Statistics undergraduates present their Capstone work

Graduating biostatistics students present and discuss their Capstone projects.

Each year, Brown’s Undergraduate Statistics Concentrators present Capstone projects to the Department of Biostatistics faculty, students and visitors. This year, 7 graduating statistics undergraduate concentrators presented their Capstone work on April 21st. This is an opportunity for graduating students to present and discuss their Capstone projects which meets a graduation requirement and also qualifies each candidate for the annual Undergraduate Statistics Concentration Capstone Poster Award of Excellence which is announced at our annual Biostatistics Commencement Celebration.

Here are the 2023 candidates:

Author: Nathan Provost

Poster Title: Asymptotic Error Bounds for Nonparametric Censoring Unbiased Estimators of Restricted Mean Survival Times
Abstract: We extend the results presented by Antos et al. [1] for nonparametric estimation methods using complete datasets to nonparametric regression estimators that use right censored (incomplete) datasets. Specifically, we establish novel asymptotic least upper error bounds for the estimators first constructed by Buckley and James [6], Koul et al. [7], and Rubin and van der Laan. [4] Moreover, we prove two corollaries that follow from our results under the Dedekind-MacNeille completion of the real numbers. [8] We then discuss the significance of these error bounds, as well as their limitations and the impact of the assumptions under which they hold. Additionally, we run simulations for all three estimators, which provide some practical support to our theoretical results. Finally, we comment on the mathematical limitations of our theoretical results and the computational limitations of our simulations, pointing out areas of improvement and potential future research avenues.

Author: Annorjan Naguleswaran

Poster Title: Predictive Investments Simulator (R Shiny App)
Abstract: Within the past few decades, there have been several interesting trends within the stock market. For example, with the 2021 GameStop craze and rise of retail investor-pumped stocks has come additional attention to the markets. Within this presentation, I outline the results of an R Shiny App created to help people see how stocks they’re interested in investing in are simulated/expected to change in the future (and thus help you gauge how your investments are expected to perform).

Author: Abigail Sinotte

Poster Title: The Effect of Race to the Top on Student Achievement
Abstract: How have Race to the Top (RTTT) education grants impacted student achievement? I analyzed testing outcomes for fourth grade and eighth grade students measured by the National Assessment of Educational Progress (NAEP). I employed an event study empirical strategy to exploit any changes in achievement after RTTT implementation. Overall, there were relative increases in standardized test scores and proficiency across both grades and exams following the RTTT competition. This work also analyzed RTTT’s impact on demographic subgroups, as a key aim of this program was to close achievement gaps. There were inconsistent results seen for proficiency changes of historically low-performing groups of students such as English language learners and students eligible for free/reduced price lunches in states that received RTTT funding.

Author: Colby Zarle

Poster Title: The Effects of California Wildfires on Student Test Scores
Abstract: This project examines the impact of California wildfires on student academic performance by analyzing test scores from the California Assessment of Student Performance and Progress (CAASPP) and the CalMatters Disaster Days Public School Closure Database. While previous research has explored the economic impacts of wildfires, this study specifically focuses on the effects of school closures due to wildfires on academic outcomes. The analysis uses a two-ways linear fixed effects regression to estimate the effects of school closures on academic performance, and the data is constrained to the five academic years from 2014/15 to 2018/19 due to the COVID-19 pandemic limiting testing protocols in later years. The findings indicate a negative effect of 1 additional day of school closures on the percentage of students meeting or exceeding the standard for both ELA and Math scores for all grades, with 6th graders showing the greatest impact. Additionally, medium-sized schools suffer more than small and large schools.

Author: Linda Wakamoto

Poster Title: The Association Between Social Determinants of Health and Risk of Hypertension
Objective: It is not clear how social determinants, primary care access and hypertension hospitalization outcomes interact. A better understanding of this correlation will help identify the most vulnerable patient subgroups that may benefit from targeted preventions for poor hypertension control.
Methods: This was a secondary data analysis, and a retrospective cohort study. This project utilized the data from Dr. Tingting Zhang and Dr. Janette Baird’s current Social Determinants of Health research project, which is based on the Rhode Island Department of Health Ecosystem data. The study population was defined as Medicaid enrollees with hypertension, who were identified using a validated case definition based on outpatient and inpatient diagnosis. The baseline period was 12 months of Medicaid enrollment, and patients were studied whether they were diagnosed with hypertension after this period. Follow-up period was until the earliest event: disenrollment or the end of their study (December 31, 2020). The incidence rate of hypertension was calculated by dividing the number of new cases by the total number of person days of follow-up.
Results/Conclusion: The incidence rate of hypertension for those who were homeless or unemployed was 17.2% higher than those who were not. Age had a significant correlation with hypertension at a significance level of 0.05, but sex and unemployment at follow-up did not. Further studies with larger sample size and more rigorous adjustment are needed to examine the association between these social determinants and the development of hypertension.