STATISTICAL MODELS FOR MANAGEMENT STUDIES

The course is one of the first term activities of the Ph.D. teaching program in Management that allows students to acquire knowledge and understanding of some of the main statistical concepts and their use in business management activities. The course aims at providing an introduction to statistical predictive modelling and its application to Management studies. The students will learn business and predictive analytics methods for visualizing, mining, and interpreting information and special emphasis will be given to critical interpretation of results. An integral part of the course is learning the basis of the Stata software for statistical computing.

At the end of the course, students will be expected to have acquired the skills to develop a critical, personal and rigorous analysis of business phenomena and their consequence for management studies through statistical methods suitable for the analysis. They must also be able to present and discuss the results and the management strategies derived from the developed models.
In particular, students should:
1. Knowledge and understanding
- know the terminology and basic concepts of probability and statistical inference
- understand the strengths and limitations of the statistical approaches used to analyze real phenomena.
- know the standard statistical models and some advanced methods for the analysis and the prediction and their application to Management studies.
2. Ability to apply knowledge and understanding
- understand the main aspects of the statistical analyses;
- know how to determine the best statistical models for analysis and prediction
- know how to present strategies for management studies based on the achieved results.
3. Making judgements
- be able to critically assess under which circumstances the analyses are reliable
- be able to assess the goodness of the estimated models
4. Communication
- know how to present, discuss and prove the information achieved by the analyses
- know how to write a report of the results in form of scientific paper
- know how to argue management decisions in an effective way.

Working knowledge of basic probability and statistics. A gentle introduction to statistics can be found in Ross. S.M.(2010) Introductory Statistics. 3d edition, Elsevier. Understanding of the topics covered in Chapters 1-9 of the above book will be assumed though the course. Students may consider alternative textbooks that cover the same topics.

1. Introduction to data science
2. Introduction to Stata
3. Review of elementary descriptive statistics with applications in Stata
3.1 Mean, mode, percentiles
3.2 Variance, asymmetry and kurtosis
4. Univariate regression models
4.1 Model specification
4.2 Why do we need models, with examples in business
4.3 Fitting and interpreting regression model, LS idea, interpretation of coefficients
4.4 Evaluation of a regression model, R squared
4.5 Signal to noise ratio, t-stat, p-value
5. Multivariate regresion with examples in Stata
5.1 Multivariate regression model
5.2 Correlation vs. causality
5.3 F-stat
6. Dummy variables and interaction terms
6.1 Interpretation of a dummy variable coefficient, with examples in Stata
6.2 Interaction term of dummy with a quantitative variable, with examples in Stata
6.3 Data transformation
6.4 Applications in Stata
7. Making the model more selective
7.1 Multicollinearity, with examples in Stata
7.2 Variable selection, with examples in Stata
8. Nonlinear modelling
8.1 Discrete-choice
8.2 Logit and probit
8.3 Multinomial logit ordered and non-ordered
9. Applications in Stata of discrete-choice models
10. Time series modelling
10.1 Peculiarity of time series data
10.2 The notion of stationarity / spurious regressions
10.3 Decomposition in trend, cycle, irregular and seasonal component
10.4 Regression with time series data
11. Fine-tuning your model
11.1 Predictive power vs. model fit
11.2 Measuring the predictive power of a model, with examples in Stata
12. Panel data modelling
12.1 Fixed-effect vs. random effects
12.1 Applications in Stata
13. Endogeneity
13.1 Testing endogeneity, with examples in Stata
13.2 Reverse causality
13.3 Instrumental variable method, with examples in Stata
14. Intro to machine learning
15. Machine learning in action: regression trees and gradient boosting with examples

To support the theoretical knowledges acquired during the course, each topic will be developed by using the Stata statistical software. In particular, Stata will be introduced and the approaches and models used in the analyses will be developed used particular packages provided in Stata.

Notes, slides, data and other material necessary to follow lectures and to attain the intended learning outcomes are downloadable from the e-learning platform moodle.unive.it.

Mandatory texts:
1. Course Slides.
2. Jank, W. (2011). Business Analytics for Managers. Springer.
3. Baum, C.F. (2006). An Introduction to modern econometics using Stata. Stata Press.

Additional readings:
Other reading material will be suggested by the teacher during the course

The exam consists of writing a report in form scientific paper about the results of the statistical analysis of a dataset assigned to each student by the lecturer and of its critical oral presentation.
The evaluation will be based to the understanding of the material, on the material prepared by the student and on the efficacy of the presentation.
Homeworks will be assigned to practice the understanding and the skills achieved during the course.

Theoretical lectures complemented by lab classes. Methods will be discussed and illustrated through applications to real data making use of dedicated software. Teaching material prepared by the lecturer will be distributed during the course. The statistical software used in the course is Stata.

oral