Linear Logistic Test Model Analysis of a Reading Comprehension Assessment

• In Event: Psychometrics and Educational Measurement Poster Session 1

Abstract

This research examined the effects of multiple sources of text difficulty on the difficulty of items on a reading comprehension assessment by means of the linear logistic test model. The sources selected were twofold: text complexity indicators measured quantitatively and story features evaluated by human judge. The effects of the sources on item difficulty were quantified and tested with statistical significance through LLTM. Word count and narrativity from the former source and character and goal identification from the latter were significantly related to item difficulty. Despite methodological limitations, the findings will attribute to providing validity evidence of a reading assessment and generating items appropriate for the intended constructs.

Author

• Sunhi Park, K-12 Public Education