ELENA POLI

I teach in two Undergraduate Courses: Statistics and Quantitative Methods for Finance and two Master Courses: Data Analysis and Principles of Business Analytics.

I am a Lecturer of Statistics in the Decision Sciences Department at Bocconi University, Milan. I obtained my master degree in 1998 in Economy at Università degli Studi di Brescia and my PhD in Methodological Statistics in 2002, at the University of Trento and Milano Bicocca after working at my PhD  thesis at Department of Social Statistics, Institute of Social Sciences, University of Southampton, Southampton (UK).

I have worked for many years at the Quantitative Methods Department, University of Brescia, at the CUSSB, University Centre for Statistics in the Biomedical Sciences, Vita e Salute San Raffaele University, Milan and since 2009 at Bocconi University.

Among the other teaching engagements, I found a very funny and formative experience the realization of the recordings for the fully on-line class of Statistics course.

I devoted particular attention to the study of non-linear multilevel models, particularly to models with ordered categorical response variables for which I studied the comparison among different estimation methods for the regression parameters (the Gauss-Hermite numerical integration method, MQL and PQL methods, and methods based on MCMC simulation). From the application point of view, I focused on two-level models to assess satisfaction and its determinants in the social services sector.

Another field of study of interest is that relating to distances between ranks, in particular Cayley, Kendall and Hamming distances. The focus, in this area, has been identified in position and dispersion measurements for unimodal and multimodal frequency distributions.

In the biomedical field, great attention has been devoted to:

-comparison of classification model performance and development of an assessment tool by appropriate modification of the weighted Cronbach's Alpha index.

- description of multidimensional categorical data and application of non-linear mixed-effects models to the evaluation of alternative treatments

-use of permutation models for the treatment of longitudinal data.

I am also interested in the detection and evaluation of individual perceptions by means of appropriate questions and survey scales, in order to properly evaluate latent variable