Dissertation Defense Announcements

In Chapter 1, we show that every shift of finite type (SFT), sofic shift, and strongly irreducible shift on locally finite groups have strong dynamical properties. We show that every sofic shift is an SFT, every SFT is strongly irreducible, and every strongly irreducible shift is an SFT, among other properties. Furthermore, we show that if any of these properties hold for a group, then the group must be locally finite. In pursuit of these results we present a formal construction of free extension shifts on a group $G$, which takes a shift on a subgroup $H$ of $G$ and naturally extends it to a shift on all of $G$.

In Chapter 2, we study measurable dynamical systems on compact metrizable spaces. We introduce the completion of a dynamical system, and show that it has many strong properties. We introduce Birkhoff systems, for which a version of the pointwise ergodic theorem holds as many classically systems do, using completions. For Birkhoff systems, we define dynamical independence, a property which induces probabilistic independence in ergodic measures. We apply the concepts to a variety of systems, and prove an extension of De Finetti's Theorem which characterizes exchangeability in terms of dynamical properties.

HEATHER ELDEB RYAN. Social Emotional Learning: A Mixed-Methods Study on Mitigating Test Anxiety and Increasing Agency. (Under the direction of DR. TINA HEAFNER)
Abstract
Middle School student academic achievement has not yet reached its potential. A multitude of educational studies have examined the causes and potential solutions that could foster long-term achievement gains for students in these pivotal adolescent years. Much of the prior research has looked at the pedagogical side of teaching, the programmatic side of curriculum, and the multitude of psychological influences experienced by the learner. Although previous work has been absolutely beneficial, these investigations have often been done in isolation. This quasi-experimental mixed-method study attempted to amalgamate pedagogy, curriculum, and psychological influences by integrating social emotional instruction techniques with targeted curriculum to address emotional and behavioral factors that affect student academic performance. A sample of 108 students aged 10-11 years old were placed into four treatment (n =59) and four control groups (n =49), by purposeful sampling. Students in the treatment classrooms received one intervention lesson on motivation and one lesson on anxiety reducing skills. By using an explanatory mixed-method approach, the research used two phases to uncover and track students’ test anxiety, math avoidance, and motivation as potential influences on fifth-grade agency for learning over a nine-week period. Surveys were administered in conjunction with weekly math exams over four separate data points. After repeated ANOVA analysis had been completed on preliminary quantitative data, a structured interview protocol was developed to obtain a deeper understanding of the participant experience for students in both the treatment and control groups. Initial survey analysis found no significant changes by either group in student test anxiety, math avoidance, or motivation towards agency over the research period. Further analysis showed no significant effect of the intervention lessons on the treatment group’s test anxiety, math avoidance, or motivation towards agency. Contrarily, interview data supported previous research findings with the majority of students stating that they had experienced test anxiety, and math avoidance, and had a desire to be more intrinsically motivated. Additionally, respondents from the treatment group perceived the motivational lesson helpful in making intrinsically motivated agentic choices with future study behaviors and the enrollment in advanced math courses. Similarly, 100% of treatment participants believed that the instruction on anxiety skills was useful and would be utilized when necessary. Moreover, interview findings discovered that over half of students in the intervention group had successfully implemented anxiety reducing skills both inside and outside of the school environment post-treatment. Overall findings justify further integration and investigation of social emotional instruction in the general education classroom as a potential solution to increased student achievement.

This quantitative study investigated factors that predict faculty’s intentions to integrate social media into language courses to facilitate student learning at four-year higher education institutions in the United States, using the Decomposed Theory of Planned Behavior (DTPB) framework. Data were collected via an online survey from faculty teaching foreign languages in four-year higher education institutions in the United States (N=135). A descriptive analysis of the faculty’s social media proficiency level was generally aligned with their use in courses. Collaborative authoring/editing tools and media-sharing tools were the most used, while microblogging and podcasting were the least used. The results of a path analysis revealed foreign language faculty’s attitudes as the strongest determinant of their intentions to integrate social media applications into language courses. The qualitative analysis of open-ended survey responses showed that the perceived usefulness of social media was driven by perceived benefits in increasing student engagement and making the class fun. Conversely, the most commonly reported barrier to using social media was the time required to prepare materials, followed by content appropriateness. In addition, peer influence was the determinant of subjective norm, while self-efficacy was the determinant of perceived behavioral controls, implying these two constructs indirectly influenced faculty’s intentions. Multiple regression analysis confirmed that faculty’s intentions towards using social media in language courses differ based on gender. The findings had significant implications for faculty teaching language courses, professional development programs, and curriculum design.

The growing integration of artificial intelligence (AI) and wearable technology in healthcare presents opportunities to reduce health inequities tied to social determinants of health (SDOH). However, low adoption rates—AI (2%) and wearable devices (4%)—raise concerns about their effectiveness in improving clinical quality outcomes (CQOs). This study explores whether AI and wearable devices moderate the impact of SDOH factors (geographic location, insurance status, race/ethnicity) on adverse CQOs.

Using secondary data from the All of Us Research Program (N = 39,632), binary logistic regression assessed AI and wearable usage on adverse CQOs (15%). Findings indicate AI adoption is minimal (2%) and does not significantly reduce rural-urban (p = 0.276) or insurance-based disparities (p = 0.194). Instead, AI usage increases the risk of adverse CQOs for people of color (OR = 2.864, p < 0.001), pointing to potential algorithmic bias or systemic inequities in AI-driven healthcare.

Wearable device usage (4%) had mixed effects. Uninsured users faced a 64% higher risk of adverse CQOs (OR = 1.641, p = 0.034), suggesting wearables are more reactive than preventive. However, Hispanic individuals using wearables saw an 87% reduction in adverse CQOs (OR = 0.126, p = 0.042), indicating population-specific benefits.

These findings highlight the need for equity-driven AI policies to ensure healthcare technologies address rather than reinforce disparities. Future research should focus on strategic AI implementation and policy-driven incentives to enhance digital health equity and clinical quality.

KEYWORDS: Artificial Intelligence, Fitbit, Social Determinants of Health, Clinical Quality Outcomes, Adverse Events, Health Equity, Digital Health, All of Us