- Databases for bibliographic research (9 hours; first-year)
The course covers the processes and methods of publishing research tools for analysis of academic literature through databases.
The first part of the course focuses on the importance of academic publications for the advancement of international knowledge and for an academic career. It examines the different types of publication, the structure of an academic paper, choosing a journal and the review and publication processes. Students are invited to share personal experience.
The second part of the course focuses on the literature review: the choice of keywords, the use of the databases available through the University of Parma (e.g., Ebsco Host, Emerald, Essper) as well as other online resources (e.g. Researchgate, Academia.edu, Scienceopen, Iris, sole24ore), and reference management software (Zotero and Mendeley). The course includes the execution of practical exercises by the students during the lectures.
- Introduction to Phyton and Latex (24 hours; first-year)
The course aims at providing doctoral students with the fundamentals of programming in Python and at studying in deep some standard tools for data analysis. The specific applications vary based on the research needs and interests of the participants.
The sessions are interactive and cover the following topics:
- : introduction to syntax and fundamental data structures; basics of object-oriented programming; control flow, modules, input/output; introduction to numpy and pandas for data manipulation; graphical representations.
- : introduction to writing and compiling; insertion of floats (tables and figures); use of cross-references for formulas; management of bibliographic citations and automatic generation of the bibliography.
- Statistical methods for the analysis of economic and business data (36 hours; first-year)
The course consolidates knowledge of advanced statistical methodologies relevant to economics and business sciences and typically applicable to big data. The course consists of two modules.
The first module focuses on statistical models and algorithms for prediction and segmentation. It analyses the following methodologies: association measures and rules, and their applications to Market Basket Analysis; linear regression for classical inferential aspects and diagnosis of violations of basic assumptions; logistic regression for prediction of individual behavior; classification trees, including overfitting and misclassification rates; Cluster Analysis and the non-hierarchical K-means segmentation algorithm. All methodologies are applied to economics or business examples using SPSS.
The second module focuses on theory and algorithms describing phenomena characterized by dynamics evolving over time. There are two aspects: the first extends multiple regression models to cases of seasonal data using ARIMA-type models. Basic assumptions are introduced, and inferential results and post-adaptation diagnostics are presented. The goal is to build models that can be used to provide reliable forecasts in the short and medium term. The regression model is subsequently extended to panel data. Practical exercises will verify student learning in the methodological sessions. There are two assignments at the end of the course.
- Robust mixture modelling and robust model based clustering with applications to artificial intelligence (24 years; first-year; imparted in English - Prof. Agustín Mayo-Iscar, University of Valladolid – Spain, Visiting Professor in University of Parma, http://www.eio.uva.es/mayo-iscar-agustin/)
In recent years, we have witnessed a proliferation of continuously expanding data. An effective synthesis capable of revealing the complex structure of interconnected relationships has become a necessity in all branches of economic and financial sciences. Artificial intelligence algorithms based on statistical methods offer extraordinary learning opportunities. Often, the synthesis of relationships is identified through the creation of homogeneous groups of statistical observations. Mixture models are a rather recent development in the statistical literature in this direction. When adapting a mixture model to economic and financial data, it is also important to consider that any anomalies can alter and undermine the entire procedure. Therefore, robust calibration of the algorithms should always be considered.
Participants in this course will learn, through computational implementation, to optimize all phases of constructing mixture models and verifying their effectiveness, both using the theoretical basis and writing code in dedicated languages (primarily R). The course content includes:
- Mixture of multivariate normal distributions and parsimonious collections of mixture models.
- Extensions related to mixtures of more flexible models (skew-normal and skew-t distributions).
- Dimensionality reduction (mixture of factor analysis and mixtures of orthogonal subspaces).by
- Mixture of linear regressions and generalized linear models with robust estimation.
- Multivariate Gaussian distribution.
- Mixture of random variables.
- Robust models.
- Likelihood for robustified mixture models.
- Unsupervised clustering.
- Expectation Maximization (EM) algorithm (convergence, diagnostics, variable selection).
- Applications to economic and financial data.
- Computational implementation of the algorithms.
- Meta-analysis for management and social sciences research (24 hours; first-year; imparted in English dal Prof. Tammo H. A. Bijmolt, Università di Groningen - Olanda, Visiting Professor at0 University of Parma, https://www.rug.nl/staff/t.h.a.bijmolt/?lang=en)
The course covers the methods for conducting a meta-analysis. The aim is to train participants to carry out and publish a high-quality scientific meta-analysis in the broad field of managerial research and social sciences. This research technique allows the summarization of existing empirical findings on relationships (effects of certain x on a specific y) studied multiple times by different researchers and it enables the obtaining of relevant empirical generalizations. Conducting a meta-analysis can, therefore, be an excellent research project for a doctoral student who needs to analyse the literature on a particular topic.
The course covers the entire process of meta-analysis research: problem formulation, literature search, effect coding, analysis, discussion, and publication of results. In addition to statistical methodology, emphasis is placed on the knowledge and skills required to conduct a meta-analysis. All stages of the meta-analysis process (including statistical analyses) will be demonstrated and implemented with assignments during the lessons. Furthermore, topics will be illustrated through concrete examples of meta-analyses. Participants will expand their knowledge of relevant literature (textbooks and journal articles) and the software that supports meta-analysis projects. Specifically, most of the analyses will be demonstrated using R, and the Metafor package.
The material of the course consists of lecture notes and slides provided by the instructor. The following textbook on meta-analysis is highly recommended: Borenstein, Hedges, Higgins, and Rothstein (2021), Introduction to Meta-analysis, second edition; Wiley.
- Qualitative research methodology (21 hours; first-year)
The module presents research design and offers in depth focus on qualitative research methodologies. Therefore, the following contents will be covered:
- Definition of the research topic;
- Identifying research questions;
- Literature analysis and the systematic literature review;
- Relationship between theory and research – basics of epistemology and ontology;
- Qualitative - rational research methods, strengths and weaknesses;
- Approaches to qualitative research - ethnography, grounded research, action research, narrative research, case studies;
- Qualitative case studies - research design and data collection methods;
- Action research - research design and management methods;
- Data collection methods - focus groups and qualitative interviews;
- Data analysis methods - text analysis and content analysis.
The module is carried out with an interactive approach with the learners and requires a final evaluation of the acquired knowledge based on the presentation of a qualitative research design related to a research topic of interest to the doctoral student.
- Quantitative research methodology (21 hours; first-year)
The course extends student knowledge on the collection and analysis of primary and secondary data through different quantitative research techniques.
The first part focuses on the use of quantitative research techniques for the collection of primary data. Two research techniques are studied. The structured questionnaire and the main methods of data analysis, with particular regard to validity and reliability measurement models are covered. The next focus is on the technique of experiments and quasi-experiments and construction process of the experimental protocol, the types of data collected and the basic analysis methodologies.
The second part delivers the theoretical knowledge and application tools to carry out quantitative systematic reviews of academic literature. The course focuses first on the use of SciMAT open-source software for longitudinal bibliometric analyses for mapping the cognitive structure and thematic evolution over time of a line of research. The course also offers an overview of the theoretical and practical tools for systematically reviewing literature by means of meta-analyses and for critical reading meta-analyses.
During the lectures published academic papers will be described and discussed, and data analysis software (mainly SPSS) will be used.
At the end of the course, students have about two months to design their research project on a topic of interest and to draft research questions and hypotheses, the methodology and the expected results. The projects are assessed individually, and strengths and weaknesses of the research designs are shared.
- Data management with STATA (21 ore; primo anno)
The objective of the course is to introduce doctoral students to STATA, providing them with the basics related to the fundamental syntactic structure and the main commands for data management, analysis, and graphical representation.
The module covers various topics such as string functions, commands for reshaping, merging, and combining data sets, foreach and forevalues commands for repetitive tasks, and the collapse command for data set aggregation. Examples are used throughout the module, with particular emphasis on patent data, but not exclusively.
The lessons take place in a computer lab to allow doctoral students to apply what they have learned during the course. All teaching materials, including datasets and script files, are provided to the students.