Identifying Differential Item Functioning (DIF) in Context of School Violence
Violence and bullying in schools are a concern for everyone. Violence in school remains a sensitive subject, particularly because of the harmful consequences it may have on the personal, social and academic development of young people, whether they are victims, perpetrators or witnesses (Blaya, 2006, Juvonen, Wang, & Espinoza, 2011). The consequences for students, parents, school or society are far too important to ignore this problem. In Quebec SEVEQ’s researchers launched a project to obtain an overview of this violence experienced at school. To fully reaching their goals, outcomes from assessments used require multiple sources of evidence of validity to be interpreted meaningfully (Downing, 2003). Messick (1989, 1995) propose five distinct sources of validity: content, response process, internal structure, relation to other variables and consequences. One of the threats to evidence of internal structure is item bias. Item bias occurs when two examinees from different groups do not have the same probability to report a violent behavior while controlling for ability (Angoff, 1993; Clauser & Mazor, 1998). This item bias is assessed by Differential Item Functioning (DIF) procedures. DIF can be caused by different characteristics that are not related to the purpose of the assessment: item or context of the evaluation. Several procedures have been developed to identify DIF but none of them is universally accepted by the research community (Clauser & Mazor, 1998; Holland & Thayer, 1986).
Published in: World Congress on Education (WCE-2018)
- Date of Conference: 15-18 July 2018
- DOI: 10.2053/WCE.2018.0051
- ISBN: 978-1-908320-91-9
- Conference Location: Dublin, Ireland