View Related Sites
Cambridge Assessment Admissions Testing
Administering our tests
Our dedicated research team conducts routine operational analyses and a wide range of validation work to ensure our tests make a useful and fair contribution to the admissions process. This continuous programme forms an important part of our test development and review cycle.
Candidate information (such as gender, age, school, and requests for access arrangements) is collected during the registration process. This information is matched with the UCAS data received by universities, which enables us to monitor the characteristics of candidates – forming the basis of our research into test fairness.
In addition to analyses of item quality, potential bias is investigated by analysing the performance of different candidate groups on each test item. This is known as differential item functioning (DIF) analysis and is carried out after each test session.
Areas of validation
We carry out a wide range of validation research, including work to inform construct validity (cognitive, context and scoring validity), criterion-related validity (concurrent and predictive validity) and test impact.
Our research also aims to explore key issues for candidates, such as the way they choose to prepare for our tests and the opinions of students from different social and educational backgrounds.
We work closely with our stakeholders and hold biannual liaison group meetings to:
Research is carried out at the whole test level and for specific universities.
A positive relationship between candidates’ test scores and performance on their academic course provides evidence that the scores can usefully differentiate between even those university applicants who present with the highest school attainment.
We also compare the predictive validity of test scores for different test-taker groups (e.g. males and females). If the same relationship exists between BMAT scores and course performance for different candidate groups then this provides strong evidence that the test is measuring ability fairly for all candidates.