MEGS - Classical Test Theory
This shows the results of using classical test theory on the MEGS, available at http://ter.ps/megsPDF
The Kuder-Richardson coefficient purports to measure the reliability of the test. Scores can be as high as 1, and a score of 0.77 indicates "Okay, sufficient for group measurements, not individuals".
The Ferguson's Delta tests for discrimination - how well the test distinguishes between students with a lot of the construct being tested and students without much of it. A score above 0.90 is considered acceptable.
The point-biserial coefficient is a correlation between an individual question and the entire test, and should be 0.2 or higher.
The discrimination of a question is the amount by which high-scoring students outperform low-scoring students on a question, and above 0.30 is considered acceptable.
reference: Engelhardt, Paula V., and K. Harper. "An introduction to classical test theory as applied to conceptual multiple-choice tests." Getting Started in Physics Education Research (2009): 1-40.