correcting for the distortion in the parameter estimation deriving from the non-exogeneity of working career, taking into account possible common unobservable. The presence of a significant random working effect and of a significant correlation between the two random components of the model means that the effect of the working career on the fertility process depends on unobservable characteristics of each woman.
4 citations
TL;DR: A non parametric classifier which combines radial basis function networks (RBF networks) and binary classification trees is proposed and it is shown that from a classification tree a class of RBF networks can be derived, whose elements differ in the degree of dependence from the discriminant rules described by the tree.
Abstract: In this paper the use of neural networks in the solution of a classification problem is described and a non parametric classifier which combines radial basis function networks (RBF networks) and binary classification trees is proposed. It is shown that from a classification tree a class of RBF networks can be derived, whose elements differ in the degree of dependence from the discriminant rules described by the tree. A simulation study based on a two-class problem shows that the proposal represents a valid solution, particularly in presence of noisy variables.
TL;DR: In this article, two parameter-free tests for Cauchy distributions are constructed based on the specific form of the characteristic function and it is shown that asymptotically both test statistics follow an F 1, 1 − distribution under the null hypothesis and that they are consistent for symmetric alternatives.
Abstract: Two parameter-free tests for Cauchy distributions are constructed based on the specific form of the characteristic function. It is shown that asymptotically both test statistics follow an F 1, 1 - distribution under the null hypothesis and that they are consistent for symmetric alternatives. The performance of both methods in finite samples is investigated by employing simulated as well as real data.
TL;DR: In this article, it was shown that an incomplete VAR model, based on variables of the long-run reduced forms, produces an approximation of the cointegrating vectors coefficients and the misspecification of the short-run dynamics.
Abstract: This paper explores the problems arising when the reduced form cointegrating vectors are obtained from an incomplete VAR with omitted endogenous variables. It is shown that an incomplete VAR model, based on variables of the long-run reduced forms, produces an approximation of the cointegrating vectors coefficients and the misspecification of the short-run dynamics. It is also shown that this problems do not arise in certain circumstances. An example concerning the estimation of the NAIRU is provided suggesting the empirical relevance of our criticism.
TL;DR: In this paper, the authors propose two definitions for conditional analysis: one based on the notion of implicit function, the other based on invariance with respect to the permutation group.
Abstract: In this paper, after a brief critical analysis emphasizing the necessity that in conditional inferences actual data should also play a role in the definition of the parameter estimator or the test, we propose two definitions for conditional analysis: one in the parametric and one in the nonparametric framework. The former is based on the notion of implicit function, the latter on that of invariance with respect to the permutation group.
TL;DR: In this article, a non-parametric multivariate permutation method was used for the comparison of cancer growth patterns in animal models, where the analyzed data come from an experiment whose purpose consists of revealing the anticancer activity of a new taxol synthetic drug.
Abstract: This paper deals with an application of non parametric multivariate permutation methods (Pesarin, 1992, 1997, 1999) for the comparison of cancer growth patterns in animal models. The analyzed data come from an experiment whose purpose consists of revealing the anticancer activity of a new taxol synthetic drug. We are interested in testing if the growth inhibition effect due to treatment is statistically significant. To compare treated with untreated growths we face some multivariate testing problems related with repeated measures. A parametric approach for the solution of these hypotheses testing problems is not possible because in our case the number of observed units is less then the number of repeated measures.
TL;DR: The identification and estimate of a statistical model to analyse the SO2 in the city of Padua, where data are collected by some fixed stations and some mobile stations moving without any specific rule in different new locations, staying in every location for a variable number of days.
Abstract: Given atmospheric measurement from a network of monitoring sites in the area of a city and over an extended period of time, an important problem is to identify the spatial and temporal structure of data. In this paper we focus on the identification and estimate of a statistical model to analyse the SO2 in the city of Padua, where data are collected by some fixed stations and some mobile stations moving without any specific rule in different new locations, staying in every location for a variable number of days. The proposed method divides the global variability in large scale and small scale using some stochastic process as component of variability. The estimate is provided using a state space formulation of the model. As applications of the model we propose the spatial and temporal prevision of the concentration of SO2 . Finally, an exercise is proposed to choose an optimal network for the mobiles monitoring stations for a fixed future time.
TL;DR: In this paper, a new time series-based approach for testing consumption risk sharing among a set of counties or regions where both the long run and short run dynamic implications of the theory are simultaneously investigated within a Vector Equilibrium Correction (VEqC) model.
Abstract: In this paper it is proposed a new time series-based approach for testing consumption risk sharing among a set of counties or regions where both the long run and short run dynamic implications of the theory are simultaneously investigated within a Vector Equilibrium Correction (VEqC) model. This allows to measure whether countries or regions insure the non-aggregate uncertainty in their resources not only with respect to transitory shocks but also with respect to permanent ones . It is suggested that conventional risk sharing test have to be properly redefined to account for the possibility of error correcting dynamics and to avoid the possibility of "endogenous regressor bias" phenomena.
TL;DR: A forecasting procedure based on a specially devised neural network aimed at computing a quality parameter of next semen batch of a given bull from the preceding time series of the same parameter appears more accurate than usual ARIMA systems on a limited bench of experiments.
Abstract: We present a forecasting procedure based on a specially devised neural network aimed at computing a quality parameter of next semen batch of a given bull from the preceding time series of the same parameter. The procedure has its rationale in a sub symbolic issue of harmonic components analysis and appears more accurate than usual ARIMA systems on a limited bench of experiments.
TL;DR: In this article, the authors deal with permutation testing for balanced and unbalanced repeated measures designs and consider a replicated homoscedastic (balanced or unbalanced) factorial design with fixed effects as the basic experimental plan.
Abstract: We deal with permutation testing for balanced and unbalanced repeated measures designs and we consider a replicated homoscedastic (balanced or unbalanced) factorial design with fixed effects (Milliken et al., 1984) as the basic experimental plan. The design responses are measured in L time occasions. The usual linear model for responses is: y = {y ij(l) r = so that, the null hypothesis H 0 is true if all partial sub hypotheses are true. In order to test separately the overall null hypotheses, we jointly consider the L measures, then the permutation solution is based on the nonparametric combination methodology (Pesarin, 2001). It is worth noting that the new permutation approach, presented here, is exact and, being conditional to the sufficient statistic represented by the data matrix it does not require normality of error terms in the linear model for responses.
TL;DR: In this paper, the authors derived necessary and/or sufficient conditions for measurement in order to be a maximising one, according to some assumptions either on the dimension of the physical space of the system or on the kind of measurement to be applied.
Abstract: Braunstein and Caves (1994) showed that Helstrom's quantum information maximises expected Fisher information over all possible measurements on a given quantum system. For one-dimensional parameter pure state system, we derived necessary and/or sufficient conditions for measurement in order to be a maximising one, according to some assumptions either on the dimension of the physical space of the system or on the kind of measurement to be applied.
TL;DR: A statistical procedure is proposed to overcome the difficulties of comparison of nucleotide substitutions rates between DNA sequences and the sample is sometimes small, which can limit the use of statistical methods.
Abstract: The comparison of nucleotide substitutions rates between DNA sequences has two major difficulties: rates are result of pairwise comparisons between sequences so they are not independent, and the statistical distribution of this rates is unknown. Moreover the sample is sometimes small, which can limit the use of statistical methods. So, the comparisons of the evolutionary behaviour of genes through analysis of nucleotide substitution rates should be done using appropriate statistical approach. In the present study a statistical procedure is proposed to overcome these difficulties.
TL;DR: In this article, a monotone B-spline smoothing method is proposed as an appropriate and flexible tool for estimation of the term structure of interest rates, and numerical experiments with italian bond data demonstrate the effectiveness of the proposed approach.
Abstract: A monotone B-spline smoothing method is proposed as an appropriate and flexible tool for estimation of the term structure of interest rates. Numerical experiments with italian bond data demonstrate the effectiveness of the proposed approach.
TL;DR: In this paper, the problem of predicting the total test time, median and range of a future random sample from the one parameter exponential distribution on the basis of an observed type II censored sample was addressed.
Abstract: This paper is concerned with the problem of predicting the total test time, median and range of a future random sample from the one parameter exponential distribution on the basis of an observed type II censored sample from the same distribution. Empirical Bayes technique is used for this purpose. A numerical example is given to illustrate the procedures and the accuracy of prediction intervals is investigated via extensive Monte Carlo simulation.
TL;DR: In this article, a statistical approach is proposed to detect the model of sustainable development, and to attribute the different types of pollution to the economic factor of pressure, which is a partial solution to the problems of making a good classification.
Abstract: In some paper whose aim is to detect the model of sustainable development, it is simple to note that the actual models present in the System of national Accounts (SEC, SNA), are not able to give the ecessary information about the relation between economy and environment. The main approaches in this field such as SEEA, NAMEA, SERIEE, SIP, still present problems as the difficulties to make a good classification, and to attribute the different types of pollution to the economic factor of pressure. This paper shows a statistical approach finalized to a partial solution of these problems. In particular, it is analysed an Italian region (Basilicata) with sufficient environmental statistical information about one kind of pollution. The aim is to make some statistic multivariate analyses of territorial typizations, of the variability of the observed phenomenon of different kinds of the pollution existing in the territory and finally, of the procedure to identify the economic variables of pressure. The results are encouraging in giving new possibilities to analyse pollution problem with a new approach (whereas in past, the dates were analysed only with the threshold value of the single monitored variables). This is useful both for the building up of integrate model of "economic and environmental National Accounts" and for the individualization of relations between economic and antropic activity and the respective environmental output.
TL;DR: Brown and Boyce as discussed by the authors introduced some aspects of survey design for a monitoring scheme established for Karner blue butterflies (Lycaeides Melissa samuelis) in Wisconsin, USA.
Abstract: A technique commonly used for surveying butterflies is the Pollard - Yates method (Pollard, 1977; Pollard and Yates, 1993) This method was originally developed for surveying butterflies at Monks Wood, UK (Pollard et al, 1975) and is now widely used in the UK The method is based on a count of the number of butterflies seen along a fixed survey route The Pollard - Yates method is easy to implement in the field but it does not allow for differences in the ability to detect butterflies among sites Line transect sampling (Buckland et al, 1993) has similar survey methods to Pollard - Yates except that distance data is collected and used to estimate a detection function Therefore, one of the main advantages of line transect sampling is that it provides estimates of abundance that are comparable among sites regardless of the sites' "butterfly-detectability" In addition, it can be used to derive unbiased estimates of local-site densities provided key assumptions are met (Brown and Boyce, 1996, 1998) We introduce some aspects of survey design for a monitoring scheme established for Karner blue butterflies (Lycaeides Melissa samuelis) in Wisconsin, USA