What is it about?
The traditional, conditional standard errors (CSEMs) and those from IRT are compared with each other and with conditional reliability coefficients. It is shown that the IRT CSEMs and conditional reliability coefficients can identify where a test measures with or without precision. The traditional CSEMs can not.
Featured Image
Why is it important?
This inquiry provides evidence that determining what the precision of measurement is for subgroups, or at various cut scores, cannot be determined with traditional CSEMs. It is also shown that building tests to have equal, CSEMs across a score scale does not insure that measurement precision is equal across the scale.
Perspectives
It was interesting for the author to find that the three, indices of measurement precision--traditional CSEMs, IRT CSEMs and conditional reliability coefficients (and, incidentally, the information function of IRT)) are all functions of the same, underlying constructs--conditional error variance and conditional true variance
Alan Nicewander
Read the Original
This page is a summary of: Conditional Precision of Measurement for Test Scores: Are Conditional Standard Errors Sufficient?, Educational and Psychological Measurement, February 2018, SAGE Publications,
DOI: 10.1177/0013164418758373.
You can read the full text:
Contributors
The following have contributed to this page
