TY - JOUR
T1 - Differential Item Functioning Across Race and Ethnicity for the Emotional and Behavioral Screener.
AU - Lambert, Matthew C.
AU - January, Stacy-Ann A.
AU - Cress, Cynthia J.
AU - Epstein, Michael H.
AU - Cullinan, Douglas
PY - 2018/1/1
Y1 - 2018/1/1
N2 - The objective of this study was to assess the measurement invariance of items from the Emotional and Behavioral Screener (EBS) across racial and ethnic groups and assess the impact of differential item functioning (DIF) on test scores from the EBS. Participants were 4,856 first-grade students (49% female) who were diverse with regard to race and ethnicity (45% African American, 41% Hispanic/Latino, and 14% Caucasian). Classroom teachers rated each student on the EBS during the fall semester screening window. Item response theory modeling was combined with ordinal regression to investigate the presence and impact of differential item functioning across 3 race and ethnicity groups. The findings suggest that items from the EBS exhibit small to negligible levels of DIF, and the limited DIF that was present does not significantly impact overall scores. Researchers and practitioners can have confidence that scores from the EBS are relatively unaffected by test bias when measuring the emotional and behavioral risk of young students from African American, Hispanic/Latino, or Caucasian backgrounds.
AB - The objective of this study was to assess the measurement invariance of items from the Emotional and Behavioral Screener (EBS) across racial and ethnic groups and assess the impact of differential item functioning (DIF) on test scores from the EBS. Participants were 4,856 first-grade students (49% female) who were diverse with regard to race and ethnicity (45% African American, 41% Hispanic/Latino, and 14% Caucasian). Classroom teachers rated each student on the EBS during the fall semester screening window. Item response theory modeling was combined with ordinal regression to investigate the presence and impact of differential item functioning across 3 race and ethnicity groups. The findings suggest that items from the EBS exhibit small to negligible levels of DIF, and the limited DIF that was present does not significantly impact overall scores. Researchers and practitioners can have confidence that scores from the EBS are relatively unaffected by test bias when measuring the emotional and behavioral risk of young students from African American, Hispanic/Latino, or Caucasian backgrounds.
UR - https://digitalcommons.usf.edu/esf_facpub/246
UR - https://doi.org/10.1037/spq0000224
U2 - 10.1037/spq0000224
DO - 10.1037/spq0000224
M3 - Article
VL - 33
JO - School Psychology Quarterly
JF - School Psychology Quarterly
ER -