Paving a Path to More Inclusive Computing
but also one of research into understanding the implications
of different models for addressing the capacity challenge. We
need to study the impact of different responses to the capacity
crisis to better inform school administrators and policy makers
about approaches that can successfully scale opportunities in
computing education without sacrificing educational quality or
student diversity—otherwise, we stand to lose both.
COMPUTING EDUCATION RESEARCH AS A
RESEARCH AREA WITHIN CS
Computing education research (CER) faces an identity crisis
in much of higher education. Faculty in many computing departments believe that CER, as educationally-focused research,
belongs in a school of education. Conversely, many schools of
education, focusing on K- 12 education, education policy, or
more “established” subjects such as mathematics education, believe that CER, with a focus on computing science, belongs in
a computing science department. Practically, this means that
very few schools have CER faculty, much less any programs in
that area, giving rise to the question: how can computing education research be more broadly accepted as a research area
within computer science, and how can CER results be more
fully utilized in practice?
In many ways, the story arc for CER matches that of Hu-man-Computer Interaction (HCI) with a 25 year lag, despite
the fact that SIGCSE predates SIGCHI by more than a decade.
Indeed, in the early 1980’s much work in the HCI community
was being done by social scientists, psychologists, and human
factors researchers. Many computing departments at the time
did not see this work as “core” computer science and, as a result,
it was difficult to justify hiring HCI faculty within a CS department. Fast forward to today and HCI is a well-established area
in many CS departments and included as a “core” knowledge
area in the CS2013 curricular recommendations [ 1].
In a similar way, we must promote CER to become a more
accepted area of research within computing departments. To
this end, CRA released a white paper aptly titled “The Importance of Computing Education Research” [ 3] outlining various
approaches to promoting CER. Generalizing one of the report’s
recommendations (originally: “Apply CER to improve departmental courses”), we would argue that for CER to become more
accepted as a research area within CS, it needs to not only define common problems that computing educators would care
about (e.g., assessing modalities for teaching), but also define
benchmarks that can consistently and objectively be measured
to make studies more comparable and show progress in the
field. The use of benchmarks, while not to be considered the
only measure of progress in a field, have been used to help show
progress in areas such as machine learning, computer vision,
information retrieval, and systems for years. CER could certainly benefit from the same.
CER needs to extend beyond the notion of an idea working
in a researcher’s own classroom to establishing practices that
ing faculty in no way matches the growth in the number of
undergraduate CS majors.”
With this backdrop, a significant question for our commu-
nity is: how can we successfully meet the capacity challenges
for computing education? Admirably, the National Academies’
report not only describes this problem, but also provides a set
of recommendations aimed at trying to address this issue. We
do not repeat their full set of recommendations here (the inter-
ested reader is strongly encouraged to read the original report).
Rather, we highlight one of the more significant issues raised
in the report, which is that the seemingly simple approach of
limiting enrollment in computing programs to deal with the
capacity challenge may lead to the unintended consequence of
limiting student diversity. This is a critical factor for programs
to consider, both for purposes of equity as well as having a di-
versity of viewpoints in the downstream workforce.
Somewhat unintuitively (at least until one examines the statistical dynamics more closely), we found at Stanford University
that as the total number of students majoring in CS increased,
the percentage of women in the program also increased (Figure
1). This indicates that overall enrollment growth caused an acceleration in the number of women relative to men choosing to
major in CS. Such accelerations may be due to a number of factors, including an increased sense of community and belonging
when there are greater numbers of women in the program. This
larger community of women results in even more women feeling a sense of common identity in pursuing CS.
Limits on enrollment typically also take the form of minimum GPA requirements in introductory computing courses or
other early performance hurdles to allow for admission into a
computing major. Such barriers tend to favor students with prior experience in computing, reinforcing the existing dominant
demographic of students and thus potentially limiting student
diversity in these programs.
Ultimately, to maintain a commitment to making computing
education broadly accessible, we must answer the question of
how we scale our educational offerings to make them available
to all students who are interested. This is a question not only
about resources—of whose limitations we are painfully aware—
Figure 1: Total number of CS major declarations, including men
and women (bar graph, left axis) and the percentage of female CS
declarations (line graph, right axis) at Stanford University from academic
year 1996/97 to 2016/17.