Please introduce yourself. What is your background and how did you become an expert in the topic?
I graduated in economics back in 1990. While studying economics, I discovered statistics as a set of methods that can extract useful information from data, support decision-making, and help understand the world. After my undergraduate degree, I started a PhD in applied statistics in Florence. During, and soon after my PhD, I worked as a research assistant and research associate in England. I then pursued a career in statistics in Italy. I started as an Assistant Professor in Statistics back in 1994, and since 2005 I have been Professor of Statistics at the University of Florence.
Statistics is a bit broader than econometrics, which is where I started from by using statistical and mathematical tools to address and model problems in economics. But, I agree with American mathematician and statistician John Tukey that "the fun of being a statistician is that you can work in other people's backyard." As a statistician, I could develop tools for investigating and address real empirical problems in a lot of different and exciting fields of science. I believe, in fact, that most of the great advances in statistics and econometrics stem from the need of using data to generate new knowledge.
Since I started doing applied research in economics and the social sciences, I was particularly interested in causal inference, namely finding or discovering causal relationships in either experimental, but most of the time, observational data – when the treatment or the cause is not under experimental control. This is where I have been focusing my research over the last 20 years.
I am not only doing research in causal inference. Nowadays, we are bombarded by data – data is everywhere. Big data is a powerful resource that allows researchers to generate new knowledge and discoveries for the benefit of society, but only when it is used, analysed, and interpreted thoughtfully, with honest uncertainty quantification.
Even though my background is in economics, I have also been applying statistical methods and causal inference methods in a wide range of settings and fields including medicine, genetics, political science, and environmental health problems. I can see there is a lot to gain from interactions within various fields and in doing interdisciplinary work.
Over the last couple of years at the EUI, as a Part-time Professor I have been teaching a second-year course on the econometrics of causality. At the University of Florence, I was Director of the Florence Centre for Data Science, an interdepartmental research centre aimed at gathering expertise in data science from different methodological fields – such as engineering, mathematics, statistics, economics, and econometrics – as well as supporting research in various fields, not only in the social sciences, but also in medicine, biological, and agricultural sciences.
Are you working on any research projects/publications at the moment? Are you undertaking any grants?
I have been working on designing experiments, but more specifically, developing methods to tackle problems that may arise in experiments where humans are involved. In so-called "broken randomised experiments", for example, people can be lost to follow-up, or they may not comply to the experimental protocol refusing to take the treatment, or discontinue the treatment because of adverse events. This line of research is very interesting because these events make experiments a lot closer to observational studies. You still need to find a way to quantify treatment effects that can be used to inform policy decisions, trying to exploit the benefits of initial randomisation. The analysis of broken randomised experiments is a broad and still open research topic, especially for settings where you have multiple complications of this type.
Another related line of research is on understanding mechanisms. So, when you find an effect, like the impact of a policy, you may also want to know what the causal processes driving the effect are, in order to improve the policy or find certain features that should be dropped or enhanced.
The other line of research I have been working on is methods to address causality in settings with interference and spillovers that arise when units interact within groups or on a network. For example, in vaccine studies, the efficacy of a vaccine not only depends on an individual taking the vaccine or not, but also on what other people around the individual are doing. This is known as interference: The treatment applied on a unit may affect the outcome of other units; developing causal methodology for settings with interference is a new and fertile area of research.
What will your role entail? Which areas of work will you be focusing on?
Data analysis is an essential component of research if we truly want to push the boundaries of discovery and innovation in economics, but also more broadly in social sciences and science in general. Due to the conceptual and practical challenges in inferring causality, I would like to equip researchers with the necessary tools and resources to advance their work. Economists and researchers at the EUI are sharp and creative, and being exposed to the types of problems they study is really stimulating for me and for improving my own research. I hope it is a mutual benefit.
What are your goals/aims during your time at the EUI?
I would like to enrich my research by interacting with researchers, research fellows, and faculty. The excellent academic environment of the EUI Economics Department will certainly be a stimulus to develop and apply novel methods, having a far-reaching impact beyond statistics and econometrics. I truly hope to train and inspire researchers to work on challenging problems with sound data science tools that can improve and support policy decisions.