This summer, I worked with Professor Judd Kessler and Professor Clayton Featherstone, of the Business, Economics and Public Policy department in Wharton. Our research project was based on the field of market design, a field of microeconomic theory that examines how the rules of matching markets can be chosen so as to maximize benefit to all parties. Our case study focused on Ensena Chile (eCh), a Chilean nonprofit similar to Teach for America, and how they can optimize their process of matching teachers to schools. We used this case study to further examine how assignment systems can be optimized without asking participants to reveal information that is difficult to collect - specifically, cardinal preference information, which is an absolute measure of how much someone prefers an outcome.
Normally, eCh only asks teachers for ordinal information - their ranking of the available options - which is much easier to collect than cardinal information. We collected cardinal information from the teachers using a unique survey design, and then used that cardinal information to test different matching algorithms to see which of them maximized overall utility.
I was fortunate because my work on the project was very substantive in nature - it required a real understanding of the data we were analyzing and the matching algorithms we were implementing. I had to implement each of the matching algorithms in Java, using linear programming concepts. I also used Stata, a statistical programming language, to analyze the preference data to see how the other matching algorithms compared to eCh’s matching process, and how much utility they provided.
Working with Professors Kessler and Featherstone has helped me explore my interest in economic theory - particularly in the growing field of market design - through a specific application of the theory that I’ve learnt through the project.
This summer, I worked with Professor Judd Kessler and Professor Clayton Featherstone, of the Business, Economics and Public Policy department in Wharton. Our research project was based on the field of market design, a field of microeconomic theory that examines how the rules of matching markets can be chosen so as to maximize benefit to all parties. Our case study focused on Ensena Chile (eCh), a Chilean nonprofit similar to Teach for America, and how they can optimize their process of matching teachers to schools. We used this case study to further examine how assignment systems can be optimized without asking participants to reveal information that is difficult to collect - specifically, cardinal preference information, which is an absolute measure of how much someone prefers an outcome.
Normally, eCh only asks teachers for ordinal information - their ranking of the available options - which is much easier to collect than cardinal information. We collected cardinal information from the teachers using a unique survey design, and then used that cardinal information to test different matching algorithms to see which of them maximized overall utility.
I was fortunate because my work on the project was very substantive in nature - it required a real understanding of the data we were analyzing and the matching algorithms we were implementing. I had to implement each of the matching algorithms in Java, using linear programming concepts. I also used Stata, a statistical programming language, to analyze the preference data to see how the other matching algorithms compared to eCh’s matching process, and how much utility they provided.
Working with Professors Kessler and Featherstone has helped me explore my interest in economic theory - particularly in the growing field of market design - through a specific application of the theory that I’ve learnt through the project.