A Simulation-Monte Carlo Study of Item Difficulty Measures Delta and D.6
- Author(s):
- Tucker, Ledyard R
- Publication Year:
- 1981
- Report Number:
- RR-81-06
- Source:
- ETS Research Report
- Document Type:
- Report
- Page Count:
- 29
- Subject/Key Words:
- Item Analysis, Item Response Theory, Monte Carlo Methods, Psychometrics, Test Reliability
Abstract
An investigation was conducted by simulation-Monte Carlo techniques of item deltas and a possible new item difficulty measure, D.6, derived from the relation of proportion correct on an item to observed score on a test. Measure D.6 was the standardized test score for which the probability of a correct response was .6, half way between a chance level of .2 and a probability of unity. A major problem involves replacement of a trait score in IRT with the less than perfectly reliable test score. A second problem concerns the unidimensional assumption. Results indicate for item deltas that differing discriminating powers of items should be taken into account in converting deltas from an experimental sample to a reference population. With respect to the D.6 measure, in addition to some computational problems which were encountered effects of control test reliability and homogeneity were observed. (29pp.)
Read More
- Request Copy (specify title and report number, if any)
- http://dx.doi.org/10.1002/j.2333-8504.1981.tb01239.x