16 September 2016 in The Journal of Educational Research
Although dual language learners (DLLs) are linguistically, culturally, and socially diverse, researchers usually study them in aggregate and compare them to non-DLLs. The authors' purpose was to identify subgroups of preschool DLLs using latent class analysis. There were 7,361 DLLs and 69,457 non-DLLs. Results revealed three distinct classes. Children in class 1 (emergent bilingual children) tended to speak mostly their native language but some English at home and in the classroom. Children in class 2 (bilingual children) tended to speak English only in the classroom, but spoke both languages at home. Children in class 3 (heritage language speakers) tended to speak only their native language at home and in the classroom. Different demographic profiles and levels of development and learning were observed across classes. The DLL subgroups and their profiles provide important information about how educational programs may be tailored to meet the diverse needs of young DLLs.
30 April 2013 in Early Education and Development
Research Findings: This study examined the measurement equivalence of the Teaching Strategies GOLD® assessment system across subgroups of children based on their primary language and disability status. This study is based on teacher-collected assessment data for 3-, 4-, and 5-year-old children for the fall of 2010, winter of 2010, and spring of 2011 checkpoint periods. Differential item functioning (DIF) analysis was conducted to determine whether items functioned differentially (a) between children with and without disabilities, (b) between English language learners (ELLs) and non-ELLs, and (c) between Spanish-speaking ELLs and non-ELLs. Results showed that the majority of items displayed little or no DIF. One item pertaining to the use of conventional grammar was consistently identified as having DIF and requires further review. Practice or Policy: Overall the findings suggest that the Teaching Strategies GOLD® is a viable observation-based, authentic assessment measure and is equally valid and reliable for children with disabilities and for those whose home language is not English. Implications for teacher education and training are discussed.
15 July 2010 in Assessment
This study examined the psychometric properties of the Schutte Self-Report Emotional Intelligence (SSREI) scale in a sample of international students studying in the U.S. universities using Rasch analysis. The results indicated that the original five-category rating structure may not function effectively for the international student sample. The results also revealed that although most of the items in the SSREI scale contributed to a single underlying construct, reverse-keyed items appeared to be problematic. Examination of the person—item map suggested that the items may be inappropriately targeted for this sample. In sum, the original version of the SSREI scale needs further improvement to adequately measure emotional intelligence.
28 March 2009 in Applied Measurement in Education
This study examined the extent to which the underlying factor structure of the 2005 South Carolina Palmetto Achievement Challenge Tests (PACT) in science for grades 3, 4, and 5 was equivalent for students who were administered the test in a regular (standard) or accommodated form. Three accommodation groups were of interest: students who received any testing accommodations, students who received any testing accommodations excluding those for whom setting was the only accommodation provided, and students who received the oral administration accommodation. Each accommodation group was compared to a matched sample of the regular group. It was found that the science test for a given grade level is appropriately described by the factor model for regular and accommodation groups. Results indicated that the factor structure, factor loadings, and error variances of the science test were similar across the regular and accommodation groups.