Assessing word-processing skills by event stream analysis

This paper derives the algorithms required to process the stream of textual events collected from a candidate's interaction with a word processor to produce an assessment of their word-processing skills. The information that can be extracted from the textual event stream is compared to that which can be deduced from a comparison of the candidate's submitted answer with the model answer(s) generated by the examiner. For many examinations, document comparison is simpler and more efficient than event stream analysis but it is not always possible to fully analyse errors from document comparison; hence a mixture of document comparison and event stream analysis is desirable for computer-based word-processing assessment.