• Automated administration
– Versant tests can be performed on a computer or over the phone. Test takers can access and complete the tests from anywhere with a landline or an internet connection. Test takers are given a Test Identification Number and are instructed to listen to a recorded examiner’s voice, which is also written verbatim on the test paper or computer screen. Test takers listen to recorded item prompts read by a range of native speakers throughout the test. Large numbers of tests can be administered and scored quickly because the test is automated.
• Automated scoring technology
– Within minutes of completing the test, Versant posts the results online. By inputting their Test Identification Number on the Versant website, test administrators and test takers can view and print their Versant test results. Sentence Mastery (i.e., grammar), Vocabulary, Fluency, and Pronunciation are the four diagnostic subscores included in the Versant test score report. On a scale of 20 to 80, the overall score and subscores are reported. A large number of speech samples from native and non-native speakers are used to optimize the automated scoring method. To obtain a sufficient number of such speech samples, extensive data collecting is usually carried out. Following that, the spoken responses are transcribed in order to train an automatic speech recognition system.
– The voice recognizer, which has been designed for non-native speech, automatically processes each incoming response. The recorded signal contains the words, pauses, syllables, and phone numbers. The existence or absence of expected proper words in expected sequences, as well as the tempo, fluency, and pronunciation of those words in phrases and sentences, are used to grade the content of the response. Based on statistical models of native and non-native speakers, base measurements are then produced from segments, syllables, and words.