A Comparison of Results from Equating O-NET Scores with Common Items and with a Three-Parameter IRT Model

Abstract

The purpose of this research was to compare the O-NET scores from 2007 and 2008 equated by using two methods: common-item equating, and a 3PL IRT model. The findings can be summarized as follows:

1. When using common items to equate O-NET scores, no significant differences were found between test means from 2007 and 2008 (p>.05).

2. When a 3PL IRT model was used for equating, test means were found to differ (p<.01).

3. The O-NET scores equated with the common-item method had a high positive relationship (p<.01)

with the original O-NET scores in every subject over both years.

4. The O-NET scores equated with 3PL IRT also had a high positive relationship (p<.01) with theoriginal O-NET scores in every subject over both years.

5. When comparing the correlation coefficients found in (3) with those found in (4), no differences were noted (p>.05)