Final conference of the research project

Advances in Multilevel
and Longitudinal Modelling

January 28, 2019
Department of Statistical Sciences, University of Padua

Room SC140


09.30 Registration
10.00 Welcome address

10.20 Keynote speech: Harvey Goldstein
An approach to fitting generalised linear models with both missing data and errors of measurement in multilevel data

11.20 Coffee Break

     Session 1: Multilevel
11.50 Work 1: L. Grilli, O. Paccagnella, C. Rampichini, R. Varriale
Aggregating level-1 information for imputing level-2 missing data

12.10 Work 2: L. Grilli, M.F. Marino, O. Paccagnella, C. Rampichini
Multiple imputation and selection of ordinal level-2 predictors in multilevel models

12.30 Work 3: M. Bernardi, O. Paccagnella
Bayesian Predictive Mean Matching imputation in multilevel models

12.50 General Discussion
13.20 Lunch

14.20 Keynote speech: Gerko Vink
The Dance of the Mechanisms: How Observed Information Influences the Validity of Missingness Assumptionsi

15.20 Coffee Break

     Session 2: Longitudinal
15.40 Work 4: F. Bassi
Students' satisfaction in higher education: the role of practices, needs and beliefs of teachers

16.00 Work 5: M. Alfò, M.F. Marino
Dependence and sensitivity in regression models for longitudinal responses subject to dropout

16.20 General Discussion
16.50 Closing



Department of Statistical Sciences, Via Cesare Battisti, 241 - Padua

How to reach the department

From the train station walk all the way of Corso del Popolo and continue on Corso Garibaldi.
After the gardens (Giardini dell'Arena) turn left into via Eremitani
Continue straight into via Zabarella and turn left into via Cesare Battisti
You'll find the campus towards the end of the street, on your left.

By metrobus or bus  
From the train station, catch the metrobus line 1 or a city bus going downtown
(there are many options, watch out for n. 3, 8, 12 or 18) and get off at the "Ponti Romani".
The bus stop is really close to via Cesare Battisti.
You must enter the street and walk down for a few minutes.
You'll find the campus towards the end of the street, on your left.


This one-day conference represents the final meeting of the project SID2016 “Advances in Multilevel and Longitudinal Modelling”, funded by the University of Padua.

Starting in January 1st, 2017 the aim of this project has been to enrich the literature on multilevel modelling along two directions.

The first is methodological and investigates the problem of missing data in a multilevel setting. In the project we compare and extend some methods proposed in the literature to perform Multiple Imputation (MI), with a particular focus on the imputation of level-2 covariates (which is a research topic that has received attention recently in the literature). The issue of model specification in complex multilevel settings is strictly connected to MI, in particular the application of regularization techniques, such as the lasso, in order to select relevant predictors in a wide set.

The second research strand involves empirical applications of advanced multilevel solutions, in particular for the analysis on the students' perception and evaluation of teaching. The University of Padua has collected information on students' satisfaction of teaching for several years. In particular, it made available students' evaluations of a large sample of didactic activities, as well as administrative data related to the educational offer, over three academic years from 2012 to 2014. Moreover, a third innovative dataset may be now linked to these data: it collects information on practices and needs of teachers of the Padua University in the academic year 2012/13. This data collection is part of a project (PRODID – Preparazione alla Professionalità Docente e Innovazione Didattica) to improve the academic educational innovation and the quality of academic teaching.

The project is composed by three units:
1) University of PaduaDepartment of Statistical Sciences: Omar Paccagnella (PI), Francesca Bassi, Mauro Bernardi, Anna Giraldo
2) University of Florence Department of Statistics, Computer Science, Applications: Leonardo Grilli, Maria Francesca Marino, Carla Rampichini
3) ISTAT – Italian National Statistical Institute: Roberta Varriale

This conference is hosted by the Department of Statistical Sciences of the University of Padua, which is located in the historical building of the "Monastero di Santa Caterina d’Alessandria", a campus consisting of lecture rooms, computer labs and a library all situated around a charming garden.

This event is totally free but you need to register using the form below.



Registration Closed


Contacts and organization

For info:

Scientific Committee:
Leonardo Grilli
Omar Paccagnella
Carla Rampichini