P1: Measurement and methods
(Globalization, institutions, power: Measurement and methodological issues)

This project intends to generate measures of globalization, gender and institutions, and world power in the specific context of some of the other projects in this RTG. While past research provides first basic concepts for the measurement of world power the measurement of world power as one aspect of globalization has been neglected in the last decades and is methodologically underdeveloped. Research in the RTG in this methodologically-focused project investigated statistical issues in the creation of composite indices, which are ubiquitous in globalization research and related macro-oriented empirical work. In particular, the use of partial least squares (PLS) as a preferred way to create weights for composite indices was investigated.

Future research in this project will focus on robust model selection to identify explanatory variables that are robustly related to the outcome. We would like to explore and compare the performance of different methods in specific social sciences applications. That will include revisiting empirical results on the impact of globalization on growth and inequality.

• Statistically robust model selection: extreme bounds analysis versus penalized regressions
• Model selection for empirical growth research