Accueil > Statistics and data analysis for policymakers


Type d'enseignement : Seminar

Semester : Autumn 2017-2018

Number of hours : 24

Language of tuition : English

Voir les plans de cours et bibliographies




CAUTRES, Bruno (Chercheur CNRS et CEVIPOF, Sciences Po)

Pedagogical format

Students are not expected to buy these books, since there will be a reading pack with a choice of chapters and texts. In case students would like to buy the books, they can buy used versions of the previous editions (for instance Agresti and Finlay third edition or Gujarati, 4th edition) that are normally much cheaper. Every session will be based on a PowerPoint presentation, made available to students through a Google workgroup. For every session, the instructor will indicate to the class the exact pages of the reading pack that must be read before the class. To fully benefit from the class, the readings must be done before coming to the class.

Course validation

The course is evaluated on the following basis: A mid term exam (30% of the grade) ; Two take-home assessment tasks : 2 assignments (2 x 10% = 20%) ; Final exam (50% of the grade). Students getting low grades in mid-term exam will be proposed remedial workshop or crash course. Remedial class can be organized before the final exam also. Attendance to class is mandatory; doing the weekly readings is a key condition for learning.


Monitoring of the course and of the students' works : A short meeting between students delegate and instructor will be organized every two/three courses, to ensure the quality of the interaction between the course and the students. Work meetings can be organized with students having difficulties with quantitative methods. Instructor offers e-office hours : every Monday, from 1:30pm to 3pm, instructor will answer by email to any question asked by student and send during this e-office hours session. Students will get immediate answers during this 90-minute electronic session. In case of need instructor can extend this electronic session by “Skype-statistics-clinic” sessions.

Required reading

  • There are several textbooks that may be of use for statistical methods but two books are excellent and instructive
  • Alan Agresti, Barbara Finlay. Statistical Methods for Social Sciences. Prentice Hall, 4th edition
  • Damodar N. Gujarati. 2003. Basic Econometrics, 5th edition. NY: McGraw-Hill/Irwin
  • Another book is most relevant : Damodar Gujarati, Econometrics by Example, Palgrave, 2011

Plans de cours et bibliographies

  • Session 1 : Introduction to the course , basic notions (or recalls) on descriptive statistics
  • Session 2: Basic notions (or recalls) on descriptive statistics and statistical association/correlation
  • Session 3 : Introduction to the linear regression model
  • Session 4 : The linear regression model : theory and basics.
  • Session 5 : OLS estimates and their properties : how the linear model functions ?
  • Session 6 : Goodness of fit and statistical inference : how to evaluate the quality of a model ?
  • Session 7 : Multiple regression analysis : the logic of controlling for and ‘ceteris paribus’, partial regression coefficients
  • Session 8 : Goodness of fit, statistical inference, models comparisons : how to choose a good model?
  • Session 9 : Residual analysis
  • Session 10 :  Dummies variables, interactions, nonlinearities (1)
  • Session 11 :  Dummies variables, interactions, nonlinearities (2)
  • Session 12 : Summary