| Summary: | 碩士 === 國立屏東教育大學 === 教育心理與輔導學系 === 98 === The Development of a Reading Comprehension Test for 4th and 5th Grade Students Using NAEP Reading Framework
Abstract
The major purposes of this study were to develop a reading comprehension test for 4th and 5th grade students, and to establish reliability and validity evidence of the test. The test was developed based on NAEP 2009 Reading Assessment specifications and includes five passages (three literary and two informational passages) separated in two booklets. Participants included were 1034 4th and 5th grade students from Kaohsiung and Pingtung Counties, Taiwan.
Rasch objective measurement model was chosen to examine the psychometric features (i.e., reliability and validity) of the test. Evidences for reliability includes person and item separation reliability indices, test-retest reliability (two weeks interval), and rater reliability; evidences for validity includes content-related, construct-related, and criterion-related validity evidence.
The main results are as follows:
1. Test results of the Reading Comprehension Test are reliable with regard to person separation, item separation, two-week test-retest consistency, and rater consistency.
2. Test results are valid in terms of:
(1) content-related validity
The relevance and representativeness of the test content was examined by content experts through judging the match between test items and test specifications based on NAEP frameworks.
(2) construct-related validity
The data of the Reading Comprehension Test fit unidimensional Rasch model; items exhibited no DIF (i.e., item difficulties are invariant across different grade and gender groups); and the three cognitive levels of the test showed hierarchical order (from the lowest to the highest levels are: locate/recall, integrate/interpret, and critique/evaluate) for the literary text but not for the informational text (factors that might affect the difficulty of the items are discussed analyzed and discussed).
(3) criterion-related validity
The Reading Comprehension Test demonstrates convergent correlation with other reading comprehension tests and discriminate correlations with other non- reading comprehension measures. The results of subgroup analyses showed that 5th grade students outperformed 4th grade students, girls performed better than boys, and students with higher verbal ability scored higher than those with lower verbal ability.
|
|---|
