An objective measure of English proficiency must be SECURE.
English3, Chief Technology Officer
Ph.D. Candidate, Cybersecurity and Information Assurance<br>
English3, Managing Partner
Test security focuses on preventing test-takers from gaining an unfair advantage and provides confidence that the score report is an accurate representation of the English ability of the test-taker.
After readily identifying threats to test security with common online English tests, the test development experts at English3 became convinced that an innovative approach was needed.
Keys to Maintaining the
Security of the Interview
REAL-TIME FREE RESPONSE
Interactive video tools simulate the academic and every-day scenarios in which the test-taker will be placed. English3’s proprietary technology assesses real-time ability to produce language under the pressure of real-life situations, such as answering questions as part of a face-to-face discussion.1 This test format is ideal for assessing a test-taker’s ability to “participate in his or her program and to function on a day-to-day basis”.2
The requirement that test-takers respond in real-time eliminates many security concerns, including the ability of the test-taker to apply strategies for gaming multiple-choice, fill-in-the-blank, and parts of speech questions. Real-time response requires authentic language production.3
FACIAL RECOGNITION SOFTWARE WITH PROCTOR REVIEW
Robust facial recognition software protects test integrity by comparing the test-taker’s photo ID and pre-test setup photos to the test-taker’s videos. In addition, photos and videos are compared across our complete test-taker database.4 This review prevents an individual from taking a test multiple times under different identities. English3 is continually improving biometric technology used to ensure test-taker identity. Facial recognition results must be approved through a manual review by the proctor before Score Reports are created.
1:1 PROCTOR TO TEST-TAKER RATIO
English3 tests are proctored and rated remotely by experienced language professionals who complete a rigorous certification process. Test videos are reviewed by raters5 who are trained to spot security concerns, such as looking off camera, reading from notes, or receiving assistance from others. Video recordings allow raters to see student expressions and give transparency to the applicant’s English ability. Attempts at cheating are readily visible and documented. Multiple test video recordings are also available for review by university staff with access to the administrator dashboard.
RANDOMIZED TEST FORMS
No two tests are the same. Questions come from large question banks with variable question types. Tests are created for specific purposes and test-takers progress through targeted activities with high levels of validity. The English3 Interview begins with introductory and beginner questions before progressing through intermediate and advanced questions. Question topics include day-to-day interactions, cultural exchange, safety, professional goals, and other relevant topics.
SECURITY QUESTIONS
Security questions are unique item types with thousands of variations. These questions are not graded directly but are used as a benchmark to approximate the language ability of the test-taker compared to the graded test questions. Security questions include picture descriptions,6 scrolling text, and elicited imitation.7 Elicited imitation “consists of having learners listen to and repeat, to the best of their ability, [sentences] of varying lengths and complexities.”8
SCREEN MONITORING
Test proctors review test questions, videos, and screen activity for the duration of tests. Monitoring screen activity increases test security and gives proctors deep visibility of the testing environment.
Why Test Security Matters
CONFIDENCE
Secure English proficiency tests give university faculty and staff confidence that international visitors have the English proficiency required to participate in their program and to function on a day-to-day basis.
CONTINUITY
Secure English proficiency tests ensure that academic work continues uninterrupted by issues arising when international visitors arrive without the ability to function in an English-speaking academic environment.
COMPLIANCE
Secure English proficiency tests offer an objective measure of English proficiency—helping universities ensure the health and safety of international students and scholars, while promoting powerful academic environments.
References
Griffiths, M. E., & Graham, C. R. (2009b). The potential of asynchronous video in online education. Distance Learning, 6(2), 13–22.
“§ Sec. 62.10 Program Administration. | USCIS.” 62.10 Exchange Visitor Program Cases. | USCIS, 1 Jan. 2015, www.uscis.gov/ilink/docView/22CFR/HTML/22CFR/0-0-0-1/0-0-0-3590/0-0-0-3859.html#0-0-0-813%20or%20
Griffiths, M. E., & Graham, C. R. (2009a). Using asynchronous video in online classes: Results from a pilot study. International Journal of Instructional Technology and Distance Learning, 6(3), 65–76. http://itdl.org/Journal/Mar_09/article06.htm.
Applying Computer Vision to Images with Amazon Rekognition, Aws Lambda, and Box Skills: Amazon Web Services. Joe Norman – https://aws.amazon.com/blogs/apn/applying-computer-vision-to-images-with-amazon-rekognition-aws-lambda-and-box-skills/.
Plymouth University (2017). E-INVIGILATION OF E-ASSESSMENTS. Available at: https://pdfs.semanticscholar.org/e748/836bf5e60dcbaa68f4ef07c5a70d6ec38208.pdf.
Lavalle, Pamela I. and Briesmaster, Mark. (2017). The study of the Use of Picture Descriptions in Enhancing Communication Skills among the 8th-Grade Students—Learners of English as a Foreign Language. i.e.: inquiry in education: Vol. 9: Iss. 1, Article 4. Retrieved from: https://digitalcommons.nl.edu/ie/vol9/iss1/4.
C. R. Graham. 2006. An analysis of elicited imitation as a technique for measuring oral language proficiency. In Teaching, pages 57–67, Taipei, Taiwan. English Teachers’ Association.Yi ju Chen and Yiu nam Leung, editors, Selected Papers from the Fifteenth International Symposium on English
Kennington, C. & Graham, C. R. Elicited Imitation as an Oral Proficiency Measure with ASR Scoring, 1.2, 1604.