IMPACT OF GENDER AND LINGUISTIC BACKGROUND ON ENGLISH LANGUAGE ARTS TEST: DIFFERENTIAL ITEM FUNCTIONING
IMPACT OF GENDER AND LINGUISTIC BACKGROUND ON ENGLISH LANGUAGE ARTS TEST: DIFFERENTIAL ITEM FUNCTIONING
No Thumbnail Available
Date
2019
Authors
Ahmed, Zahya
Journal Title
Journal ISSN
Volume Title
Publisher
Middle Tennessee State University
Abstract
The purpose of this study was to investigate the impact of gender and ELL status in students’ performance, and to compare the consistency of the results of the differential item functioning (DIF) determining from applying Mantel-Haenszel (MH) procedure and the DIF results determining from applying item response theory (IRT)-likelihood ratio statistics. In literature, DIF has been applied as a statistical tool to investigate bias items against subgroups in a particular population and to evaluate the fairness and validity of the educational and/or psychological assessments. Gender and different linguistic backgrounds are the most studied variables in DIF literature. A version of an English language art test was examined for evidence of DIF based on gender difference and different linguistic backgrounds (ELL vs. non-ELL). DIF analyses were implemented through six sets: (1) gender differences (male vs. female), (2) ELL status (non-ELL students vs. ELL students), (3) gender within non-ELL group (male non-ELL vs. female non-ELL), (4) gender within ELL group (male ELL vs. female ELL), male cross linguistic background groups (male non-ELL vs. male ELL), and female cross linguistic background groups (female non-ELL vs. female ELL). The sample of this study consisted of students from 7th (N = 12,658) grade from 11 states who took the standardized English Language Arts (ELA) test based on common core state standards (CCSS) at the beginning of the 2014-2015 academic year. Data were analyzed using classical test theory (CTT) and IRT to detect DIF.
The results revealed one DIF item when the MH procedure was applied. On the other hand, IRT-likelihood ratio flagged 4 items out of 34 test items. The results from both detection methods were inconsistent. Implications of these analyses were discussed in accordance with the previous findings for providing linguistically diverse students in regard to their gender with effective literacy programs to meet their academic needs.
Description
Keywords
Educational tests & measurements,
Gender studies,
Language arts,
English as a second language--ESL