 |
|
 |
|
 |
|
 |
Validity |
 |
Reliability |
Printer-friendly page
 |
Evaluating Tests
Created by:
Dr. Sonya Lynn Dinero
Professor Emeritus, Evaluation and Measurement
Kent State University
Dr. Thomas Edward Dinero
Professor Emeritus, Evaluation and Measurement
Kent State University
Nuria M. Cuevas, Ph.D.
Acting Associate Vice President for Academic Affairs;
Director, Institutional Effectiveness and Assessment
Norfolk State University
Objectives
Evaluating any process in a meaningful way means selecting criteria before the fact and applying these criteria in an objective way. Tests aren't art or pornography: we can't evaluate them by saying, "I'll know good art when I see it." Or "Whatever offends the citizenry...." On the other hand, tests are used by fallible humans to evaluate other people using criteria which are often little understood or well defined, and so there is a bit of impressionistic influence involved. Your efforts should be to develop a test which can be used in as fair and consistent way as possible so that you can be both personally and professionally accountable to your students.
Upon completion of this module, you will be able to:
new objectives
- evaluate tests for reliability and validity
- design or redesign tests to be more reliable and more valid
old objectives
- distinguish between reliability and validity
- distinguish between breadth and depth in a test
- distinguish between internal and external test criteria
- discuss the relationship between a domain and the resulting test
- know the four types of scales and how they relate to the classroom test
- be familiar with three correlation coefficients and how they are used in tests
- understand the need for well-thought-out validity in classroom tests
- understand how to study the statistical properties of your test
- improve the reliability of your classroom test, and
- improve the validity considerations of a classroom test.