Despite expectations in the late 1990s that the Internet would lead to frictionless commerce, empirical research shows that electronic commerce is characterized by persistent price dispersion. This paper examines reasons why online price dispersion exists in the books, CDs, and consumer electronics retail sectors. The Internet allows for multi-channel retailing, and it influences supply, demand, and market-related factors for both retailers and consumers. These factors shape retailers’ price-setting abilities which, in turn, shape their pricing strategies. I find that different pricing strategies result in online price dispersion, and I ultimately predict that retail e-commerce markets, which are oligopolistic in nature, will continue to exhibit persistent price dispersion. In particular, price dispersion will decrease the most in the CDs sector, followed by the books and then the consumer electronics sectors.

The IT age is marked by innovative approaches to the online commerce. Technology as the core of innovation has undergone numerous evolutions through the “creative destruction." Motivated by the phenomena and the challenges in the technology-driven markets, I explore the economic role of innovation from different angles in the following essays. Chapter 1 focuses on firms’ competitive strategies while constructing novel business models in delivering online services. In particular, I am interested in their bundling of marketing services with the core business. In a game theoretic model, I derive competing firms' equilibrium strategies with choices between three business models, no ad-support, ad-support with the optional advertising strategy, and the mandatory advertising strategy, and find that competitive business models can be differentiation-driven or advertising driven depending on market ad aversion. Interestingly, mandatory advertising weakly dominates optional advertising under certain market conditions. My findings offer new insights to the bundling literature. Chapter 2 examines the performance-based auction model in the iconic online advertising innovation, keyword auctions. I analyze advertisers' decision of utilizing their existing reputation from a primary auction upon entering a new auction. The short-term and long-term setups are modeled for analyzing seasonal marketing in a new auction and branding a new product, in examining the impact of new market size, performance, and risk on advertisers' decisions. While an optimistic new market encourages reputation stretching, in the long-term setup it further depends on the performance difference between the two markets. A higher risk is found to induce stretching under intensive competition for both cases; in the long-term, stretching decision is determined by the market size. Chapter 3 examines the connection between business cycles and innovation and offers insights for regulatory innovation policies. Combining endogenous market structure with the dynamic game framework, I study the Markov perfect equilibrium where heterogeneous firms choose their innovation rates. I find that increased per-capita income tends to improve aggregate innovation, while income inequality shocks may reduce innovation conditional on the market structure. I also find subsidies to dampen innovation incentives, and policies such as tax credits that reduce the variable R&D costs to have positive effects on innovation. / text

Traffic information, now available through a number of different sources, is re-shaping the way planners, operators and users think about the transportation network. It provides a powerful tool to mitigate the negative impacts of uncertainty, and an invaluable resource to manage and operate the network in real-time. More information also invites to think about traditional transportation problems from a different perspective, searching for a better utilization of the improved knowledge of the network state. This dissertation is concerned with modeling and evaluating the system-level impacts of providing information to network users, assuming that the data is utilized to guide an Adaptive System-Optimum (ASO) routing behavior. Within this context, it studies the optimal deployment of sensors for the support of ASO strategies, and it introduces a novel SO assignment approach, the Information-Based System Optimum (IBSO) assignment paradigm. The proposed sensor deployment model explicitly captures the impact of sensors' location on the expected cost of ASO assignment strategies. Under such strategies, a-priori routing decisions may be adjusted based on real-time information. The IBSO assignment paradigm leads to optimal flow patterns which take into account the ability of vehicles to collect information as they travel. The approach regards a subset of the system's assets as probes, which may face higher expected costs than regular vehicles in the search for information. The collected data is utilized to adjust routing decisions in real time, improving the expected system performance. The proposed problem captures the system-level impact of adaptive route choices on stochastic networks. The models developed in this work are rigorously formulated, and their properties analyzed to support the generation of specialized solution methodologies based on state-space partitioning and Tabu Search principles. Solution techniques are tested under a variety of scenarios, and implemented to the solution of several case studies. The magnitude and nature of the information impacts observed in this study illustrate problem characteristics with important theoretical, methodological and practical implications. The findings presented in this dissertation allow envisioning a number of practical applications which may promote a more efficient utilization of novel sensing and communication technologies, allowing the full realization of their potential. / text

In my dissertation, I argue that rationality, for real humans, is best understood as a strategy for communication and interacting in a social environment. In particular, I argue that humans are rational to the extent that they are able to understand and be understood by others, to the extent that they can give and accept reasons and explanations. This raises a pair of questions. The first concerns the source of the norms for giving and accepting reasons. The second is why we should accept and follow these norms if they are not guaranteed to preserve truth or optimize outcomes. I address the first question by arguing that these norms function as constraints on our imaginations, on the ways in which we can think about or understand the world. This goes beyond the traditional view that these norms govern acceptable inferences. Rather, I argue, the more significant function of these norms is to govern the structure of our reasoning in the sense of guiding considerations about the relevance and form of our understandings of situations. This suggests an answer to the second question. We ought to accept these norms because they are self-confirming. Following them allows us to communicate and interact with others who follow these same norms. In those endeavors that require interaction and coordination in a social group, being understood is frequently more important than being right. / text

Using a multiple gating procedure, five elementary principals were identified as special education leaders. This descriptive study uses a multiple-case study design to explore principal’s (a) understanding and perceptions of instructional strategies associated with improved outcomes for students with learning disabilities (LD), and (b) their instructional leadership practices utilized to promote educators’ instruction of students with LD. Findings reveal that the majority of participants (n = 4) expressed at least a moderate understanding of effective instruction for students with LD, perceived implementation of these practices as necessary (n = 5), but reported mixed perceptions regarding feasibility. As expected, the five participants also described different levels of utilizing instructional leadership practices to promote educator’s instruction of students with LD. However, the two participants with advanced special education degrees were distinct from the others with respect to their combined high understanding, positive perceptions, and instructional leadership practices utilized to promote educators’ instruction of students with LD. Overall propositions indicate: (a) principals in this study who possessed higher understanding of effective instructional practices for students with LD and were interested in improving this understanding, were more apt to engage in instructional leadership practices to promote effective instruction for students with LD, (b) principal’s prioritization of developing a collaborative vision and practices among educators to promote effective instruction of students with LD may be beneficial to improving instruction for students with LD, and (c) principal’s intentional interaction and support with both general and special educators may lead to higher levels of collaboration among educators as well as more effective instruction for students with LD. / text

Bounded Verification is an effective technique for finding subtle bugs in object-oriented programs. Given a program, its correctness specification and bounds on the input domain size, scope bounded checking translates bounded code segments into formulas in boolean logic and uses off the shelf satisfiability solvers to search for correctness violations. However, scalability is a key issue of the technique, since for non-trivial programs, the formulas are often complex and can choke the solvers. This thesis describes approaches which aim to scale scope bounded checking by utilizing syntactic and semantic information from the code to split a program into sub-programs which can be checked incrementally. It presents a thorough evaluation of the approaches and compares their performance with existing bounded verification techniques. Novel ideas for future work, specifically a specification slicing driven splitting approach, are proposed to further improve the scalability of bounded verification. / text

我們在此文中檢視了股市流動性的動能效果，並將此效果連結到相對應股票的報酬表現上。我們發現過去六個月平均流動性較高的股票，在未來三年中也會具有較高的流動性。此外，我們發現買入較高流動性的股票，賣出流動性較低的股票，會有正的報酬。我們希望此研究能夠幫助投資人獲取更多有用的資訊。 / We examine the predictability of liquidity, the momentum effect in liquidity, and we also would like to link this effect to expected stock returns. We find that stocks with high liquidity in the past six month will be traded with high liquidity in the future (within 3 years) and that all of the zero-cost portfolios, which buy high liquidity stocks and sell low liquidity stocks, have positive returns. We hope the results in this study will help uninformed trader to obtain more information in the stock market.

The need to give educators the ability to enter a particular teaching strategy of their choice into an Intelligent Tutoring System (ITS) and have the ITS respond appropriately to each student has been stated by many researchers. For example, an educator could tell the ITS to keep students within a certain help level ratio (how much help they request), or to introduce a new topic in a particular manner and the ITS simply carries this out at each learning point of interest. Educators could then try new strategies, ones that unaided are impossible to try out in class (such as keeping a student within a help-seeking range) or difficult within an ITS (as the ITS would have to be specially programmed in that way). Current ITSs provide adaptivity to the student at the domain level but not necessarily at the pedagogical level. While a variety of pedagogical strategies have been implemented (e.g. apprenticeship, socratic, practice), there is no system that offers parts or all of these strategies with the ability to choose between them dynamically. In this project, we designed a new framework for an ITS to include multiple, potentially adaptable pedagogical strategies. This was done by breaking up the pedagogical module into separate components. The Pedagogical Strategy Set (PSS) contains all the strategies, written as constraints. The Pedagogical Student Model (PSM) keeps track of which pedagogical strategies were used by each student. Within the ITS, there is still a smaller, separate pedagogical module to deal with domain-specific strategies. The Pedagogical Control Centre (PCC) contains the logic of when and how to use the pedagogical strategies. It gathers its information from the other modules and uses decision logic to trigger strategies. We implemented and evaluated this framework within the context of SQL-Tutor and found that the framework could be used to enter pedagogical strategies, which in turn compared favourably to the original SQL-Tutor. This proof of concept opens up the possibility of the logic and algorithms that could be implemented (e.g. in the PCC) in future ITSs. The PSS is a separate module, written in a different language, independent of ITSs. This could lead to sharing of pedagogical strategies between tutors. Furthermore, students learn differently to each other; this framework allows them to do so.

Collaborative learning activities are those that involve students in jointly constructing meaning and solving academic tasks (Damon & Phelps, 1989). Collaborative approaches to learning are particularly appropriate for learning disabled (LD) students who tend not to actively engage in learning activities (Torgesen, 1978; Wong, 1980). Activities of the interactive semantic mapping (ISM) strategy (Bos & Anders, 1989; Scanlon, Gallego, Duran, & Reyes, in press) provide students with opportunities to engage in collaborative learning. The purpose of this study was to determine how adding collaborative learning skills instruction to ISM strategy instruction would affect LD adolescents' comprehension of texts, collaborative skills performance, and participation in the ISM strategy and collaborative group activities. Subjects for the study were 32 LD adolescents with fourth grade or higher reading skills and IQs in the average range. Subjects were assigned to one of two treatment conditions and groups (n's = 4) within conditions. In one condition, subjects only learned the ISM strategy; in the other, subjects were instructed in both the ISM strategy and effective collaborative learning skills. Comprehension, collaborative skills performance, ISM skills performance and group interaction patterns were assessed at baseline, post test, and long term application, as well as at domain generalization and situation generalization. Findings indicate that virtually no significant differences occurred between conditions for comprehension or performance of collaborative or ISM skills. Significant differences were found for time. Comprehension, collaborative learning and ISM skills increased significantly from baseline to posttest. Students in both conditions were better able to generalize their skills at domain generalization than at situation generalization. Comprehension and skills use generally decreased at long term application. Interaction process analyses indicated that groups taught collaborative learning skills interacted in a marginally more collaborative manner than did ISM condition groups.

This dissertation contains three essays describing experiments in game theory and economics. Chapter one studies mixed strategies by asking whether game players are willing to pay for randomization. A natural intuition for mixed strategies is randomization for unpredictability, but this is theoretically fragile. A player should only randomize between strategies if indifferent, and then could choose a disequilibrium strategy. Various theories describe mixed strategies not as random play, but heterogeneous pure-strategy play. I conduct experiments in which players can choose to pay a fee to use a randomization device, applied to O’Neill’s zero-sum game. If subjects did so, it would show a strict preference for randomization over any available pure strategy. In fact, very few chose to use the randomization device. Subjects’ descriptions of their decision process were consistent with the notion of purification. Chapter two also studies mixed strategies, asking whether randomization is a property of individual choice or game play. In two experiments, game players are mirrored by guessers who make predictions about game play, distinguishing best-responding from game playing. In a Matching Pennies game, I find that game players are they are both more interested in unpredictability and actually more random. In an Asymmetric Matching Pennies game, I look at whether players are willing to forgo expected payoff in order to be unpredictable, and find little difference between players and guessers, with players being somewhat better at exploiting disequilibrium play. Chapter three experimentally implements markets for competing goods with network effects. Markets with strong network effects often have multiple equilibria, including winner-take-all equilibria in which one firm has a monopoly. Firms may compete dynamically with the aim of locking-in to a favorable equilibrium. In this paper we create an experimental market with differentiated products and network effects. When lock-in is created by simulating naïve buyers, monopoly does arise with sellers setting high prices. With human buyers, markets without switching costs are extremely competitive, with no support for stories of lock-in and monopoly. Markets with switching costs are inefficient, but this is overwhelmingly due to the individual switching costs rather than monopoly.