This dissertation contributes to the ongoing debate on how educational inputs make a difference and how to allocate them efficiently. Educational inputs could be broadly defined as any personnel inputs such as teachers and career service staff, learning environment that includes peers and school facilities, and policies that facilitate learning. This dissertation explores three topics: peer effects in higher education, the consequences of college expansion, and the impacts of school closures. Chapter I estimates the peer effects of non-cognitive skills. I show how peers’ non-cognitive skills influence students' academic outcomes and own non-cognitive skills. I use a unique dataset that includes information on student non-cognitive skills, course grades, and friendship from a university in China that randomly assigns students to dormitories. My first main finding is that peers’ non-cognitive skills affect students’ academic outcomes positively but differentially. All students benefit from exposure to “persistent” peers, while students with low baseline academic ability also benefit from exposure to “motivated” peers. My second main finding is that peers also affect the development of students’ self-control and willingness to socialize. These findings have important implications in evaluating the social returns to interventions that improve non-cognitive skills and education policies that change peer group composition. Chapter II summarizes the current literature on college expansions, which change the education resource for many students. Studies have explored the impact of College Expansions that happened worldwide and this chapter summarizes literature in the field of economics of education. This chapter pays special attention to studies that explore the impact on wages and employment and how current studies identify causal relationships. Meanwhile, this chapter reviews how current studies examine the impacts of college expansion in China starting from 1999, which was unparalleled in magnitude. Finally, I discuss how future studies could improve to identify causal effects of the impact of the tremendous college expansion in China. Chapter III, a joint work with Ying Xu, estimates the effect of school closures causedby wildfires. School closures are a common and disruptive feature of education systems when sudden shocks from weather, natural disasters, or infectious disease require that students remain at home rather than in the classroom. Indeed, since January 2020, school closures have happened all around the world due to the COVID-19 pandemic. In the United States, more than 50 million students are currently out of school due to COVID-related closures. This raises an important question: How do sudden school closures affect student development in the short and medium term? In this chapter, we use administrative data to examine the causal effect of unexpected school closures, exploiting sudden variations in these closures due to wildfires in California. We show that unexpected closures have negative effects on student test scores, and the loss of school time is one of the most important mechanisms of decline in student achievement. Meanwhile, minority students and students from school districts with low socioeconomic status experience larger negative effects from such unexpected closures. We argue that these results can help inform policy to identify and address the negative impacts of such closures.

High PM2.5-related mortality and morbidity has become a public health concern in China. To date, there have been few studies investigating individual implementation and use of protective behaviors to counter outdoor PM2.5 air pollution levels in China. This study, therefore, aimed to investigate adults’ attitudes toward evidence-based protective behaviors against outdoor PM2.5 air pollution, the results of which might inform health educators and health promotion specialists about what they should emphasize when educating communities about protective behaviors in the PM2.5 air pollution days. Utilizing the purposing sampling method, a WeChat-based survey was conducted among 300 adult participants living in B residential area in Yuhuatai district of Nanjing, which is a representative city of high urbanization level and PM2.5 air pollution in China. The survey included total 16 items assessing participants’ background information, attitudes toward four evidence-based protective behaviors against PM2.5 air pollution, and knowledge regarding possibly effective protective approaches specifically related to Chinese medicine in the PM2.5 air pollution. The results showed that the importance adults in Nanjing attached to wearing N95 respirator when walking outside, putting air filter (HEPA) at home, and avoiding unprotected outdoor sports activities in the PM2.5 air pollution was greater than closing all the doors, windows, and many openings in the PM2.5 air pollution. Regarding comparison of attitudes between pre- and post-controlled COVID-19 periods, the value adults gave to closing all the doors, windows, and many openings in the PM2.5 air pollution was higher during pre-COVID-19 period, whereas the opposite of circumstance took place when referring to other three protective behaviors. Regarding comparison of attitudes between biographic variables, adults aged 18-50 (including 50) attached greater importance to all the four protective behaviors than those aged 50+. In addition, adults having the habit of checking daily AQI ranked the behaviors of avoiding outdoor sports activities and wearing N95 respirator when walking outside in the PM2.5 air pollution in a higher position than those not. Implications for future research and practice are discussed, based on a critique of the present work.

The purpose of this multi-site comparative study was to engage music educators in a process to uncover broader perspectives on their pedagogy by breaking down the barriers between general education pedagogy and music education. The curriculum planning and instruction of music teachers were observed through Schwab's Commonplaces framework to identify connections between their initial approaches and changes made during the beginning of the 2020-2021 school year. Participants were seven New York City middle school general music teachers. Data were collected from participants in two sets, each consisting of one questionnaire in Qualtrics, and one interview on Zoom for a total of four instruments. The data analysis process was as follows; (a) data organization, (b) first cycle structural coding, (c) second cycle coding, and (d) synthesis and cross-case analysis. The study addressed the following research questions: (a) How can the curriculum planning, and instruction of music teachers be observed in relation to Schwab's commonplaces? (b) What connections might be inferred between these observations and any later curriculum or instructional changes (or lack thereof) made by teachers? (c) How might the schooling changes resulting from the Covid-19 outbreak have impacted these decisions? (d) What impact and/or changes in student engagement and learning might be observed by teachers during the period of this study? The findings were as follows; (a) Commonplace lens/es for curriculum planning and instruction were misidentified by participants, Learner was the most emphasized Commonplace instruction lens and four participants were unable to differentiate between curriculum and instruction, (b) Teachers' more accurately identified the Commonplace lens/es in the second data set, Learner was the most emphasized Commonplace lens for curriculum planning and instruction, and student feedback and/or engagement influenced curriculum changes, (c) COVID-19 affected participants' emotions, attitudes, and decision-making, school reopening structures frequently changed, participants simplified curriculum content for remote and reduced instruction time, and altered curriculum and instruction to prioritize students' social-emotional well-being and engagement, and (d) Student engagement and learning looked different due to COVID-19 schooling changes, in-person students showed improved engagement and quality of work, other subjects affected student engagement and learning, which improved after curriculum changes.

Science education has maintained a longstanding goal of enhancing societal interest, values, and understandings of science. Despite a series of public education reforms and efforts by scientific researchers, scientific literacy and passion remain sparse across the American public. In fact, many students demonstrate a lack of interest in the sciences as early as first grade, with major drop-offs occurring by the age of 14. This is further exacerbated for youth of color, as science is deeply rooted in pervasive and institutionalized racism. When accessible, out-of-school science experiences are uniquely positioned to promote youth agency, leverage students’ current values, and challenge structural inequities. Therefore, this work sought to highlight the narratives of three young science learners who identify as youth of color and graduated from an eight-year-long museum science program. A narrative inquiry was implemented, guided by Critical Race Theory and Cultural Learning Pathways as frameworks. Data from semi-structured interviews, questionnaires, and drawings provided insight into the students’ orientations towards science and the development of those orientations. The study took place during the Covid-19 outbreak. Implications of the pandemic on the study are discussed. Findings from the study suggest that students positively identify with science and feel at home in The Museum. It was also noticed, however, that the students maintained ideologies consistent with Western perspectives. Recommendations include creating homeplaces, making out-of-school learning more easily accessible, and creating justice-centered curricula.

The study of the three-dimensional structures of biological macromolecules has given us significant insight into life and its mechanisms. Understanding these structures in their native contexts, a challenging but important goal, came closer to reality with the development of electron microscopy. After many years of technological development, we are now starting to understand previously intractable biological phenomena at an unprecedented resolution. One such phenomenon is how neighboring cells interact, both to communicate and send signals, and to adhere and form complex tissue structures. While the molecules that mediate such processes have long been studied in isolation, electron microscopy allows us to examine them in a more native biophysical environment; as hydrated, dynamic molecules tethered to opposed cellular membranes.Imaging unadulterated biological material using electron microscopy requires that the sample be embedded in a thin layer of vitreous ice to immobilize the molecules and protect them from the vacuum of the microscope, and thus is generally referred to as cryogenic electron microscopy (cryo-EM). Samples can be imaged using two common cryo-EM modalities: single particle analysis (SPA), where many two-dimensional projection images of molecules in solution are collected, and cryo-electron tomography (cryo-ET), where the sample is tilted as it is imaged at multiple angles to reconstruct a three-dimensional volume. In this work, I will describe how I have used both SPA and cryo-ET to understand cell surface interactions involving a variety of proteins. The first chapter will look at the cell surface molecules known as the Toll receptors, a family of molecules found in Drosophila melanogaster, with orthologs in mammals known as the Toll-like receptors (TLRs). I will focus on their role in the development of the Drosophila embryo during germ band extension, a kind of convergent extension that is a conserved process through all metazoans. Biophysical assays of the three implicated Toll receptors, Toll-2, -6, and -8, revealed both homophilic and heterophilic interactions. SPA was used to determine the structure of monomeric Toll-2 which closely resembles the overall fold of Toll, whose structure was previously solved by x-ray crystallography. Surface plasmon resonance (SPR) spectroscopy and analytical ultracentrifugation (AUC) showed Toll-6 is a dimer in solution, which I visualized using cryo-EM. The Toll-6 homodimer is a novel dimer interface for Tolls and TLRs, where molecules on the same cell surface have been shown to dimerize in the presence of a wide variety of ligands. In contrast, the Toll-6 dimer is formed in the absence of any ligand and exists in an antiparallel arrangement that could be formed by molecules on opposing cell surfaces. Together, these results provide a biochemical basis for germ band extension which may be further explored through the study of structure-based mutations. While cryo-EM SPA is a powerful tool, cryo-ET allows one to reconstruct three dimensional volumes of highly heterogeneous samples, such as the interior of cells, where molecules of interest may not exist in enough copies to facilitate averaging. This technique, where the sample is imaged multiple times as it is tilted to obtain three-dimensional information of a region of interest, was used to study cell adhesion of a different type: that mediated by the classical cadherins. These calcium-dependent adhesion molecules cluster into adherens junctions, spot-like protein densities found in a wide variety of tissues. In the second chapter, these junctions are recapitulated between synthetic liposome membranes by tethering the adherent cadherin molecules to chemically functionalized lipids. They are then imaged using cryo-ET to reveal higher-order structural details. First, this method is applied to the clustered protocadherins, a family of cadherins that mediate neuronal self-avoidance in mammals. Cryo-ET in combination with x-ray crystallography revealed that clustered protocadherins form extended one-dimensional zippers between membranes, which are a combination of strictly homophilic trans interactions coupled with promiscuous cis interactions. Neurons express unique subsets of the ~50-60 possible isoforms, and when two neuronal processes express identical subsets, which happens only when those processes are a part of the same cell, these linear chains grow and initiate a repulsive signal. If the subsets are different, the chains terminate and no repulsive signal is generated. The same technique has been used previously to study the type I classical cadherins, perhaps the most well-studied members of the cadherin superfamily. In the second half of this chapter, we extend our analysis to include the type II classical cadherins, which possess more complex expression patterns and binding specificities. Cryo-ET of type II cadherin ectodomains tethered to synthetic liposomes revealed that several representative members of this family form only moderately ordered arrays between liposomes, a finding in agreement with their role in cell sorting and migration. However, VE-cadherin, an outlier type II expressed in vascular endothelial cells where it withstands blood pressure, forms extraordinarily ordered junctions. Subtomogram averaging reveals the regularity of this two-dimensional array. In the final chapter, I describe my work on a membrane surface molecule of a different kind, one not involved in cell adhesion but viral infection. The global COVID-19 pandemic gave me the opportunity to contribute to our understanding of SARS-CoV-2 by studying the structure of neutralizing antibodies bound to the viral spike protein, perhaps the most infamous membrane surface protein. The first subchapter describes the initial isolation, neutralization, and structural analysis of antibodies isolated from convalescent COVID-19 patients. This work revealed that patients with severe COVID-19 produce potently neutralizing antibodies that target two spike protein domains: the receptor binding domain (RBD) and the N-terminal domain (NTD). RBD-directed antibodies occlude binding to ACE2, the human receptor that mediates viral fusion, but the neutralization mechanism of NTD-directed antibodies is unknown. The following two subchapters are more detailed structural studies of two specific types of antibodies. The first looks at a class of RBD-directed antibodies derived from the VH1-2 gene, which are some of the most potent and common antibodies against SARS-CoV-2. The heavy chains of these antibodies recognize almost identical epitopes, but the antibodies employ a modular approach to recognize the RBD in either of its possible conformations. The second class are antibodies that target the NTD, which our work revealed all bind to a single antigenic supersite. The final subchapter focuses on emerging SARS-CoV-2 variants and includes the structures of two antibodies that are still capable of neutralizing these new variants. They are also infrequent in the human antibody response to SARS-CoV-2, meaning they put little selective pressure on the virus to produce escape mutations, making them good candidates for monoclonal antibody therapies. Though Drosophila embryogenesis, adherens junction formation, and SARS-CoV-2 neutralization are seemingly unrelated systems, they are united by the incredible flexibility of cryo-EM to visualize biological molecules in more native environments. Whether it is the ability to study multiprotein complexes or assemblies formed between membranes, cryo-EM is a powerful technique that promises to help bridge the divide between structure and function.

Markov Decision Processes (MDP) is a common framework for modeling sequential decision-making problems, with applications ranging from inventory and supply chains to healthcare applications, autonomous driving and solving repeated games. Despite its modeling power, two fundamental challenges arise when using the MDP framework in real-worldapplications. First, the optimal decision rule may be highly dependent on the MDP parameters (e.g., transition rates across states and rewards for each state-action pair). When the parameters are miss-estimated, the resulting decision rule may be suboptimal when deployed in practice. Additionally, the optimal policies computed by state-of-the-art algorithms may not be interpretable and can be seen as a black-box. Among other reasons, this is problematic as the policy may not be understandable for the people that are supposed to operationalize it. In this thesis, we aim to broaden the applicability of the MDP framework by addressing the challenges of robustness and interpretability. In the first part of the thesis, we focus on robustness. We introduce a novel model for parameter uncertainty in Chapter 2, that is significantly less pessimistic than prior models of uncertainty while enabling the efficient computation of a robust policy. In Chapter 3, we consider a healthcare application, where we focus on proactively transferring patients of a hospital to the Intensive Care Unit, to ameliorate the overall survival rates and patients’ flow. In the second part of this thesis, we focus on interpretable algorithms, with an emphasis on the application to find novel triage protocols for ventilator allocations for COVID-19 patients. In Chapter 4, we introduce a simulation model to estimate the performance of the official New York State (NYS) triage protocol at various levels of shortages of ventilators, using a real data set of patients intubated during Spring 2020 because of COVID-19 complications. In Chapter 5, we introduce our algorithmic framework for computing interpretable (tree) policies and apply our methods to learn novel triage protocols.

Due to the COVID-19 pandemic many schools were forced to interrupt in-person delivery of educational services and switched to delivery of instruction in a remote setting. The educational impacts of school closures and remote instructional delivery have become a concern for the impact on an entire generation of students. Although delivery of behavioral interventions remotely is a topic that has been reported on in behavior analytic literature for over 15 years before this pandemic, few studies directly compared delivery of equivalent services across in-person and remote settings. Further, no studies included in recent literature reviews included comparative analyses between acquisition of novel instructional objectives across in-person and remote settings. Additionally, no studies are reported using verbal behavior developmental measures to identify potential prerequisites for benefitting from remote instruction. I therefore present a series of experiments to investigate the relative effectiveness of in-person and remote instruction as well as comparing outcomes for students grouped by level of verbal behavior development. In Experiment I, I conducted a carefully controlled experiment to compare rate of learning, rate of instructional presentation, and maintenance of objectives mastered across in-person and remote settings. I used a reversal design across 6 preschool aged participants with disabilities. The results indicate that some participants reliably mastered objectives and completed instruction faster in-person for 3 of 6 participants while the results for the other 3 participants were mixed. Overall, participants mastered objectives and completed instruction faster in-person in approximately half of the comparisons while showing no difference or learning faster remotely in the remaining comparisons. No consistent difference was shown in 14 and 21-day follow up maintenance measures. No consistent difference in the outcomes of students who demonstrated Naming compared to those who did not demonstrate any Naming. In Experiment II, I extend the findings of the previous experiments by comparing system wide educational outcomes of a hybrid in-person and remote educational model to the outcomes of the same model in a pure in-person setting in terms of fidelity of instruction, educational outcomes, and a cost analysis to determine how much the transition to remote provision of instruction costs stakeholders. Further, I compared educational outcomes across students categorized by level of verbal behavior development. The results indicate that the total number of learning opportunities and objectives mastered are significantly higher during a fully in-person model when compared to a hybrid educational model containing a remote instructional component. Further, when comparing educational outcomes across groups of students categorized by level of verbal behavior development, the results indicate that the rate of learning and objectives masted are significantly increased once students demonstrate joining of the listener and speaker repertoires as indicated by the presence of the Naming capability. The implication of the results are discussed in terms of feasibility of remote instruction as an alternative to in-person instruction as well as the importance of identifying and establishing the Naming capability for students to best benefit from remote instruction.Keywords: Telehealth, COVID-19, remote instruction, virtual instruction, verbal behavior development

This dissertation combines micro-level data and partial equilibrium models to understand how financial policies affect non-financial firms, with a particular focus on the role played by banks in such transmission. In the first chapter we study a large-scale intervention in the Brazilian banking sector characterized by a sudden increase in the supply of credit provided by commercial government banks. Theoretically, the effect of this type of policy is ambiguous: while it can increase the total amount of credit in the economy, it can also increase misallocation if government banks finance riskier firms with unproductive projects. We leverage credit registry data to document a series of empirical facts and test if the intervention alleviated inefficient underprovision of credit. We find that while the intervention led to a reduction in interest rates and to an increase in total credit from both government and private banks, government banks faced a significant increase in defaults. Loans to ex-ante indebted firms explain this increase in default for government banks. Moreover, neither the increase in total credit nor the reduction in interest rates had any observable effects on output or employment. Our results suggest that the intervention increased credit misallocation, and that adverse selection did not play a significant role in the allocation of credit in Brazil. In the second chapter, we assess the role of banks in the Paycheck Protection Program (PPP), a large and unprecedented small-business support program instituted as a response to the COVID-19 crisis in the United States. In 2020, the PPP administered more than $525 billion in loans and grants to small businesses through the banking system. First, we provide empirical evidence of heterogeneity in the allocation of PPP loans. Firms that were larger and less affected by the COVID-19 crisis received loans earlier, even in a within-bank analysis. Second, we develop a model of PPP allocation through banks that is consistent with the data. We show that research designs based on bank or regional shocks in PPP disbursement, common in the empirical literature, cannot directly identify the overall effect of the program. Bank targeting implies that these designs can, at best, recover the effect of the PPP on a set of firms that is endogenous, changes over time, and is systematically different from the overall set of firms that ultimately receive PPP loans. We propose and implement a model-based method to estimate the overall effect of the program and find that the PPP saved 7.5 million jobs. In the third chapter we further explore the Paycheck Protection Program (PPP) by asking what is the optimal allocation of funds across firms and the distortions caused by allocating these funds through banks. We show that it can be optimal to allocate funds to the least or most affected firms depending on the underlying distribution of the shock that firms face, the firms' financial position, and the total budget available for the program. In the model, as in the data, banks distort the allocation toward firms with more pre-pandemic debt and those less affected by the COVID-19 crisis. We characterize how this misallocation depends on the degree of asymmetric information between banks and the government. In an empirical application of our model, we estimate the PPP's effectiveness and compare it with alternative policies. A policy targeted at the smallest firms could have increased the program's effectiveness significantly.

This dissertation applies place-based assimilation theories to understand the role of the school and other community-based institutions in the lives of immigrant families in a small rural town. The rate of immigration is increasing globally and over time, more and more children and families will be immigrants, finding themselves in a new community, making it imperative to understand the lived experiences of immigrant children and families. For most migrant families with school-aged children the school is the first point of contact in a new community. Thus, the school is well positioned to assist families in the integration process providing them with vital information and connections to resource-rich community-based institutions. This dissertation explores the relationship between families and community-based institutions in Provincetown Massachusetts, a small, coastal, rural community with a significant immigrant population. It is a narrative inquiry that employs qualitative research methods, specifically semi-structured interviews and visual research methods including photographs taken by immigrant students and photo elicitation interviews to answer the following questions: 1) What role does the school play in the process of immigrant families integrating into a new community?; 2) How do community-based institutions help or hinder immigrant families accessing resources and developing a sense of belonging?; 3) In what ways has the current COVID-19 health pandemic affected the work of community-based institutions and immigrant families’ interactions with them? Newly arrived families to Provincetown face food and housing insecurity and a lack of access to health care. There is however, a comprehensive web of community-based institutions with programs and resources to meet those needs. Access to most of these resources requires a referral or connection from an agency like the school, so families are reliant on schools for connection to these institutions. The school has formal mechanisms in place to help families. There are also informal mechanisms in the school to help families. This consists of individual teachers who develop deep and lasting relationships with a particular student and assist this student and his or her family using their own time and resources. This dissertation also explored the ways in which immigrant children in Provincetown find belonging. The children reported that they find belonging in the natural environment, through enrichment activities such as art clubs and sports teams, and through participation in the tourism work force, either by helping family members or beginning to work on their own. There are many institutions that work with the school and families to provide access to this enrichment programming, but there are barriers to participation. Immigrant children are often prevented from participating in enrichment activities outside of school hours because they have to care for younger siblings or lack transportation to and from afterschool events. There is also a disconnect between institutions and families because some institutions struggle to communicate with families. Some institutions have tried to respond to these barriers by providing transportation and parallel programming for siblings. This study also found that the school was the most successful way for institutions to communicate with families because of the well established communication patterns, available translation services and presence of school personnel who have taken an active interest in the outside lives of students. Many solutions in Provincetown are place-specific and the experiences of families in Provincetown are atypical because there are several factors that make Provincetown unique. It is a tourist town with access to financial resources that can fund many institutions and opportunities. The town is small, making the relationship between families and institutions more personal so that individuals and institutions become more invested in the lives and outcomes of individual families in a way that would not be possible in an urban area.

Efforts to improve learning outcomes in undergraduate mathematics are backed by calls for instructors to move away from persistent pedagogical norms, such as traditional “chalk-talk lectures” and timed closed-book exams. Although the movement towards active learning pedagogies and alternative assessments is gaining ground, uptake has been slow. But when traditional practices became virtually impossible to maintain during the coronavirus pandemic, many instructors quickly gave up old methods for new ones. This qualitative study sheds light on how twenty-eight instructors of undergraduate mathematics experienced teaching through the coronavirus pandemic. It documents stories they told to describe their experiences and explores how they adjusted their teaching practices. A purposeful sample of instructors whose teaching experience spans 58.5 years participated in the study. They taught lower- and upper-division courses during the pandemic and came from colleges and universities across the United States. The primary data collection method was semi-structured interviews. Four salient storylines emerged from data: (1) Remembering other times of change and disruption, (2) Pivoting to emergency online instruction, (3) “Like flying blind:” navigating a new normal, and (4) Coping. In addition, analysis of participants’ stories revealed that they adjusted five dimensions of practice for teaching during the pandemic: (1) Representing Mathematical Content, (2) Choosing Tasks and Content, (3) Monitoring: “Leaning Over Their Shoulders,” (4) Building a Community of Mathematics Learners and Doers, (5) Assessing for Learning During the Pandemic. The lessons we learn from this emergency will be “educative” for this crisis and the next. Moreover, the coronavirus pandemic also offers an opportunity to examine longstanding norms in undergraduate mathematics education. This study offers recommendations for practitioners, leadership, and further research.