The purpose of this study is the validation of the GA Classic IESOL examination at CEFR level B2 (Reading, Listening, Writing and Speaking), as a criterion-referenced achievement test, using the multi-trait, multi-method approach. The data analysed at the moment pertain to the Listening unit and have been studied to examine the content validity, reliability of the scores and unidimensionality of the construct. The Listening unit consists of three tasks (22 items) in four different versions (D-E-F-G). Four separate statistical analyses were performed and all revealed excellent values of internal consistency. As for unidimensionality of the construct, excellent values were revealed for versions E, F and G and acceptable for version D. Similar analyses on the reading test are planned, with the aim of reaching a multitrait-multimethod analysis of the examination as a whole. Further steps of the study will be the IRT equating procedure of the four versions.
Validation of a High-Stakes Test: GA IESOL Listening Unit / Mitrovic, Snezana; Botta, Emanuela; Castellana, Giusi. - (2022), pp. 1001-1008. (Intervento presentato al convegno HEAd’22 - 8th International Conference on Higher Education Advances tenutosi a València) [10.4995/HEAd22.2022.14514].
Validation of a High-Stakes Test: GA IESOL Listening Unit
Snezana Mitrovic
Primo
Writing – Original Draft Preparation
;Emanuela BottaSecondo
Writing – Original Draft Preparation
;
2022
Abstract
The purpose of this study is the validation of the GA Classic IESOL examination at CEFR level B2 (Reading, Listening, Writing and Speaking), as a criterion-referenced achievement test, using the multi-trait, multi-method approach. The data analysed at the moment pertain to the Listening unit and have been studied to examine the content validity, reliability of the scores and unidimensionality of the construct. The Listening unit consists of three tasks (22 items) in four different versions (D-E-F-G). Four separate statistical analyses were performed and all revealed excellent values of internal consistency. As for unidimensionality of the construct, excellent values were revealed for versions E, F and G and acceptable for version D. Similar analyses on the reading test are planned, with the aim of reaching a multitrait-multimethod analysis of the examination as a whole. Further steps of the study will be the IRT equating procedure of the four versions.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.