Henry Maier

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2022-128

May 14, 2022

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-128.pdf

At UC Berkeley, unprecedented growth and demand for Computer Science education has resulted in difficulties assessing how students navigate introductory Computer Science courses. Instructors are left using intuition, low quality survey data and information relayed by large course staffs to determine pain points and identify students at risk of underperforming. In response, previous work developed Snaps, a tool for CS61B (CS 2) at UC Berkeley that captures fine-grained snapshots of student work as they complete assignments. We build simple metrics for assignment completion time and working habits using the Snaps dataset from CS61B Spring 2021. Using these metrics, we investigate individual associations between these metrics and performance, as measured by scores on exams and final course grade. We find rough associations that confirm intuitions relating working time and working habits to performance. Specifically, we find that students who spend longer and start later on assignments tend to perform worse in the course. Additionally, we develop and evaluate models that use our metrics, as well as grade data from CS61A (CS 1) at UC Berkeley as features and find an early prediction model based on data from the first 2 weeks of the course for exam points R^2 = 0.656) and final course grade (R^2 = 0.585) with limited accuracy.

Advisors: Joshua Hug

---

BibTeX citation:

@mastersthesis{Maier:EECS-2022-128,
    Author= {Maier, Henry},
    Title= {Investigating Intuitions and Predicting Success using Fine Grained Student Code Snapshot Data},
    School= {EECS Department, University of California, Berkeley},
    Year= {2022},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-128.html},
    Number= {UCB/EECS-2022-128},
    Abstract= {At UC Berkeley, unprecedented growth and demand for Computer Science education has resulted in difficulties assessing how students navigate introductory Computer Science courses. Instructors are left using intuition, low quality survey data and information relayed by large course staffs to determine pain points and identify students at risk of underperforming. In response, previous work developed Snaps, a tool for CS61B (CS 2) at UC Berkeley that captures fine-grained snapshots of student work as they complete assignments. We build simple metrics for assignment completion time and working habits using the Snaps dataset from CS61B Spring 2021. Using these metrics, we investigate individual associations between these metrics and performance, as measured by scores on exams and final course grade. We find rough associations that confirm intuitions relating working time and working habits to performance. Specifically, we find that students who spend longer and start later on assignments tend to perform worse in the course. Additionally, we develop and evaluate models that use our metrics, as well as grade data from CS61A (CS 1) at UC Berkeley as features and find an early prediction model based on data from the first 2 weeks of the course for exam points R^2 = 0.656) and final course grade (R^2 = 0.585) with limited accuracy.},
}

EndNote citation:

%0 Thesis
%A Maier, Henry 
%T Investigating Intuitions and Predicting Success using Fine Grained Student Code Snapshot Data
%I EECS Department, University of California, Berkeley
%D 2022
%8 May 14
%@ UCB/EECS-2022-128
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-128.html
%F Maier:EECS-2022-128