Maria Michela Dickson

Incomplete geocoding and spatial sampling: The effects of locational errors on population total estimation

Due to the increasing availability of georeferenced microdata in several fields of research, surveys can benefit greatly from the use of the most recent spatial sampling methods. These methods allow to select spatially balanced samples, which lead to particularly efficient estimates, by incorporating the distances among the exact locations of statistical units into the design. Unfortunately, since locations of units are rarely exact in practice due to imperfections in the geocoding processes, the implementation of spatial sampling designs is actually often limited. This paper aims at demonstrating that spatial sampling designs can be implemented even when spatial information is not completely accurate. In particular, by means of a Montecarlo sampling simulation study about the estimation of water pollution, it is proved that the use of spatial sampling methods still lead to more spatially balanced samples, and more efficient estimates, also when the geocoding of population is not exact.

Effects of missing data and locational errors on spatial concentration measures based on Ripley’s K-function

ABSTRACT Effects of missing data and locational errors on spatial concentration measures based on Ripley’s K-function. Spatial Economic Analysis. Measures based on Ripley’s K-function are the preferred tools to test the concentration of individual agents in an economic space. In many empirical cases, however, the datasets contain different inaccuracies due to missing data or uncertainty about the location of the agents. Little is known thus far about the effects of these inaccuracies on the K-function. This paper sheds light on the problem through a theoretical analysis supported by Monte Carlo experiments. The results show that patterns of clustering or inhibition may be observed not as genuine phenomena but only as the effect of data imperfections.

Design-based estimation in environmental surveys with positional errors

The recent increased availability of information about the micro-geographic positions of population units in environmental surveys has led to important developments in spatial sampling methodologies and, as a result, has improved the estimation accuracy. In real data, however, information about the location of units is often affected by inaccuracy about their exact spatial positions, and these non-sampling errors can affect the estimation procedure. This paper aims to investigate the effects of positional errors on total estimation through a Monte-Carlo simulation study based on real populations of trees. Starting from perfect positioning, we examine two typical types of coarsening that frequently impact two different species of trees. The simulation results show that the exploitation of spatial information to estimate population totals continues to be relevant in the context of environmental surveys, even in the presence of inaccuracies.

A Graphical Tool for Interpreting Regression Coefficients of Trinomial Logit Models

ABSTRACT Multinomial logit (also termed multi-logit) models permit the analysis of the statistical relation between a categorical response variable and a set of explicative variables (called covariates or regressors). Although multinomial logit is widely used in both the social and economic sciences, the interpretation of regression coefficients may be tricky, as the effect of covariates on the probability distribution of the response variable is nonconstant and difficult to quantify. The ternary plots illustrated in this article aim at facilitating the interpretation of regression coefficients and permit the effect of covariates (either singularly or jointly considered) on the probability distribution of the dependent variable to be quantified. Ternary plots can be drawn both for ordered and for unordered categorical dependent variables, when the number of possible outcomes equals three (trinomial response variable); these plots allow not only to represent the covariate effects over the whole parameter space of the dependent variable but also to compare the covariate effects of any given individual profile. The method is illustrated and discussed through analysis of a dataset concerning the transition of master’s graduates of the University of Trento (Italy) from university to employment.

Likelihood-based risk estimation for variance-gamma models

Although the variance-gamma distribution is a flexible model for log-returns of financial assets, so far it has found rather limited applications in finance and risk management. One of the reasons is that maximum likelihood estimation of its parameters is not straightforward. We develop an EM-type algorithm based on Nitithumbundit and Chan (An ECM algorithm for skewed multivariate variance gamma distribution in normal mean–variance representation, arXiv:1504.01239, 2015) that bypasses the evaluation of the full likelihood, which may be difficult because the density is not in closed form and is unbounded for small values of the shape parameter. Moreover, we study the relative efficiency of our approach with respect to the maximum likelihood estimation procedures implemented in the VarianceGamma and ghyp R packages. Extensive simulation experiments and real-data analyses suggest that the multicycle ECM algorithm gives the best results in terms of root-mean-squared-error, for both parameter and value-at-risk estimation. The performance of the routines in the ghyp R package is similar but not as good, whereas the VarianceGamma package produces worse results, especially when the shape parameter is small.

La sopravvivenza immediata delle start-up italiane del settore manifatturiero sanitario: un’analisi multilevel

L’obiettivo del presente lavoro e quello di fornire nuove evidenze circa le determinanti della probabilita di sopravvivenza di breve periodo delle start-up italiane attive nel settore farmaceutico e nel settore della produzione di dispositivi medico-sanitari. Al fine di valutare l’effetto di caratteristiche specifiche delle singole imprese, e di tener conto delle variabili di contesto osservate e non osservate, la probabilita di sopravvivenza a tre anni viene descritta mediante un modello logistico multilevel. L’analisi si basa sulle osservazioni a livello di popolazione raccolte e gestite dall’ISTAT in conformita con le direttive dell’OCSE e di EUROSTAT sulla demografia d’impresa, in grado di garantire la coerenza delle informazioni raccolte con particolare riferimento alle entrate e alle uscite delle imprese dal mercato. L’elevato numero di effetti random e la conseguente elevata dimensionalita dell’in¬tegrazione richiesta dal processo di stima rendono le tecniche di stima standard poco affidabili. Le stime sono state quindi effettuate mediante il metodo del-l’entropia relativa per l’ottimizzazione di funzioni con rumore (Bee et al., 2015).

Metodi di campionamento spaziale per la selezione di campioni rappresentativi di imprese

La maggior parte degli studi campionari sulle imprese si fondano sulla selezione di campioni quanto piu possibile rappresentativi della popolazione oggetto di indagine. Tale popolazione puo, per esempio, essere costituita dalle imprese di un certo settore di attivita economica o dagli impianti produttivi (unita locali) che insistono in una determinata regione del paese di interesse per l’analisi. La prassi standard in quest’ambito e quella di affidarsi a disegni campionari stratificati, scelta dovuta sia alla grande adattabilita e velocita computazionale di questo metodo, sia al-l’alto livello di rappresentativita che esso garantisce. In alcune circostanze, pero, la stratificazione della popolazione risulta difficoltosa, soprattutto in presenza di un alto numero di strati o, al limite, di strati vuoti. In questi casi, un’alternativa al campionamento stratificato puo essere l’utilizzo delle recenti metodologie di campionamento spaziale. Mediante uno studio simulato condotto a partire da dati reali, in questo lavoro verra valutata l’efficienza delle metodologie di campionamento spaziale per la conduzione di indagini su imprese e si discutera dell’alta rappresentativita dei campioni cosi selezionati e delle condizioni che tale rappresentativita garantiscono.

Administrative burdens evaluation: A smart perspective in the Province of Trento

Since 2006, Red Tape has been a hot topic for the European Commission. All the European Countries agreed in reducing the costs that enterprises have to face due to Administrative Burdens (AB). Each Country has implemented his own simplification plans in order to move from the Red Tape to better regulation. Trentinos local governments both 1) issued laws to cut down the total amount of information obligations that the firms have to fulfill by legislation and 2) implemented a set of new online services to simplify the communication between firms and local PAs (Public Administrations). Even if e-government has proved to be the right answer to reduce AB in several EU Countries, it hides issues and pitfalls, especially in terms of accessibility, usage and user-friendliness, that could transform a best practice into a costly (for PAs and for firms too) unused online structure. By the means of a sample investigation, this research aims to gather data about the feelings of the entrepreneurs located in Trentino concerning the simplification actions and the online services undertaken by national and local governments.

Teaching statistics in the context of social foresight. An applied approach based on the use of an open-source software

Purpose – The purpose of this paper is to present the contents and the didactic approach that characterize, respectively, the “Introductory Statistics with R” and “Statistics and Foresight” courses of the Master in Social Foresight. Design/methodology/approach – The two courses “Introductory Statistics with R” and “Statistics and Foresight” are designed to provide an introduction to quantitative methods in the social sciences with specific applications to social foresight. In particular, the first course introduces students to data analysis providing the necessary tools to study and represent socio-economic phenomena through graphical summaries and numerical measures. During the course, example applications based on the use of the open-source software R are shown. At the end, the students should be able to perform data management, conduct descriptive analysis of categorical and quantitative variables and analyze bivariate distributions. The subsequent course “Statistics and Foresight” presents the most efficient methods to make decisions in a context of uncertainty while visualizing the potential errors of wrong decisions and computing the probability of their occurrence. Findings – This paper is a description of an interesting and promising way of teaching applied statistics in social sciences. Practical implications – With the main aim of learning the correct use of statistics, specific attention is devoted to the use and interpretation of the aforementioned methods rather than to their theoretical aspects. Even in the second course, an important role is played by the treatment of real data by the use of the R software. Originality/value – This paper attempts to systematize a method of teaching statistics based on the practical use of open-source software.