The only way learners can really know whether they have understood the concepts they have studied is by trying to apply those concepts, ie through some form of self assessment. Instructors too need to gauge learner understanding of course content in order to tailor review and revision activities and ultimately to award certification.
Computers are very efficient at administering a wide range of assessment types, eg multiple choice, multiple answer, numerical and short answer questions. Short answer questions require a word or short phrase as an answer and are marked by pattern-matching learner input against required answer(s).
Questions may be drawn randomly from question banks according to pre-defined criteria. Similarly, question parameters may be randomly generated at run-time.
The result is that learners may take the “same” test repeatedly, but each time be presented with different questions. Thus the memorization of techniques, rather than answers, is encouraged.
In assessed tests each learner is presented with a different set of questions, but covering the same topics and of equivalent difficulty. Thus peering over a neighbour’s shoulder is no longer beneficial.
Results may be analysed in a variety of ways. Instructors can not only see which learners are performing best but also which questions and topics are causing most problems. Thus later activities can be designed to address these deficiencies.
To encourage maximum use of self assessments instructors may wish to design tests so that individual results are not stored, or are only stored on an anonymous basis.
Computer based assessment can (in the long run) save instructor’s time. Creating questions for computerized delivery can be time consuming, though probably no more so than for traditional tests. The benefits come from the fact questions may be shared and re-used in different contexts, and can be marked automatically.
From the learner’s perspective automatic marking means instant feedback which can be absorbed and acted upon while questions are still fresh in the mind. Feedback need not be limited to displaying the correct answer, it can contain hyperlinks to more detailed explanatory material.
Those weaker learners who are disinclined to alert their instructors to their difficulties can use computer based assessment both to identify problem areas and to independently locate remedial resources.
AiM Assessment in Mathematics is an excellent example of an open source model of computer-aided assessment.