FINDING COUPLING MISCONCEPTIONS IN UNDERGRADUATE COMPUTER SCIENCE STUDENTS’ CODE AND SOFTWARE DESIGN

Abstract

Many studies have been devoted to software design, the concept of abstraction and how
students, and novice software developers, have handled these concepts. Furthermore,
there are studies that have been dedicated to coupling and cohesion, as well as, research
in regards to reviewing these studies. Which showed that understanding cohesion is not so
much a problem for computer science students and novice software developers. However,
for coupling that seems to be a different story. A major portion of the studies have mostly
been focusing on the quantitative aspect of coupling, while the qualitative aspects have not
(or hardly) been the focus. Therefore, within our own research we focused on such missing
qualitative aspects of coupling, and narrowed down on misconceptions. Misconceptions
that students have held in regards to coupling and its application. In particular undergraduate computer science students, studying at the Open Universiteit.
We were provided with the possibility to analyse the submissions the students created for
the course "Practicum Onderwerpen en Implementeren" (in short, POI), where they designed
and created the code for a vaccination computer application. With that we primarly
focused on the software design and implementation tasks within the assignments of the
course. By studying the work of Fregnan et al. [2019], Eder and Schrefl [1995], and Passier
[2021] we were able to create an analytical framework, which we utilized to detect, and
annotate, coupling types in the student submissions. Additionally, we proposed 2 propositions,
with one of them where we assumed that the domain models, created by the students,
were direct copies of the data models. With the other proposition we assumed that
the there would be rather significant differences between the software designs and the implementations (code) of these designs, created by the students.
Through our analysis we were able to make observations in regards to detecting coupling
patterns, to validate the propositions, and point out 4 coupling misconceptions. Furthermore,
we were able to create an advise for the teachers of the POI course. However, we also
encountered limitations and provided new research possibilities from them.

Date of Award	21 Aug 2023
Original language	English
Supervisor	Ebrahim Rahimi (Examiner), Sylvia Stuurman (Co-assessor) & Harrie Passier (Co-assessor)