Student motivation is a well-established predictor of academic engagement and achievement. Despite its importance, motivation is difficult to assess accurately, as it is shaped by multiple, overlapping constructs and contextual factors. This complexity continues to pose challenges for researchers and practitioners seeking measurement tools with valid and reliable scores. While open-ended responses provide meaningful information about students’ motivation, traditional human scoring is time-consuming, subject to inconsistencies, and not easily scalable. Advances in natural language processing, especially with transformer-based large language models, offer promising new approaches for automated scoring. However, their use in scoring student motivation assessments remains largely undocumented, and practical applications are in the early stages of development. This dissertation addresses this gap through two studies. The first study presents a scoping review of transformer-based approaches to motivation scoring, examining model types, scoring methods, target constructs, and available psychometric evidence. Guided by the PRISMA-ScR protocol, the review identified only nine studies, underscoring the limited research in this area. The second study empirically evaluates transformer models for motivation assessment, comparing fine-tuned encoder models (BERT, DeBERTa) with rubric-prompted decoder models (GPT-3.5 Turbo, GPT-4.1). Model performance is evaluated through a psychometric lens, including analyses of human–machine agreement, construct validity, subgroup fairness, and predictive validity with academic outcomes such as GPA and retention. Notably, DeBERTa outperformed human raters in predicting academic outcomes, suggesting that certain transformer models may capture motivational signals with greater predictive value—even when agreement with human scores is modest. However, validity analyses revealed important limitations: models demonstrated inconsistent construct validity and notable subgroup disparities, raising concerns about fairness. Overall, while transformer-based approaches, specifically DeBERTa, hold promise for scalable motivation assessment, challenges remain in ensuring interpretability, validity, and equity. This study also discusses the challenges and limitations encountered, outlines practical implications for educational assessment, and identifies directions for future research in developing robust and fair applications of transformer models.
The acetone-to-isobutene (ATIB) conversion is part of the ethanol cascade reaction to produce highly valuable compounds. Lewis acid sites can lead to selective C–C coupling and self-deoxygenation pathways in the acetone aldol condensation reaction. However, the reaction mechanism, nature of active sites, and the role of local environment are not well understood. This work demonstrates that well-dispersed isolated Ce atoms within the dealuminated BEA zeolite (CedeAlBEA) exhibit exceptional isobutene formation during the ATIB reaction by providing optimal Lewis acidity, which facilitates the direct decomposition of diacetone alcohol (DAA) to isobutene. In contrast, elevated cerium loadings lead to the formation of CeO2 nanoparticles, which favor side reactions such as DAA dehydration to mesityl oxide (MSO). In situ DRIFTS experiments with acetone adsorption reveal that adjacent silanol (Si–OH) groups in open Ce sites play an essential role in acetone activation, enabling the concerted adsorption of two acetone molecules, on the OH group and the Lewis active site, thus facilitating selective C–C bond formation and stabilizing the transition state for isobutene production. Kinetic assessments, including reactions with DAA as the reactant and kinetic isotope effect measurements, demonstrated that the direct decomposition of DAA to isobutene and acetic acid is the rate-limiting step. Accordingly, a Langmuir-Hinshelwood mechanism is proposed, involving the cooperative adsorption of two acetones molecules on isolated Ce species and the adjacent Si–OH groups. Building on these insights, bimetallic Ce–M catalysts were developed by introduction of a second metal with distinct electronic properties to enhance metal dispersion. In this system, the isolated Ce atoms serve as anchors for the second metal, generating dual Lewis acid sites. Combinations such as Ce–Ga and Ce–Sc preserved high dispersion and achieved 100% theoretical selectivity to isobutene (88.9% yield) even at high total metal loadings. These findings demonstrate that pairing confined Ce atoms with a secondary metal can overcome dispersion limitations, increase active site density, and mitigate metal aggregation, offering a novel strategy for the rational design of bifunctional zeolite catalysts for selective C–C coupling and deoxygenation reactions.
Abstract In everyday life, people intuitively use space to make meaningful distinctions between objects. In this paper, we demonstrate how geometric, spatial representations can capture a wide variety of cognitive phenomena. To conduct these experiments, we developed GRIS (Generating Representations In Space), an advancement of prior spatial-arrangement paradigms. In GRIS experiments, participants manipulate a set of objects (text, audio, images) and place them on canvases. Following an introduction to the paradigm, we present three studies which demonstrate how experiments in the GRIS paradigm can both a) reveal nuanced insights about human and computational representations, and b) capture low-level behavioral data of spatial representations as they are constructed.
Abstract How do urban environments shape the way people conceptualize and communicate space? Using a cultural transmission experiment in virtual reality, we asked participants to navigate one of two contrasting cityscapes—a regular, grid-like “Manhattan” or an irregular, landmark-rich “Barcelona”—and to produce route descriptions and sketch maps. We examined participants’ spatial conceptualizations and tracked how spatial communication was transmitted across six simulated generations. Results show that city layout shaped spatial language and map production. Participants used more horizontal (e.g., numbered routes and intersections) descriptors in the “Manhattan” environment and more vertical (e.g., buildings distinguished by colors and textures) descriptors in the “Barcelona” environment. These strategies were also more effective: participants completed navigation tasks faster when the descriptions they received matched the structure of the environment. We observed robust cultural transmission of spatial conventions, with participants’ productions predicted by the linguistic input they received. However, rather than showing cumulative elaboration over time, these conventions stabilized quickly. Individual-level analyses revealed consistency between language and maps, as well as across environments: participants formed coherent spatial models that persisted and shaped adaptation to a new cityscape. This suggests that individual cognitive stability may act as a mechanism for preserving, rather than progressively refining, cultural conventions. Together, these findings highlight urban environments as cognitive–cultural niches that shape how people perceive, conceptualize, and communicate space. We discuss implications for theories of spatial language, cultural evolution, and urban design, showing how city layout shapes not only navigation behavior, but also the structure of spatial thought.
