Projektdetails
Hochschule
Pädagogische Hochschule Salzburg
Sprache
Projektleitung gesamt
Siller, Klaus; Mag.
Projektleitung intern
Siller, Klaus; Mag. Dr.
Interne Projektmitarbeiter/innen
Externe Projektmitarbeiter/innen
Kooperationspartner
Laufzeit
2015 – 2016
Beschreibung
siehe englische Beschreibung
Beschreibung (engl.)
The process of validation is one of the necessary cornerstones of good testing practice. The prospective doctoral thesis tries to provide evidence for the construct validity of the Austrian ”E8 Baseline Reading Test” by identifying predictor variables influencing item difficulty in the multiple-choice tasks used in this test.
The research project is based on Weir’s socio-cognitive validation model (2005), which represents a unified approach of validation, as suggested by Messick (1987). It describes the use of language as the interaction between cognitive abilities of a test taker and social phenomena when performing language tasks. Two aspects of the model, context validity and cognitive validity, will be used for the study.
The first question to be answered is whether the cognitive processes required to complete the test tasks are appropriate, i.e. if the tasks elicit the kind of processes that test takers would use in a real-world (reading) context. This is to be done by matching the processes produced by a test against a model of reading from psychology. The model used for the research project is the reading model by Khalifa and Weir (2009), which describes different types of reading and cognitive processes.
The second aspect to be considered is context validity. The study tries to identify the key contextual features both test designers and teachers need to think about when setting tasks or selecting texts to be read and comprehended. As suggested in the literature, text comprehension (here: task completion in a test situation) is influenced by linguistic demands, content demands and/or features of the task setting.
About 170 multiple-choice reading tasks and their difficulty measures are provided by the Bundesinstitut für Bildungsforschung, Innovation und Entwicklung des österreichischen Schulwesens (BIFIE) and both theoretical and empirical activities will be used to determine how well item difficulty can be accounted for by cognitive and contextual features, thereby providing evidence favouring
the construct validity of the E8 reading test.
The results of the analyses are useful both for item writers and language teachers, e. g. by supporting the development of items of a range of difficulties and by improving the understanding of what is being assessed in the Austrian
English reading tests, therefore supporting teachers to identify underlying factors influencing reading comprehension that should be taught to improve the students’ reading skills.
The research project is based on Weir’s socio-cognitive validation model (2005), which represents a unified approach of validation, as suggested by Messick (1987). It describes the use of language as the interaction between cognitive abilities of a test taker and social phenomena when performing language tasks. Two aspects of the model, context validity and cognitive validity, will be used for the study.
The first question to be answered is whether the cognitive processes required to complete the test tasks are appropriate, i.e. if the tasks elicit the kind of processes that test takers would use in a real-world (reading) context. This is to be done by matching the processes produced by a test against a model of reading from psychology. The model used for the research project is the reading model by Khalifa and Weir (2009), which describes different types of reading and cognitive processes.
The second aspect to be considered is context validity. The study tries to identify the key contextual features both test designers and teachers need to think about when setting tasks or selecting texts to be read and comprehended. As suggested in the literature, text comprehension (here: task completion in a test situation) is influenced by linguistic demands, content demands and/or features of the task setting.
About 170 multiple-choice reading tasks and their difficulty measures are provided by the Bundesinstitut für Bildungsforschung, Innovation und Entwicklung des österreichischen Schulwesens (BIFIE) and both theoretical and empirical activities will be used to determine how well item difficulty can be accounted for by cognitive and contextual features, thereby providing evidence favouring
the construct validity of the E8 reading test.
The results of the analyses are useful both for item writers and language teachers, e. g. by supporting the development of items of a range of difficulties and by improving the understanding of what is being assessed in the Austrian
English reading tests, therefore supporting teachers to identify underlying factors influencing reading comprehension that should be taught to improve the students’ reading skills.
Bericht