Calibration of consumer and homemaking test items using the Rasch procedure

摘要

The purpose of the study was to use the Rasch procedure of item calibration to select test items in consumer and homemaking education, and to compile the items in a form usable by secondary teachers for assessing student achievement;From a pool of 230 items, five forms of a comprehensive achievement test were developed and administered to 1039 students who each responded to one form. There were 17 items deleted from the pool;The total t fit and the error impact were inspected to identify possible items for deletion. Items failing to meet the expectations of one or both criteria were submitted to additional criteria;For the Item Characteristic Curve and the department from expected Item Characteristic Curve which reflect response patterns of the ability groups, the deleted items showed irregular response patterns for the ability groups represented in the sample. Ten of the deleted items were found to be distributed above the mean ability of the sample on which the particular item was calibrated and seven items fell below the mean;Discrimination indices ranged from .02 to .49, with the expected index being 1.0. Most of the declared items had fit between, measured in standard deviations, significantly greater than the standard deviation of the group in which each item was a member;The person separability indices, equivalent to Kuder-Richardson 20 were .70, .68, .77, .73, and .68 for Forms A, B, C, D, and E, respectively in the first analysis. In the second analysis after the 17 items were deleted, they changed to .71 for Form A, .69 for Form B, .76 for Form D, .71 for Form E, and remained constant for Form C;Although some remaining items had irregularities, they were judged to fit the model. In all, 213 or 93%, of the 230 items fit the model.