T.E.A. – the Test of English for Aviation – is designed to assess the plain English language proficiency of pilots and air traffic controllers in an aviation context. It is designed to effectively elicit language
Research has been carried out to measure the validity and effectiveness of TEA. A summary of this research follows:
assessable by the ICAO band descriptors and is conducted in form of a one-to-one interview between candidate and examiner. As of May 2018, approximately 48,000 TEA tests have been delivered by 294 examiners in 139 test centres around the world. TEA has been developed in accordance with the ICAO guidelines outlined in Document 9835 (2010). TEA Ltd has conducted research and produced a variety of reports pertaining to test specifications, language competences, task design, item content, assessment and administration: taken together, they present a strong argument for test validity. Test Design and Construct A designated report has been produced to describe the test purpose and the principles underlying TEA’s development. Each aspect of construction is correlated to ICAO’s language proficiency requirements. The test measures plain language proficiency in an aviation context only, and not intelligence, logical thinking, operational knowledge or any other construct which would unfairly affect the assessment. TEA incorporates both direct and semi-direct testing methods to integrate a CD-delivered listening section, with individual test items, into the test. Candidates respond to all items orally and there is no discrete-point testing. The final score for each test-taker is the lowest of the scores on each of the six ICAO language proficiency skills. Test Validity and Reliability TEA Ltd has conducted a number of studies which provide a picture of the test’s validity and reliability. A designated report illustrating the development process is available, as are reports describing item writing, item trialling and item revision. Reports into the language abilities measured, and the functions and topic domains elicited by T.E.A. show the range to be wide and appropriate. Expert judges confirm that key linguistic competences – grammatical, lexical and phonological – are engaged in every task. A study of the performance of different test versions post- operalisation demonstrates that, in each part of the test, different versions perform equally giving neither an advantage nor disadvantage to candidates. Rating Reliability T.E.A. Examiners are selected based on language and operational experience. Following training (rating and interlocution) and certification, their tests are then monitored by Senior Examiners. A third rater is consulted in the case of divergent scores. In high-stakes language testing, examiner reliability of 0.9 is considered a minimum required level of reliability. In 2013, there was a high level of agreement between the Senior Examiners on Overall Score rating ranging from 0.97 to 1.00. In 2013, approximately 1 in 4 tests were double-marked and in 98% of cases there was agreement in the overall score between fully-trained Examiners and Senior Examiners. Examiners complete annual standardisation and must re- certify every 2 years. Test Administration and Security Since security is a major issue in the high-stakes environment of aviation testing, TEA has a number of security features to reduce these risks including: numerous test-day identification checks multiple test versions centralised certificate printing (a test-day photograph appears on the certificate) certificate verification website a secure database (all candidate and test data is stored within the TEA Database, an online MySQL secure database accessed by TEA Centre Administrators)