UNPROVEN INNOVATION: IPADS IN K-12 ENVIRONMENTS

Nationally, the education sector spends more than 5 billion dollars annually on digital tools, “yet seldom are technology solutions factored into any viable equation for improving student academic achievement” (Moersch, 2014, p. ix). Consider the following case in point: In July 2014, Apple announced that in just 3 years, the company had sold more than13 million iPads to educational institutions worldwide (Cavanagh, 2014). Put into perspective, that represents more than 5.2 billion dollars spent by the education industry to purchase iPads, which is the equivalent of the annual salaries of 89,655 teachers (“High School Teacher: Salary,” 2014). Despite such vast expenditures, there have been very few attempts to evaluate the efficacy of these digital tools on improving academic achievement. This research involved a quantitative data review of participant (student and teacher) survey data to explore one of the country’s largest K-12 iPad implementation undertakings in an effort to identify (a) best practices and (b) lessons learned from implementing the iPad into K-12 educational environments. It should be noted that the school system forming the basis of this research already had administered and collated the surveys used in this study.

Essays in Market Design

In this dissertation, I study three problems in market design: the allocation of resources to schools using deferred acceptance algorithms, the demand reduction of employees on centralized labor markets, and the alleviation of traffic congestion. I show how institutional and behavioral considerations specific to each problem can alleviate several practical limitations faced by current solutions. For the case of traffic congestion, I show experimentally that the proposed solution is effective. In Chapter 1, I investigate how school districts could assign resources to schools when it is desirable to provide stable assignments. An assignment is stable if there is no student currently assigned to a school that would prefer to be assigned to a different school that would admit him if it had the resources. Current assignment algorithms assume resources are fixed. I show how simple modifications to these algorithms produce stable allocations of resources and students to schools. In Chapter 2, I show how the negotiation of salaries within centralized labor markets using deferred acceptance algorithms eliminates the incentives of the hiring firms to strategically reduce their demand. It is well-known that it is impossible to eliminate these incentives for the hiring firms in markets without negotiation of salaries. Chapter 3 investigates how to achieve an efficient distribution of traffic congestion on a road network. Traffic congestion is the product of an externality: drivers do not consider the cost they impose on other drivers by entering a road. In theory, Pigouvian prices would solve the problem. In practice, however, these prices face two important limitations: i) the information required to calculate these prices is unavailable to policy makers and ii) these prices would effectively be new taxes that would transfer resources from the public to the government. I show how to construct congestion prices that retrieve the required information from the drivers and do not transfer resources to the government. I circumvent the limitations of Pigouvian prices by assuming that individuals make some mistakes when selecting routes and have a tendency towards truth-telling. Both assumptions are very robust observations in experimental economics.