Factors Affecting the Sample Invariant Properties of Linear and Curvilinear Observed- and True-Score Equating Procedures IRT
- Author(s):
- Stocking, Martha L.; Eignor, Daniel R.; Cook, Linda L.
- Publication Year:
- 1988
- Report Number:
- RR-88-41
- Source:
- ETS Research Report
- Document Type:
- Report
- Page Count:
- 72
- Subject/Key Words:
- Equated Scores, Item Response Theory (IRT), Sampling, Scores, Statistical Bias, True Scores
Abstract
A sequence of simulations was carried out to aid in the diagnosis and interpretation of equating differences found between random and matched (nonrandom) samples for four commonly used equating procedures: Tucker linear observed-score equating; Levine equally reliable linear observed-score equating; Equipercentile curvilinear observed-score equating; and IRT curvilinear true-score equating. The results support the prediction based on theoretical grounds that observed-score equating methods are more affected by sample variation than are true-score equating methods. These results further suggest that matching equating samples on the basis of fallible measures of ability may not be advisable for any conventional equating method except the Tucker method. In addition, the results support a particular hypothesis about IRT equating, suggesting that the use of matched samples cannot be recommended for this equating method either. (72pp.)
Read More
- Request Copy (specify title and report number, if any)
- http://dx.doi.org/10.1002/j.2330-8516.1988.tb00297.x