Change Magazine May/June 2008

www.changemag.org

How Effective are the NSSE Benchmarks in Predicting Important Educational Outcomes?

Comment on Recent Articles

Comment on Recent Articles
Title: How Effective are the NSSE Benchmarks in Predicting Important Educational Outcomes?
Keywords:

by Ernest T. Pascarella, Tricia A. Seifert, and Charles Blaich

The widely used NSSE benchmarks, which measure student experiences and institutional practices, demonstrably predict undergraduate cognitive and personal development. (full text)

Caitlin Rogers


Comments:
Dan Bernstein
January 26, 2010 04:49:48 PM
NSSE is an outstanding vehicle for measuring students' perceptions of educational experiences, especially when the experiences are well defined. I am pleased that NSSE is a part of our national conversation on higher education. However, I think the impact of this well conducted empirical analysis will depend a bit on the criteria used to decide if one measure is really a proxy for another. The authors, noting the low power of their sample of 19 institutions, point out that several relations (10 of 35 possible) are unlikely to have occurred by chance (statistically different from zero, with a 5% false alarm rate). Those partial correlations range from .41 to .73, with six of them in the .4's. Using a crude approximation of the importance of the correlations (squaring the correlation to estimate the percentage of variance accounted for) suggests most of the NSSE benchmark scales highlighted account for less than 30% of the variation in the test scores offered as direct measures of important outcomes. While not random, the relations do not allow very tight specification of the range of performance on the measures for which they might serve as proxies. The remaining 25 relations, even if additional data added to their statistical power, do not show great predictive utility. All of them account for less than 15% of the variation in the direct measures, and 15 of them account for less than 10% of the variance in the predicted score. If my criterion for accepting something as a proxy for cognitive growth and personal development were simply "related more than by chance," then these results might be encouraging. If my criterion for proxy status is more strict, and NSSE would serve to substitute for the more expensive measures, I would want more of the benchmark scales to show that they capture large portions of the variability in the measures identified as direct indicators of my intended goals. Perhaps we should just use NSSE in the very fine ways it was intended and developed, rather than put it forward as a fair-to-middling proxy for other entities.

Post a comment:

Keywords
Calendar
Jan 2010
12
3456789
10111213141516
17181920212223
24252627282930
31

About Us Information

©2010 Taylor & Francis Group · 325 Chestut Street, Suite 800, Philadelphia, PA · 19106 · heldref@taylorandfrancis.com