In a very practical addition to the Performance Test Development Series, Wallace Judd, Ph.D., shows how to calculate p-values for various testing situations.
The p-value, in the context of examinations, is an indicator of how difficult a test item is. For simple correct-incorrect items, it is the percentage of test takers who answered the question or completed the task correctly.
In the paper, “Performance Test p-value Calculations,” Dr. Judd explains how to extend this concept to items that may have multiple points and to situations where not every candidate is shown every item. An accompanying Excel spreadsheet provides examples.
The papers in this series are intended to be a resource for test developers, illustrating best practices in the performance-testing industry. They provide advice on performance testing topics from initial concept to test delivery.
These topics are presented in language that does not assume study in psychometrics.
The mathematics of any topics are expressed in abbreviated terms instead of the Greek alphabet, and all computations are illustrated in Excel examples.
The papers are published by Authentic Testing, leaders in the field of performance testing, in the hope that they will expand access to performance testing to a wider audience of practitioners, inspiring them to explore the possibilities inherent in performance testing.