Next
Steps
The NSSE project staff is continuing to examine the psychometric
properties of the instrument as a whole and on the five
benchmarks of effective educational practice featured
in NSSE reports. We are also working with some partner
institutions and organizations on these some of these
efforts. For example:
- Peter Ewell of the National Center on Higher Education
Management Systems is doing a special analysis of NSSE
results from the universities in the South Dakota system
as a cross validation study, comparing NSSE data with
direct outcome measures from students’ ACT and CAAP
scores.
- NSSE is also examining information collected by the
University of South Carolina National Resource Center
for First Year Programs and Students in Transition to
gauge whether students at institutions that have “model”
first year experience programs are more engaged than
their peers elsewhere.
- Selected NSSE questions will be included on the collegiate
oversample as part of the National Assessment of Adult
Learning that will be administered during 2003.
- Finally, NSSE was co-administered with experimental
outcomes assessment instrumentation that was field tested
during spring 2002 by a CAE-RAND research team in a
study funded by several foundations. (Benjamin &
Hersh, 2002).
|
We will update this psychometric
report when the results of these analyses become available.
Conclusion
In general, the psychometric properties of the NSSE are
very good, as the vast majority of items equal or exceed
recommended measurement levels. Those items that are not
in the normal range on certain indicators, such as kurtosis
and skewness, are due to the nature of the student experience,
not because of psychometric shortcomings of the instrument.
The face and construct validity of the survey are strong.
This is not surprising because national assessment experts
designed the instrument and most of the items have been
used for years in established college student assessment
programs. In addition, we made improvements to individual
items and the overall instrument based on what was learned
from focus groups, cognitive testing, and the psychometric
analyses on the results from the spring 1999 field test,
the inaugural national administration in spring 2000,
and the spring 2001 administration. The results seem to
be relatively stable from one year to the next and non-respondents
are generally comparable respondents in many ways, though
contrary to popular belief non-respondents appear to be
slightly more engaged than respondents.
|
