9 February 2024
Panel Data Analysisonline course
"Learn a wide range of panel data methods estimated with Ordinary Least Squares Regression and how to apply these to substantive research questions for dependent variables that are continuous.
Need to know
This course builds on Ordinary Least Squares (OLS) Regression and extends it to data with a panel or TSCS structure. You should be familiar with basic theories of OLS, up to multiple regression.
To participate meaningfully in the lab sessions, you should also have basic knowledge of Stata and/or R.
Some background in linear algebra would be helpful but is not required.
You will need to watch nine pre-recorded video lectures before you cover the relevant materials and their application in class.
Key topics covered
The course consists of three parts:
We discuss the logic and assumptions underlying panel data methods. You’ll learn how the development of more advanced methods is driven by the need to address potential violations of these assumptions.
We focus on the various statistical approaches and tricks available to help you deal with such violations and problems hidden in your data. This will allow you to obtain estimates of effects as close as possible to the true causal effects.
We apply the wide range of panel data methods discussed in the previous parts to substantive research questions. You will learn how these methods can provide answers to your own research questions.
A short review of OLS regression, emphasising key assumptions required for the OLS estimator to be the ‘best’ estimator, followed by simple panel data methods: two-period panel data analysis and first differencing.
Focusing on more advanced methods for estimating unobserved effects in the context of panel data analysis, we introduce fixed and random effect estimators. We discuss their properties and the assumptions needed for them to be valid. With these foundations, you will then study a relatively new correlated random effects approach; a synthesis of fixed-effects and random-effects methods which has been proven very useful. You'll apply these techniques using R and Stata during the hands on-session.
A lecture on the instrumental variable (IV) method, which deals with violations to the strict exogeneity assumption, followed by a connected group activity and a Q&A with the Instructors.
We move on to more advanced panel data methods that address further violations of the standard OLS assumptions, including clustered and robust standard error, panel-corrected standard error (PCSE) estimates, and dynamic panel methods (Arellano-Bond and system GMM estimators). You will apply these techniques using R and Stata during the hands-on session.
To earn extra credits, you can present your research or research proposal that uses panel data methods, and receive feedback from the Instructor and fellow participants.
How the course will work online
The course combines pre-course assignments, such as readings and pre-recorded videos, as well as daily three-hour live lectures in Zoom, where you will interact with the Instructor and fellow participants in real time.
All readings and video materials, along with all R and Stata code and data, are uploaded to Canvas, the e-learning platform supporting course delivery.
For the R practice, we will give you RStudio Cloud accounts, and use this R environment. Access details will be provided on the learning management system."
Akos Mate is a research fellow at the Centre for Social Sciences in Hungary. His key research area is the political economy of the European Union and its members’ fiscal governance. Andrew X.Li is an assistant professor at CEU's Department of Internationa
The course is designed for a demanding audience (researchers, professional analysts, advanced students).
"This course focuses on panel data models estimated with OLS; namely, linear models for dependent variables that are continuous or can be reasonably assumed to be continuous.
Although the basic concepts and ideas also apply to models designed for binary or ordinal dependent variables such as probit or logit, these models are complicated by maximum likelihood estimation (MLE) and require a different set of foundations.
Limited course hours and the complexity of MLE means we won't cover non-continuous panel data models for dependent variables, such as fixed-effect logit and 2SLS (IV) probit models."
"You can earn up to four credits for attending this course.
3 ECTS credits – Attend 100% of live sessions and engage fully with class activities.
4 ECTS credits – Attend 100% of live sessions, engage fully with class activities and complete a post-class assignment."
GBP 492: ECPR Member - check if your institution is a member on our website.
GBP 985: ECPR Non-Member
Funding applications for the 2024 ECPR Methods School Winter instalment are now closed. For more details on funding opportunities for ECPR's other events, please visit our website.