June 5, 2005
Making the grade
Computers have been grading multiple-choice exams for some time; now they’re assessing style and content:
Dr James Christie, a lecturer in computing at the Robert Gordon University, Aberdeen, who created the program, said it would cut the volume of work that academics have to mark and, if applied to exam scripts, could replace thousands of traditional examiners. “In the past few years there has been a growth in its use and an acceptance of the principle,” he said. “We have now moved from simple ‘yes or no, true or false’ computerised marking into free text—sentences and paragraphs.”
To use the software to judge style, a lecturer grades a small sample of essays in the usual manner. Information about where and why his marks are allocated is then fed into the computer. It analyses elements such as vocabulary, sentence length and grammar and creates a formula by which the remaining essays are marked. To mark content, a lecturer provides details of the possible texts that could be used to answer the question.
Dr Christie said that the Schema Extract Assess Report (SEAR) programme worked best on “content-rich” essays. “Questions such as ‘describe the operation of the heart’ or ‘apply the Data Protection Act to this case study’ are most appropriate,” he said, whereas “explain irony in the novels of Jane Austen”, he admitted, would be more difficult.
Examiners, of course, are not exactly happy about all this (or about having exam papers processed in India). But to quote Austen, those who do not complain are never pitied.
TrackBack URL for this entry: