Abstraction Transition in Computer Science Education: A Framework for Functions First Pedagogy and Deep Computational Understanding in CS0
Sparad:
| I publikationen: | ProQuest Dissertations and Theses (2025) |
|---|---|
| Huvudupphov: | |
| Utgiven: |
ProQuest Dissertations & Theses
|
| Ämnen: | |
| Länkar: | Citation/Abstract Full Text - PDF |
| Taggar: |
Inga taggar, Lägg till första taggen!
|
| Abstrakt: | Research ProblemComputer science education has long struggled to design effective introductory curricula that balance technical proficiency with conceptual understanding. Despite advances in pedagogy, first-year CS courses continue to suffer from high dropout rates, particularly among students with little prior exposure to programming. A key factor in this attrition is the reliance on syntax-heavy, industry-driven programming languages, which often overwhelm beginners and detract from deeper computational thinking and abstraction skills.Although the specific programming languages used in introductory courses have changed—from Algol60 to Python, the underlying instructional approach has remained largely static, focusing on syntax and language-specific details rather than transferable computational problem-solving skills. This stagnation runs counter to Alan J. Perlis’s vision (Greenberger, 1964), which argued that computation should be recognized as the third pillar of literacy, alongside mathematics and literature. For non-majors, this traditional language-first approach creates a misalignment between academic objectives and the broader applications of computational thinking across disciplines. To address these challenges, CS education must shift from teaching programming languages as an end goal to fostering abstraction, computational problem-solving, and cross-disciplinary adaptability (Guzdial, 2019).MethodologyThis study introduces an Abstraction Transition Framework, integrating a Functions First approach that trains students to transition between problem statements, pseudocode, and executable code (Cutts et al., 2012). The curriculum leverages reduced-syntax programming languages, particularly functional and declarative paradigms, to minimize cognitive load and emphasize problem decomposition, algorithmic reasoning, and abstraction. Functional programming was chosen because its emphasis on function composition, immutability, and structured recursion naturally aligns with abstraction centered instruction.To evaluate the effectiveness of abstraction transition instruction, this research formally followed student cohorts in secondary school settings, tracking their progression from introductory computer science courses into subsequent CS coursework and concurrent mathematics classes. Additionally, the framework was informally adapted for use in both undergraduate CS0 courses and primary school settings, where its applicability was explored in early computational thinking activities. These informal implementations provided preliminary insights into how abstraction transition principles might support learners across different educational stages, though the primary focus of this study remains on secondary school curricula.Key objectives included measuring the transferability of computational problem-solving skills, particularly in:Mathematical problem-solving: Assessing students' ability to apply computational reasoning to algebraic concepts, particularly in word problems, symbolic manipulation, and logical reasoning.Algorithmic Reasoning: Tracking students' ability to break down complex problems into structured steps, formulate solutions in pseudocode, and express them in a language-independent manner.Adaptability across paradigms: Evaluating whether students trained in abstraction transition techniques demonstrate greater flexibility when encountering new programming languages and computational models.The research employed a mixed-methods approach, combining quantitative and qualitative assessments:Pre- and post-intervention assessments: Evaluated students' ability to transition between abstraction levels, analyze algorithmic building blocks, and apply computational thinking principles to problem-solving tasks.Structured project-based learning: Emphasized scaffolded problem-solving exercises, gradually refining students’ understanding of functional composition, recursion, and modular decomposition.Diagnostic question batteries: Inspired by Simon Peyton Jones’s NeurIPS 2020 Education Challenge, these assessments identified conceptual misconceptions and provided targeted learning feedback (Wang et al., 2021).Observational and qualitative analysis: Classroom observations, concurrent math course instructor interviews, and student survey data provided insights into engagement, confidence, and cognitive challenges.An essential component of the methodology was the iterative feedback loop, which tracked student performance across multiple courses and disciplines. By following student cohorts through concurrent mathematics classes and subsequent programming courses, the study established a direct measure of the framework’s short and mid-term effectiveness. These feedback loops allowed for real-time curriculum refinement, ensuring instructional strategies remained aligned with student needs and learning trajectories. This data-driven approach provides compelling evidence supporting the Abstraction Transition Framework as a scalable and adaptable pedagogical model for introductory CS education.ResultsThe findings demonstrate that students instructed using the Abstraction Transition Framework consistently outperformed their peers trained in syntax-heavy, language-dependent curricula. Notably:Students in the abstraction transition group demonstrated superior performance in concurrent mathematics courses, particularly in pre-algebra and algebra, reinforcing the transferability of computational thinking to formal mathematical reasoning.Students trained in abstraction centered instruction performed better in subsequent computer science courses, showing increased proficiency in algorithmic reasoning, problem decomposition, and functional composition.The framework effectively reduced cognitive overload, as students could focus on conceptual problem-solving rather than struggling with verbose syntax or arbitrary language constraints.Beyond programming and mathematics, students exposed to the abstraction transition approach displayed greater adaptability when introduced to new programming paradigms. This aligns with Simon Peyton Jones’s (Computing at School UK, 2018) and Mark Guzdial’s (Guzdial, 2019) argument that computing should be treated as a foundational subject discipline, rather than merely a vocational skill set.ConclusionThis research highlights a fundamental issue in traditional CS education: the continued emphasis on industry-driven programming languages has diverted attention away from core computational principles, undermining the transferability and long-term applicability of CS instruction. By prioritizing syntax over abstraction, current curricula fail to prepare students for the intellectual demands of computational literacy as envisioned by Perlis (1964).The Abstraction Transition Framework shifts the focus from programming syntax to cognitive problem-solving, equipping students with computational literacy that extends beyond coding. By using reduced-syntax languages, the framework reduces cognitive load, allowing students to develop algorithmic intuition and transferable problem-solving skills.Tracking student performance across multiple disciplines and educational levels established a feedback loop for refining CS pedagogy, supporting scalable adoption. The findings underscore the need to decouple introductory CS instruction from popular programming technologies and instead emphasize computational literacy as a core intellectual skill.This study demonstrates that the Abstraction Transition Framework provides an effective model for cultivating adaptive problem solvers. By reorienting introductory CS education toward computational thinking and abstraction, this framework prepares students to navigate an evolving technological landscape with computational fluency. |
|---|---|
| ISBN: | 9798280709942 |
| Källa: | ProQuest Dissertations & Theses Global |