Quantitative Methods for Causal Analysis

DPU, Aarhus University
Dates: 7-9 February 2011
Deadline for Application:  21 January 2011
DPU, Aarhus University, 164 Tuborgvej, 2400 Copenhagen NV, Room D 168a
ECTS-points: 3
Course responsible: Anders Holm, aholm@dpu.dk

Course description:
This course is aimed at PhD students and young researchers in the educational and social sciences who seek to estimate causal effects from on experimental and non-experimental data. The course begins with a review of standard linear regression analysis and then explains the shortcomings of this approach in the context of non-experimental data. Second, the course addresses the benefits of experimental data in terms of improving causal inference. Third, the course introduces several recent methods developed by statisticians and econometricians designed to improve causal statements from non-experimental and observational data (the course focuses principally on difference-in-difference and matching methods). The course combines lectures and pratical exercises and focuses on intuitive explanations of how the methods work and how they are applied in practice.
Requirements: Basic understanding of quantitative methods and of OLS regression.

Location:
Danish School of Education, Aarhus University, Tuborgvej 164, 2400 Copenhagen NV.  Room d168A (computer lab), Building D. See map

How to get there:
The School of Education is conveniently located next to Emdrup S-train station (F line). Trains run to and from Nørreport station and Copenhagen Central Station ca. every 10 minutes. Building D where the course takes place is the first building you see when you leave the S train station and walk toward the School of Education.

Tentative course schedule
Monday February 7:
10:00-12:00 Review of statistical control and regression analysis
12:00-13:00 Lunch break
13:00-14:15 Randomization and selection bias
14:15-14:30 Coffee break
14:30-16:00 Exercise: Use Stata to estimate causal effects in regression analysis

Tuesday February 8:
9:00-10:45 Matching analysis.
10:45-11:00 Coffee break
11:00-12:00 Statistical control through matching
12:00-13:00 Lunch break
13:00-14:45 Exercise: Use Stata to estimate the causal effect from statistical matching
14:45-15:00 Coffee break
15:00-16:00 Double differences

Wednesday February 9:
10:00-12:00 Instrumental variables and natural experiments
12:00-13:00 Lunch break
13:00-15:00 Exercise: Using Stata to estimate causal effects from natural experiments
 
Reading list:
Textbooks:
Khandker, Koolwal and Samad (2009) "handbook on Impact  evaluation – quantitative methods and practices", The world bank.+
Angrist and PIschke (2009) "Almost harmless econometrics – an empiricist's companion", Princeton University Press.*
Schneider et al. (2007) "Estimating causal effects using experimental and observational designs", American educational research association.
Morgan, Stephen L. and Christopher Winship. 2007. "Counterfactuals and Causal Inference: Methods and Principles for Social Research". Cambridge University Press.
Review of linear regression analysis and statistical control:
Stolzenberg (2004). "Multiple regression analysis", in Hardy and Bryman (eds.) "handbook of data analysis", Sage.+
Freedmand, D. (1999) "From association to causation: some remarks on the history of statistics", Statistical science, 14.
Dougherty, C. (2007) Introduction to econometrics, thirds ed. (Chapter 4, Multiple regression analysis). Oxford University press.+
Lieberson, S. (1995) "Making it count" (chapter 6: Control variables). University of California press. +
Randomization and selection bias:
Chapter 2 and 3 in Khandler et al.+
Dougherty, C. (2007) Introduction to econometrics, thirds ed. (Chapter 6, Specification of regression variables: a preliminary skirmish). Oxford University press.+*
Sobel, M. (1998) "Causal inference in statistical models of process of socioeconomic achievement – a case study". Sociological methods and research, 27.*
Moffit, R. (2004) "The role of randomized field trials in social science research", American behavioral scientists, 47.
Cook, T. D. (2004) "Randomized experiments in educational policy research: a critical examination of the reasons the educational evaluation community has offered for not doing them", Educational evaluation and policy analysis, 24.
Jones, Stephen R. G.  (1992)  "Was There a Hawthorne Effect?"  The American Journal of Sociology, Vol. 98,
Statistical control through matching and regression analysis:
Chapter 4 in Khandler et al.+
Jeffrey A. Smitha and  Petra E. Toddb (2005) "Does matching overcome LaLonde's critique of
nonexperimental estimators?", Journal of Econometrics, 125.+*

Morgan, S. (200?) Counterfactuals, causal effects heterogeneity and chtholic school effect on learning, Sociology of education, 74.+

Double differences:
Chapter 5 in Khandler et al.+
Imbens/Wooldridge, Lecture Notes 10, (2007), What's New in Econometrics? Difference-in-Differences Estimation. NBER, Summer 2007 Lecture 10.+*

Instrumental variables and natural experiments:
Chapter 6, 7 and 8 in Khandler et al.+
Ward and Johnson (2008) "Addressing counfounding errors when using non-experimental, observational data to make causal claims. Synthese, 163.+
Krueger and Angrist (2001) "Instrumental variables and the search for identification: from supply and demand to natural experiments", Journal of economic perspectives, 15.+
Angrist, J. D. (2005) "Instrumental variables methods in experimental criminological research: what, why and how". Journal of experimental criminology. 2.+
Angrist, J. D. (2004) "American educational research changes track", Oxford review of economic policy, 20.+*
Emilia Del Bono and Fernando Galindo-Rueda (2007) "The Long Term Impacts Of Compulsory Schooling: Evidence from a Natural Experiment in School Leaving Dates", Centre for the conomics of Education.

Siedler, S. (2007) "Family and Politics: Does Parental Unemployment cause Right-Wing Extemism?" Working paper, Institute for Social and Economic Research, University of Essex

+: used in course as main text
*: Technical of otherwise hard to read

Application:
Completed Application forms should be sent to: Laila Parbst, PhD Board of Studies, Danish School of Education, 164 Tuborgvej, 2400 Copenhagen NV, Denmark. E-mail: phd@dpu.dk. Use Application form from www.dpu.dk/phdcourses.