Abstract
Students are constantly asked to make sense of numbers about their performance. Some come from standardized tests, while some come from course grading systems. What interests me is not just the information itself, but what students do with it once they have it. A number can seem small and factual when it first appears on a screen or in a document, but it does not stay there. Students carry it into later decisions. They use it to judge whether they are doing well, whether a goal still feels possible, and whether a challenge seems worth the risk. My thesis portfolio is about that process. Both of my projects look at how academic performance information gets interpreted and acted on. One project focuses on course syllabi and grade forecasting, while the other focuses on percentile rankings and the role they play in shaping student effort, risk-taking, and educational opportunity. While the two projects work at different scales, they are connected by the same concern: once performance information begins to shape how students see themselves, it can also shape what they believe is realistic to pursue.
My technical project, Syllabus-Aware Grade Forecasting: NLP-Based Syllabus Parsing and Scenario Grade Modeling, comes from a very practical problem. Students often rely on syllabi to understand how a course works, but syllabi are not always easy to use. Important grading policies may be spread across different sections, written in inconsistent language, or buried in long paragraphs that students do not fully process until they are already stressed and trying to calculate what happens next. That can make it difficult to answer basic questions. A student may want to know how much a missed quiz matters, what grade they need on the final to reach a certain target, or whether a late penalty changes the semester outcome in a serious way. Existing tools do not always help much with this. Learning management systems may leave out course-specific rules, and manual spreadsheets depend on students entering everything correctly themselves. My project proposes a system that parses a syllabus, extracts its grading rules, and turns those rules into a form that can be used for transparent grade forecasting. The goal is to help students understand how that number is being produced and what choices still matter. In that sense, the project is really about making course policy easier to read and easier to use when students are trying to make decisions under pressure.
My STS research paper looks at another kind of academic signal: percentile rankings on standardized and national tests. The paper asks how those rankings affect later student effort and willingness to take academic risks. I was interested in the fact that a percentile can seem like a simple statement of comparison while still taking on much more meaning once it starts moving through school systems. A score report may present the percentile as straightforward information, but students do not read it in isolation. Families, counselors, and schools use it in placement decisions and other kinds of sorting. Over time, that number can start to feel like evidence of what kind of student someone is. In the paper, I argue that percentile rankings should not be understood as purely technical measures of comparative performance. They function more like signals that gain force through the institutions and relationships around them. Drawing on scholarship about academic self-concept, tracking, and educational stratification, I show how percentile rankings can shape what students believe is worth attempting. Through the STS concept of surveillance as social sorting, I examine how these rankings can help organize opportunity by influencing confidence, ambition, and the perceived risk of reaching for something harder. What matters to me in this argument is that the ranking does not just describe a student’s place. It can start helping produce that place.
These two projects are connected by that shared interest in what happens after performance information is received. My technical portion asks how students can be given tools that make grading systems easier to understand, so they are not forced to make important decisions based on confusion or guesswork. My STS paper steps back and looks at how comparative performance signals can shape student self-understanding over time. While one project is focused on design and usability, the other is focused on interpretational power. However, both ask versions of the same question. What happens when students are given numbers that seem objective, and those numbers begin shaping what they think they can do? That is the thread that ties this portfolio together. I am interested in the point where academic information stops being just information and starts affecting possibility.
