Internet Learning Volume 6, Number 1, Spring 2017/Summer 2017 | Page 16

The Value of Common Definitions in Student Success Research: Setting the Stage for Adoption and Scale
grams, services, actions, interventions,
and policies. The common structure
in SSM X for categorizing interventions
makes it possible to link them to participating
students, and through the
PAR dataset explore their efficacy. By
providing common predictors to classify
interventions across institutions,
particularly effective approaches can
be identified and shared among PAR
institutions, paving the way for better
understanding, measuring, and scaling
the highest impact tools for improving
student outcomes.
Looking for Scale within the
Dataset
The combination of a commonly
defined dataset and a common
framework to measure interventions
provides powerful information
for institutions to identify potential
challenges or opportunities to improve
student outcomes. Since the dataset
and common framework is shared by
all PAR Framework members, the opportunity
to have a broader discussion
with other community members who
experience similar challenges can start
immediately. The ability to measure
across this continuum of universities
allows data scientists and researchers to
identify whether challenges are unique
or shared by members. The notion of
achieving scale across the industry is
tantalizing and can be equated to finding
generalizability in research. The
dataset allows PAR data scientists and
community researchers an opportunity
to explore important research questions.
To date, analyses have been completed
on whether there is generalizability
of positive predictors of success across
multiple institutions’ transfer students,
what the PAR Framework dataset tells
us about post-traditional students, and
whether taking online classes is detrimental
to retention and progression.
Are the Predictors of Successful
Transfer Students the Same at
Multiple Institutions?
One of the analyses performed
by PAR Framework researchers
involved identifying similar
predictors of transfer student success
across institutions. One institution,
University of Maryland University College
(UMUC), conducted a research
study that found both positive and negative
predictors for transfer students.
In order to replicate the study, an institution
or system needed to report data
on transfer students. The University of
Hawaii system dataset included students
who transferred from community
colleges to 4-year universities. The
PAR Framework data scientist looked
at 18 variables at both institutions to
address the research questions. Since
the variables were comparable because
of shared common data definitions, the
research could proceed.
There were three variables that
predicted a student’s grade point average
at the bachelor’s level that were
shared at both institutions (James,
2015) including:
• Did the student complete his or her
15