How we develop the test

Teams of language specialists based in Australia, Canada, New Zealand, the UK and the USA are regularly commissioned to write test questions. The writers work from test specifications that detail the characteristics of the four IELTS components, and outline specific requirements. They also guide writers on how to approach the item writing process, including the selection of appropriate material and the development of suitable questions.

Pre-editing is the first stage of the editing process and takes place when commissioned materials are initially submitted by item writers. It ensures that submitted material conforms to the specifications in every respect (e.g. topic, level of language, style of writing, level of task, focus of task). At this stage, guidance on the revision of materials for resubmission is provided to the item writers.

Based on pre-editing feedback, materials are revised and submitted for editing. Materials are then approved for pretesting or are sent back to a writer for further revision.

New materials are pretested on representative groups of test takers from around the world who are preparing to take an IELTS test. The Validation and Data Services team at Cambridge Assessment English collates and analyses resulting data to determine how difficult the items are, and how well they distinguish between stronger and weaker test takers.  They can then make informed decisions on whether materials can be accepted for use in IELTS tests.

Standards fixing involves the testing new Listening or Reading materials with representative groups of test takers. The aim is to gauge the difficulty of new Listening and Reading tests in order to ensure that band scores on every version of the IELTS Listening and Reading tests indicate the same measure of ability. Following standards fixing, the new materials are ready to be used in live tests.

Papers are constructed for all four test components according to established principles. Among the factors taken into account are:

• mean difficulty of the test version and the range of individual item difficulties
• range of language skills tested
• balance of task types
• balance of topic and genre
• range of cultural perspectives
• range of voices and accents in the Listening versions.

Data is routinely collected from live tests and analysed to confirm the accuracy of the initial grading process and to support on-going quality assurance processes.