Differential Item Functioning in the aspects of Country and Gender Differences in the Computer and Information Literacy Test: Based on ICILS 2013

Zusammenfassung. ICT Literacy legt eine performanzbasierte Erfassung mit simulierten und interaktiven Testaufgaben nahe. Der vorliegende Beitrag untersucht, ob mit Multiple-Choice (MC)-Aufgaben ein vergleichbares Konstrukt wie mit Simulationsaufgaben erfasst wird. Hierfür wurden die Testergebnisse zweier Instrumente aus aktuellen Large-Scale-Studien gegenübergestellt, die an N = 2 075 Jugendlichen erhoben wurden: der auf MC-Aufgaben basierende ICT Literacy-Test für Klasse 9 des Nationalen Bildungspanels (National Educational Panel Study, NEPS) und der simulationsbasierte Kompetenztest der internationalen Schulleistungsstudie ICILS 2013 (International Computer and Information Literacy Study). Die Analysen unterstützen die Gültigkeit der Konstruktinterpretation des MC-basierten Tests in NEPS. Im Sinne der konvergenten Evidenz korrelieren die MC-Aufgaben substanziell mit den computer- und simulationsbasierten Aufgaben in ICILS 2013 (.68 ≤ r ≤ .90). Weiterhin ergeben sich positive und für beide Tests vergleichbar hohe Korrelationen mit ICT-bezogenen Schülermerkmalen (z. B. Selbstwirksamkeit). Weiterführende Analysen zum Zusammenhang mit allgemeinen kognitiven Fähigkeiten zeigen zudem, dass ICT Literacy und kognitive Grundfähigkeiten distinkte Faktoren repräsentieren.

Download Full-text

Proficiency level and Gender Difference in Computer and Information Literacy

Indian Journal of Science and Technology ◽

10.17485/ijst/2016/v9i24/96114 ◽

2016 ◽

Vol 9 (24) ◽

Author(s):

JiHyun Park ◽

Soojin Kim ◽

Eunkyoung Lee

Keyword(s):

Gender Difference ◽

Information Literacy ◽

Proficiency Level ◽

And Gender ◽

Computer And Information Literacy

Download Full-text

Cross-Cultural and Gender Differences in ADHD Among Young Adults

Journal of Attention Disorders ◽

10.1177/1087054715611748 ◽

2015 ◽

Vol 23 (1) ◽

pp. 22-31 ◽

Cited By ~ 1

Author(s):

Juana Gómez-Benito ◽

Fons J. R. Van de Vijver ◽

Nekane Balluerka ◽

Linda Caterino

Keyword(s):

Gender Differences ◽

Young Adults ◽

European American ◽

Hispanic American ◽

Cultural Groups ◽

American Group ◽

Item Functioning ◽

Structural Equivalence ◽

And Gender ◽

Phi Coefficient

Objective: This study explored the effect of cultural and gender differences in ADHD among Spanish, African American, Hispanic American, and European American young adults. Method: Structural equivalence between the four groups was examined by Tucker’s phi coefficient. A MANCOVA was carried out with cultural groups and gender as factors and age as covariate. Results: Structural equivalence was observed across all groups, and no differential item functioning was found. No significant effect was found for gender, although, with the exception of the Hispanic group, males scored higher than females. Furthermore, small, though significant, cultural differences were found. The lowest levels of ADHD were observed in the European American group and the highest in the Hispanic American group. ADHD symptoms, notably inattention, showed some decline with age. Conclusion: Findings extend existing data and suggest a relationship between culture and the development of ADHD, which might be mediated by parenting style.

Download Full-text

Influences on the Mantel-Haenszel Chi-Square in Detection of Differential Item Functioning under Rasch Conditions

Perceptual and Motor Skills ◽

10.2466/pms.1995.80.3c.1071 ◽

1995 ◽

Vol 80 (3_suppl) ◽

pp. 1071-1074 ◽

Cited By ~ 1

Author(s):

Thomas Uttaro

Keyword(s):

Sample Size ◽

Differential Item Functioning ◽

Item Bias ◽

Reference Groups ◽

Chi Square ◽

Item Functioning ◽

Size Number ◽

Gender Based ◽

And Gender ◽

Present Simulation

The Mantel-Haenszel chi-square (χ2MH) is widely used to detect differential item functioning (item bias) between ethnic and gender-based subgroups on educational and psychological tests. The empirical behavior of χ2MH has been incompletely understood; previous research is inconclusive. The present simulation study explored the effects of sample size, number of items, and trait distributions on the power of χ2MH to detect modeled differential item functioning. A significant effect was obtained for sample size with unacceptably low power for 250 subjects each in the focal and reference groups. The discussion supports the 1990 recommendations of Swaminathan and Rogers, opposes the 1993 view of Zieky that a sample size of 250 for each group is adequate.

Download Full-text