Asymptotic robustness in multi-sample analysis of multivariate linear relations

Standard methods for the analysis of linear latent variable models oftenrely on the assumption that the vector of observed variables is normallydistributed. This normality assumption (NA) plays a crucial role inassessingoptimality of estimates, in computing standard errors, and in designinganasymptotic chi-square goodness-of-fit test. The asymptotic validity of NAinferences when the data deviates from normality has been calledasymptoticrobustness. In the present paper we extend previous work on asymptoticrobustnessto a general context of multi-sample analysis of linear latent variablemodels,with a latent component of the model allowed to be fixed across(hypothetical)sample replications, and with the asymptotic covariance matrix of thesamplemoments not necessarily finite. We will show that, under certainconditions,the matrix $\Gamma$ of asymptotic variances of the analyzed samplemomentscan be substituted by a matrix $\Omega$ that is a function only of thecross-product moments of the observed variables. The main advantage of thisis thatinferences based on $\Omega$ are readily available in standard softwareforcovariance structure analysis, and do not require to compute samplefourth-order moments. An illustration with simulated data in the context ofregressionwith errors in variables will be presented.

Statistics in archaeology: New directions

Connections between Statistics and Archaeology have always appeared veryfruitful. The objective of this paper is to offer an outlook of somestatistical techniques that are being developed in the most recentyears and that can be of interest for archaeologists in the short run.

Univariate versus multivariate modeling of panel data

Panel data can be arranged into a matrix in two ways, called 'long' and 'wide' formats (LFand WF). The two formats suggest two alternative model approaches for analyzing paneldata: (i) univariate regression with varying intercept; and (ii) multivariate regression withlatent variables (a particular case of structural equation model, SEM). The present papercompares the two approaches showing in which circumstances they yield equivalent?insome cases, even numerically equal?results. We show that the univariate approach givesresults equivalent to the multivariate approach when restrictions of time invariance (inthe paper, the TI assumption) are imposed on the parameters of the multivariate model.It is shown that the restrictions implicit in the univariate approach can be assessed bychi-square difference testing of two nested multivariate models. In addition, commontests encountered in the econometric analysis of panel data, such as the Hausman test, areshown to have an equivalent representation as chi-square difference tests. Commonalitiesand differences between the univariate and multivariate approaches are illustrated usingan empirical panel data set of firms' profitability as well as a simulated panel data.

Análisis morfométrico de la extremidad anterior en Carnivora

Durante los cuatro años de disfrute de la beca (2006 – 2009) se ha consolidado una base de datos de medidas osteológicas del esqueleto apendicular de numerosas especies del O. Carnivora. Concretamente, se han medido 364 individuos de 126 especies. Los ejemplares pertenecían a las colecciones del Phyletisches Museum (Jena, Alemania), el Museum für Naturkunde (Berlín, Alemania), el Museu de Ciències Naturals de la Ciutadella (Barcelona, España), el Múseum National d'Histoire Naturelle (París, Francia), y el Museo Nacional de Ciencias Naturales (Madrid, España). Asimismo, con estos datos se han estado preparando tres artículos sobre la morfología de ciertos elementos del esqueleto apendicular en carnívoros, dos de los cuales se encuentran actualmente en estado de revisión para su publicación científica. Dos de ellos, "Scapula, habitat and locomotion in Carnivora" y "Size and shape in the carnivore scapula", relacionan la morfología escapular con factores como el tamaño del animal, el tipo de locomoción que presenta y el hábitat en el que se encuentra; el primero mediante metodología multivariante (análisis funcional) y el segundo bajo las nuevas técnicas de morfometría geométrica. El tercer artículo, "Scaling and mechanics in the carnivore calcaneus: A comparison of natural and artificial selection", evalúa el efecto de diferentes tipos de selección, natural frente a artificial, sobre la morfología del calcáneo y su influencia en la biomecánica de este hueso. Finalmente, también se ha desarrollado un estudio experimental sobre la búsqueda de estabilidad durante la locomoción arbórea, cuyos resultados han dado lugar al artículo "The search for stability on narrow supports: An experimental study in cats and dogs", que también se halla bajo revisión actualmente.

Variability of North Atlantic hurricanes: seasonal versus individual-event features

Tropical cyclones are affected by a large number of climatic factors, which translates into complex patterns of occurrence. The variability of annual metrics of tropical-cyclone activity has been intensively studied, in particular since the sudden activation of the North Atlantic in the mid 1990’s. We provide first a swift overview on previous work by diverse authors about these annual metrics for the North-Atlantic basin, where the natural variability of the phenomenon, the existence of trends, the drawbacks of the records, and the influence of global warming have been the subject of interesting debates. Next, we present an alternative approach that does not focus on seasonal features but on the characteristics of single events [Corral et al., Nature Phys. 6, 693 (2010)]. It is argued that the individual-storm power dissipation index (PDI) constitutes a natural way to describe each event, and further, that the PDI statistics yields a robust law for the occurrence of tropical cyclones in terms of a power law. In this context, methods of fitting these distributions are discussed. As an important extension to this work we introduce a distribution function that models the whole range of the PDI density (excluding incompleteness effects at the smallest values), the gamma distribution, consisting in a powerlaw with an exponential decay at the tail. The characteristic scale of this decay, represented by the cutoff parameter, provides very valuable information on the finiteness size of the basin, via the largest values of the PDIs that the basin can sustain. We use the gamma fit to evaluate the influence of sea surface temperature (SST) on the occurrence of extreme PDI values, for which we find an increase around 50 % in the values of these basin-wide events for a 0.49 C SST average difference. Similar findings are observed for the effects of the positive phase of the Atlantic multidecadal oscillation and the number of hurricanes in a season on the PDI distribution. In the case of the El Niño Southern oscillation (ENSO), positive and negative values of the multivariate ENSO index do not have a significant effect on the PDI distribution; however, when only extreme values of the index are used, it is found that the presence of El Niño decreases the PDI of the most extreme hurricanes.

Segmentació d'imatge de ressonància magnètica de pit utilitzant atlas probabilístic

La localització d'òrgans és un tòpic important en l'àmbit de la imatge mèdica per l'ajuda del tractament i diagnosi del càncer. Un exemple es pot trobar en la cal•libració de models farmacoquinètics. Aquesta pot ésser realitzada utilitzant un teixit de referència, on, per exemple en imatges de ressonància magnètica de pit, una correcta segmentació del múscul pectoral és necessària per a la detecció de signes de malignitat. Els mètodes de segmentació basat en atlas han estat altament avaluats en imatge de ressonància magnètica de cervell, obtenint resultats satisfactoris. En aquest projecte, en col•laboració amb el el Diagnostic Image Analysis Group de la Radboud University Nijmegen Medical Centre i la supervisió del Dr. N.Karssemeijer, es presenta la primera aproximació d'un mètode de segmentació basat en atlas per segmentar els diferents teixits visibles en imatges de ressonància magnètica (T1) del pit femení. L'atlas consisteix en 5 estructures (teixit greixòs, teixit dens, cor, pulmons i múscul pectoral) i ha estat utilitzat en un algorisme de segmentació Bayesià per tal de delinear les esmentades estructures. A més a més, s'ha dut a terme una comparació entre un mètode de registre global i un de local, utilitzats tant en la construcció de l'atlas com en la fase de segmentació, essent el primer el que ha presentat millors resultats en termes d'eficiència i precisió. Per a l'avaluació, s'ha dut a terme una comparació visual i numèrica entre les segmentacions obtingudes i les realitzades manualment pels experts col•laboradors. Pel que fa a la numèrica, s'ha emprat el coeficient de similitud de Dice ( mesura que dóna valors entre 0 i 1, on 0 significa no similitud i 1 similitud màxima) i s'ha obtingut una mitjana general de 0.8. Aquest resultat confirma la validesa del mètode presentat per a la segmentació d'imatges de ressonància magnètica del pit.

Coherent forecasting of multiple-decrement life tables: a test using Japanese cause of death data

Planners in public and private institutions would like coherent forecasts of the components of age-specic mortality, such as causes of death. This has been di cult toachieve because the relative values of the forecast components often fail to behave ina way that is coherent with historical experience. In addition, when the group forecasts are combined the result is often incompatible with an all-groups forecast. It hasbeen shown that cause-specic mortality forecasts are pessimistic when compared withall-cause forecasts (Wilmoth, 1995). This paper abandons the conventional approachof using log mortality rates and forecasts the density of deaths in the life table. Sincethese values obey a unit sum constraint for both conventional single-decrement life tables (only one absorbing state) and multiple-decrement tables (more than one absorbingstate), they are intrinsically relative rather than absolute values across decrements aswell as ages. Using the methods of Compositional Data Analysis pioneered by Aitchison(1986), death densities are transformed into the real space so that the full range of multivariate statistics can be applied, then back-transformed to positive values so that theunit sum constraint is honoured. The structure of the best-known, single-decrementmortality-rate forecasting model, devised by Lee and Carter (1992), is expressed incompositional form and the results from the two models are compared. The compositional model is extended to a multiple-decrement form and used to forecast mortalityby cause of death for Japan

Incongruence between test statistics and P values in medical papers

Given an observed test statistic and its degrees of freedom, one may compute the observed P value with most statistical packages. It is unknown to what extent test statistics and P values are congruent in published medical papers. Methods:We checked the congruence of statistical results reported in all the papers of volumes 409–412 of Nature (2001) and a random sample of 63 results from volumes 322–323 of BMJ (2001). We also tested whether the frequencies of the last digit of a sample of 610 test statistics deviated from a uniform distribution (i.e., equally probable digits).Results: 11.6% (21 of 181) and 11.1% (7 of 63) of the statistical results published in Nature and BMJ respectively during 2001 were incongruent, probably mostly due to rounding, transcription, or type-setting errors. At least one such error appeared in 38% and 25% of the papers of Nature and BMJ, respectively. In 12% of the cases, the significance level might change one or more orders of magnitude. The frequencies of the last digit of statistics deviated from the uniform distribution and suggested digit preference in rounding and reporting.Conclusions: this incongruence of test statistics and P values is another example that statistical practice is generally poor, even in the most renowned scientific journals, and that quality of papers should be more controlled and valued

A subsampling approach to estimating the distribution of diversing statistics with application to assessing financial market risks

In this paper we propose a subsampling estimator for the distribution ofstatistics diverging at either known rates when the underlying timeseries in strictly stationary abd strong mixing. Based on our results weprovide a detailed discussion how to estimate extreme order statisticswith dependent data and present two applications to assessing financialmarket risk. Our method performs well in estimating Value at Risk andprovides a superior alternative to Hill's estimator in operationalizingSafety First portofolio selection.

Canonical correspondence analysis in social science research

The use of simple and multiple correspondence analysis is well-established in socialscience research for understanding relationships between two or more categorical variables.By contrast, canonical correspondence analysis, which is a correspondence analysis with linearrestrictions on the solution, has become one of the most popular multivariate techniques inecological research. Multivariate ecological data typically consist of frequencies of observedspecies across a set of sampling locations, as well as a set of observed environmental variablesat the same locations. In this context the principal dimensions of the biological variables aresought in a space that is constrained to be related to the environmental variables. Thisrestricted form of correspondence analysis has many uses in social science research as well,as is demonstrated in this paper. We first illustrate the result that canonical correspondenceanalysis of an indicator matrix, restricted to be related an external categorical variable, reducesto a simple correspondence analysis of a set of concatenated (or stacked ) tables. Then weshow how canonical correspondence analysis can be used to focus on, or partial out, aparticular set of response categories in sample survey data. For example, the method can beused to partial out the influence of missing responses, which usually dominate the results of amultiple correspondence analysis.

A simple permutation test for clusteredness

Hierarchical clustering is a popular method for finding structure in multivariate data,resulting in a binary tree constructed on the particular objects of the study, usually samplingunits. The user faces the decision where to cut the binary tree in order to determine the numberof clusters to interpret and there are various ad hoc rules for arriving at a decision. A simplepermutation test is presented that diagnoses whether non-random levels of clustering are presentin the set of objects and, if so, indicates the specific level at which the tree can be cut. The test isvalidated against random matrices to verify the type I error probability and a power study isperformed on data sets with known clusteredness to study the type II error.

Asymptotic robust inferences in the analysis of mean and covariance structures

Structural equation models are widely used in economic, socialand behavioral studies to analyze linear interrelationships amongvariables, some of which may be unobservable or subject to measurementerror. Alternative estimation methods that exploit different distributionalassumptions are now available. The present paper deals with issues ofasymptotic statistical inferences, such as the evaluation of standarderrors of estimates and chi--square goodness--of--fit statistics,in the general context of mean and covariance structures. The emphasisis on drawing correct statistical inferences regardless of thedistribution of the data and the method of estimation employed. A(distribution--free) consistent estimate of $\Gamma$, the matrix ofasymptotic variances of the vector of sample second--order moments,will be used to compute robust standard errors and a robust chi--squaregoodness--of--fit squares. Simple modifications of the usual estimateof $\Gamma$ will also permit correct inferences in the case of multi--stage complex samples. We will also discuss the conditions under which,regardless of the distribution of the data, one can rely on the usual(non--robust) inferential statistics. Finally, a multivariate regressionmodel with errors--in--variables will be used to illustrate, by meansof simulated data, various theoretical aspects of the paper.

Geometric and long run aspects of Granger causality

This paper extends multivariate Granger causality to take into account the subspacesalong which Granger causality occurs as well as long run Granger causality. The propertiesof these new notions of Granger causality, along with the requisite restrictions, are derivedand extensively studied for a wide variety of time series processes including linear invertibleprocess and VARMA. Using the proposed extensions, the paper demonstrates that: (i) meanreversion in L2 is an instance of long run Granger non-causality, (ii) cointegration is a specialcase of long run Granger non-causality along a subspace, (iii) controllability is a special caseof Granger causality, and finally (iv) linear rational expectations entail (possibly testable)Granger causality restriction along subspaces.

Unfolding a symmetric matrix

Graphical displays which show inter--sample distances are importantfor the interpretation and presentation of multivariate data. Except whenthe displays are two--dimensional, however, they are often difficult tovisualize as a whole. A device, based on multidimensional unfolding, isdescribed for presenting some intrinsically high--dimensional displays infewer, usually two, dimensions. This goal is achieved by representing eachsample by a pair of points, say $R_i$ and $r_i$, so that a theoreticaldistance between the $i$-th and $j$-th samples is represented twice, onceby the distance between $R_i$ and $r_j$ and once by the distance between$R_j$ and $r_i$. Self--distances between $R_i$ and $r_i$ need not be zero.The mathematical conditions for unfolding to exhibit symmetry are established.Algorithms for finding approximate fits, not constrained to be symmetric,are discussed and some examples are given.