Internet Learning Volume 6, Number 2, Fall 2017/Winter 2018 | Page 19
Internet Learning Journal
that are distinguishing the student populations
persistence from one another.
This led us to compare some of the most
common data feature values shown in
Table 2.
The finding that modality-blending
students consistently have
larger credit hour loads, and enroll a
bit earlier than other groups may be
important indicators pointing us in the
direction of addressing the why question.
One hypothesis is that students
who blend modalities tend to be those
who are strategic and planful in their
academic career, where taking advantage
of online offerings helps them get
the courses they need when they need
them, or work around scheduling conflicts
due to the asynchronous nature of
most online courses. The question then
arises as to whether what we are seeing
is simply an attribute of good students
taking advantage of online course offerings
or whether the availability of
online courses could potentially help
more marginal students improve their
persistence, credit loads, or enrollment
behaviors.
Based on these findings, a clear
area for further research is to investigate
if the results shown hold for
students across a range of prediction
scores. In other words, do students who
are predicted to have a lower likelihood
of persistence also benefit from taking
a blended set of courses, or is this an
effect isolated to high-performing students?
It is possible to match on prediction
scores to build that analysis, and
that may be a fertile ground for future
exploration.
Another research area is to
perform causal impact analysis with
modality as the treatment variable to
understand what types of students benefit
from different modality options for
more personalized learning by holding
the rest of the success factors and prediction
scores identical between students
of different modalities. Future
work includes examining social psychological
factors to isolate and disambiguate
the effects of financial aid,
academic factors, and non-academic
factors on student success as a function
of modality.
Moving beyond the somewhat
simplistic findings above, it is possible
to go deeper by examining the populations
not only by course modality, but
then comparing those by persistence
scores. In other words, looking at the
differences between students taking fully
on-ground, fully-online, and blended
curriculums for the lowest quartile
of persistence predictions, then the
next quartile and so on. In pilot analyses
conducted by the data scientists at
Civitas Learning, they have confirmed
the trends detailed in this paper, but the
persistence gaps are smaller, which is to
be expected.
References
Allen, I. E., Seaman, J., Poulin, R.,
& Strout, T. T. (2016). Online report
card. Retrieved from https://www.
onlinelearningsurvey.com/reports/
onlinereportcard.pdf
Bettinger, E., & Loeb, S. (2017).
Promises and pitfalls of online education.
18