How Reliable Are Our Assessment Data?: A Comparison of the Reliability of Data Produced in Graded and Un-Graded Conditions

Authors: Anthony R. Napoli¹; Lanette A. Raymond²

Source: Research in Higher Education, Volume 45, Number 8, December 2004 , pp. 921-929(9)

Publisher: Springer

Buy & download fulltext article:

Price: $47.00 plus tax (Refund Policy)

Abstract:

Motivating students to perform well on assessment tests is difficult when students know the results have no academic consequence. The present study evaluates the influence of assessment context (graded vs. non-graded) on the reliability of an assessment measure. Results indicate the graded condition produces higher reliability (r= .71) than the non-graded condition (r = .29), which leads to unacceptably low reliability. Moreover, the graded condition produces significantly higher scores (M = 64%), than the non-graded condition (M = 43%). Only students in the graded condition (41%) obtained passing scores of 70% or above.

Keywords: assessment; higher education; reliability; authentic assessment; embedded assessment; student motivation

Document Type: Research article

DOI: http://dx.doi.org/10.1007/s11162-004-5954-y

Affiliations: 1: Office of Institutional Research and Assessment Suffolk County Community College, 533 College Road, Selden, NY 11784, USA 2: Office of Institutional Research and Assessment Suffolk County Community College, USA

Publication date: 2004-12-01