My colleague and I had a satisfying experience today. Students have taken a low-stakes test where they are marked automatically. The key benefit we have found to this approach is that students can get immediate feedback after the test, and we can provide a meaningful debrief at the same time. Students get closure and move on.
The tests were a month ago though, and a student has contacted us about the marking. My colleague had implemented ‘error carried forward’ in a question of this style:
- (a) What is ‘x’?
- (b) Using your value of ‘x’, what is ‘y’?
- (c) what is ‘z’? [implicitly requiring ‘x’ and ‘y’]
These ‘x’, ‘y’, and’z’ are physical quantities relating to an engineering problem. We mark part (a) on merit, and then we mark parts (b) and (c) assuming the student used their value from (a). This approach is to credit all students for parts (b) and (c), including those whose answer to (a) was wrong. This is known as error carried forward.
Today was different – a student pointed out that they had answered (a) incorrectly, but (b) and (c) ‘correctly’ in the absolute sense, i.e. using the correct value for (a) despite having written the wrong value in the box.
The first step to solving this problem was to make a decision on the mark scheme – should credit be given in this case? The answer was yes. This is an academic issue.
The next step was the satisfying part. We updated the marking algorithm to consider either the relative answer or the absolutely correct answer. Either case would be given marks for part (b) and (c).
We regraded all the tests at the touch of a button, across two different questions, and about 15 or so students saw their marks increase slightly (none dropped).
The impact of this one event was small and of negligible meaning. But the principle is very important. We were tempted to manually adjust the mark for the student who had raised the issue. However, I insisted that we should, as a matter of principle, modify the algorithm and apply it consistently to all students. The result was fair treatment for all.
This small story shows the benefits that are possible when using technology for assessment. It’s a complex issue with many components, but one of them is to assess consistently, and to update assessments and reapply them to all students fairly. This was really satisfying.
Teaching Engineers 