Uses Of Rasch Calibration Parameters

1187 Words Dec 3rd, 2015 5 Pages
Following the use of Rasch calibration parameters, start points as well as both reverse and discontinue rules were established according to the results of the national standardization process (Kaufman & Kaufman, 2014). Two parallel forms of the KTEA-3 were created, Forms A and B. Scores on both forms were calibrated using equating strategies, including common-person, equipercentile, combining sets of grade-appropriate items, and special linking studies. For the Reading Comprehension, Listening Comprehension, Written Expression, Oral Expression, and Word Recognition Fluency subtests, vertical scaling using item response theory (IRT) concurrent calibration was conducted individually for each form to determine start and stop points. Vertical scaling involves linking the total raw scores from different item sets (within subtests). Next, parallel form equating was completed using the equipercentile method and the entire grade norm sample. As a result, item order is not strictly based on level of difficulty. Instead, these subtests contain item sets with decision points to continue or discontinue testing.
Item analyses were completed using IRT, specifically a Rasch analysis in conjunction with the software program WINSTEPS (Linacre, 2005). Items were dropped from many of the subtests due to gender and ethnicity biases (e.g., words and terminology) as recommended by expert reviewers representing specific minority groups. Other methods for detecting and reducing bias…
Open Document