several introductory IT course sequences in Estonian universities. They found that students with pre-college experience
received higher grades in their first semester than those without experience, and that they also had higher external motivation for their studies.
Prior experience can also affect students’ sense of belonging
and stress level. In a qualitative study, Tafliovich et al. found
that students believe prior experience boosts confidence and
that discussion forums can be stressful for those with no experience [ 23]. Indeed, Rogerson and Scott report that stress and
fear are regularly associated with learning to program and can
interfere with a student’s ability to succeed [ 21].
The idea of Learning Edge Momentum as applied to CS1
[ 1, 20] provides another explanation for performance differences due to prior experience. This theory suggests that rather than
disappear, performance differences are likely to persist into later courses because once students learn one concept, it becomes
easier for them to learn subsequent concepts. However, most
prior work on the effect of pre-college experience on college CS
success stops at CS1. Our study complements this prior work
by looking at courses much further into the CS major, and by
examining the role of several specific pre-college experiences
students might have.
Recent work has also studied how pre-college experiences
(among other factors) affect students’ choice to major in computer science. McGill et al. find that pre-college outreach activities have some effect on students’ choice of major in college, but
that this effect appears to be stronger for men than for women.
[ 17]. In a 2015 survey, Wang et al. found that pre-college academic opportunities, both formal and informal, were an important factor in girls’ choice to major in computing in college,
and that taking a pre-college CS course was more significant for
girls than for boys in their decision to major in computing in
college [ 26]. A natural follow-up study to the work we present
here would examine whether the results we report are consistent for both men and women.
Based on previous work and our anecdotal interactions with
students in the classroom, we postulated the following hypoth-
• H1: Students with more pre-college CS experience will do
better initially than their less experienced peers, but these
differences will disappear as course level increases.
• H2: Many types of pre-college experience will provide an
advantage over having no pre-college experience, but some
experiences will stand out as particularly beneficial on their
own regardless of other experience.
We collected data over four quarters (fall 2014 to fall 2015)
from courses at all levels in our undergraduate curriculum. We
included our two introductory programming tracks: a two-
course sequence designed for students with no prior CS expe-
rience (CS1:n=1258, CS1.5:n=543), and a one quarter course
men [ 13, 14]. Students in these groups are also less likely to take
Advanced Placement (AP) CS [ 6], in no small part because of
lack of access [ 13, 16]. Despite the many current efforts to level
the playing field in K- 12 CS education, different pre-college ex-
perience levels, whether due to access or to student choice, are
likely to persist into college for some time.
At the college level, educators have developed a number
of curricular approaches to level the playing field for students
with less pre-college experience. Several universities have introduced a separate introductory course or track specifically
to allow students with less pre-college experience to catch up
[ 4, 11, 15]. A complementary approach uses curriculum that
will not disadvantage students with no experience in introductory courses. For example, some redesigned introductory
courses begin with content that students are unlikely to have
had exposure to in high school (e.g. [ 5, 25]). Others use a contextualized computing approach such as media computation,
which has been shown to increase retention and decrease fail
rates for non-majors and other students with no pre-college
CS experience [ 7, 22, 24]. Other programs have courses that
let students choose their own focus area based on their interest [ 9].
Little work has studied the outcome in later courses for students with no pre-college experience, and there is evidence that
these students under-perform their experienced peers even in
courses that employ a host of best practices [ 2]. We wished to
explore whether students with less pre-college experience are
less likely to do well not just in introductory courses, but also
in courses later in the major? If performance differences disappear, at what point do they do so?
We examine the relationship between different pre-college
CS experiences and students’ grades and persistence in undergraduate CS classes at a large US public research university.
We find that both students’ self-assessed level of pre-college
experience and their participation in specific pre-college experiences are associated with a difference in course performance
at levels. We notice a particularly strong effect with students
who take AP CS in high school. However, we find no significant
difference in performance on low-stakes assessments of CS
knowledge between students with pre-college experience and
those with none. These findings have implications for ensuring
all students get a high-quality computer science background in
high school and on the design of our college-level courses and
Prior experience in CS (and related subjects) is widely shown
to be an important factor of success and persistence in introductory CS at the college level [ 8, 10]. One explanation
for why prior experience is helpful is that students with prior experiences often have different motivations, confidence
levels and goals. Recent work by Kori et al. [ 12] compared
the attitudes, persistence and success of students with and
without pre-college programming experience throughout