Summary
Objectives:
The ExaMe system for the evaluation of targeted knowledge has been in development
since 1998. The new features of the ExaMe system are introduced in this paper. Especially,
the new three-layer architecture is described.
Besides the system itself, the properties of fixed testsin the ExaMe system are studied. In special detail, the reliabilityof the fixed tests is discussed. The theory background is explained and some limitations
of the reliability are pointed out.
Methods for Estimation of the Reliability:
Three characteristics used for estimation of reliability of educational tests are
discussed: Cronbach’s alpha, standardized item alphaand split half coefficient. The relation between these characteristics and reliability and between characteristics
themselves is investigated. In more detail, the properties of Cronbach’s alpha, the
characteristics mostly used for the estimation of reliability, are discussed. A confidence
interval is introduced for the characteristics.
Results and Conclusions:
Since 2000, the serviceability of the ExaMe evaluation system as the supporting evaluation
tool has been repeatedly shown at the courses of Ph.D. studies in biomedical informatics
at Charles University in Prague. The ExaMe system also opens new possibilities for
self-evaluation and distance learning, especially when connected with electronic books
on the Internet.
The estimation of reliability of tests contains some limitations. Keeping them in
mind, we can still get some information about the quality of certain educational tests.
Therefore, the estimation of reliability of the fixed tests is implemented in the
ExaMe system.
Keywords
Education - evaluation - Internet - reliability - bioinformatics
