The Role of Ordinal Data in Economics Open Access
Downloadable ContentDownload PDF
This thesis is a contribution to the understanding of the use and role of ordinal information in economics. It raises questions about, and in certain instances provides assurance for, current ordinal data practices in the measurement and evaluation of impact. The goals of this thesis are to i) formally explore and organize certain robustness issues that are inherent in analysis with ordinal data ii) introduce methodologies that address these issues and can be applied to specific ordinal data environments and iii) present related applications.I begin with an exploration of the use of ordinal data in poverty measurement and develop ordinal counterparts to existing cardinal poverty measures. This class of measures is then extended to the multidimensional case and is adjusted in order to be used for analysis that goes beyond poverty measurement, in particular for the context of randomized evaluation of impact on multidimensional ordinal variables. These multidimensional measures are then demonstrated with a randomized controlled trial in Nicaragua where impact on ordinal psychosocial variables, pertaining to aspirations, is evaluated. Finally, the dissertation ends with a pilot project that studies aspirations in Washington DC’s underserved youth and is designed in such a way that it takes advantage of the additional insights that my proposed measures offer.