Umalusi meeting with language experts
Umalusi meeting, Cape Town 2013
Agreement with Radboud University
Agreement - Radboud University, 2012
Albert Weideman
AILA Conference 2011, Beijing
NWU signs ICELDA agreement
US signs ICELDA agreement
UFS signs ICELDA agreement
Test workshop
Workshop with Prof. Alan Davies (R)


Our tests

ICELDA has five tests that have been developed for various purposes and levels, while four more tests are currently under development. The most well known are:

  1. Test of Academic Literacy Levels (TALL)
  2. Toets van Akademiese Geletterdheidsvlakke (TAG)
  3. Academic Listening Test (ALT)
  4. Test of Academic Literacy for Postgraduate Students (TALPS)
  5. Test of Academic Literacy (TAL) of the Free State School of Nursing for prospective students of nursing

TALL and TAG make up the bulk of the annual sales, with some 31000 being purchased in 2010. TALPS is another currently popular test. ICELDA not only makes these tests available to its partners and to others, but also has the capacity to help scholars in other parts of the world to develop tests appropriate for their environments. We have no doubt that the tests we design will fulfil a critically important function not only in South Africa, but elsewhere as well.


Eligibility or Placement?

The tests are primarily used to determine academic literacy levels of first-time entering students. The results are used either to place students those whose academic literacy level is too low on appropriate academic literacy support courses, or, more rarely, as part of an index to determine eligibility to higher education (access to university). In the latter case the results are used for high stakes purposes, and we recommend that the ability to handle academic discourse at university level should not make up more than 15% of such an index.

Our tests all measure a construct that relies on our definition of academic literacy.


How reliable are our tests?

The reliability of our tests is discussed in accredited journals (see bibliography). TALL, for example, has reliability levels (measured in terms of a very conservative measure, Cronbach's alpha) that averages at 0.93, which is more than 20% higher than international benchmarks for high stakes tests.