Internet Learning Volume 6, Number 1, Spring 2017/Summer 2017 | Página 12

The Value of Common Definitions in Student Success Research: Setting the Stage for Adoption and Scale
gate factors affecting the retention and
progression of undergraduate students.
The creation of such a common dataset
clearly required the creation of common
data definitions that were specific
enough to ensure reliable findings
but that could also be applied to most
undergraduate programs. The original
six participating institutions included
a community college, 4-year college,
university system, community college
system, and two for-profit universities,
which met to establish common ground
for all members of the post-secondary
community to engage in a common conversation.
All the participating members
offered both fully on-ground and
fully online classes. Participating programs
ranged from traditional semester
terms to eight- and five-week terms
with start dates happening every week.
Thus, the initial work of the collaborative
was to find ways of defining such
seemingly simple outcome variables as
retention and progression relative to a
common time frame, something with
which the Federal government continues
to struggle. Table 1 presents categories
of input variables defined.
11
Common definitions are a key
feature of the PAR dataset. Through the
original grant work, a group of researchers
identified and then openly published
a set of common data definitions
(https://community.datacookbook.
com/public/institutions/par). Because
all of the data that were and are provided
by PAR member institutions utilize
these common definitions, cross-institutional
apples to apples analyses on the
combined dataset can be performed to
better understand the factors that impact
student success generally as well as
locally.
The success of these original PAR
researchers was due, in a large part, to
their willingness to collaborate and
share data and analytic approaches in a
safe, supportive environment, a benefit
that continues today. PAR member institutions
comprise a range of the many
diverse options for post-secondary education,
including traditional, open
admission community colleges; 4-year,
traditional, selective admission, public
institutions; and nontraditional, open
admission, primarily online institutions,
both for-profit and nonprofit.
Since all of the data provided by
PAR member institutions meet the parameters
of the common definitions,
PAR researchers were able to do both
aggregated and cross-institutional comparisons
and analyses on the combined
data. Having relatively comprehensive,
detailed data for all credential-seeking
students (as opposed to a sample from
each institution) creates a more accurate
understanding of the student- and
institutional-level factors that impact
risk and success. It also makes it possible
to more effectively control for
confounding variables that might be
contributing to observed differences
between student groups.
The PAR Framework data modeling
yielded positive, negative, and
variable predictors for being retained
after 12 months. The positive predictors
included high school GPA (when available),
dual enrollment (high school/college),
prior college credits, community
college GPA, successful course comple-