The use of text and process mining techniques to study the impact of feedback on students’ writing processes

Understanding the impact of feedback in complex learning activities, such as writing, is challenging. We contribute a combination of writing environments and data and process mining tools that can provide new ways of measuring this impact. We use the tools in a field experiment in an engineering course (N=45). Responses (timing, amount and types of text changes) were examined using log data and process mining techniques. Two experimental conditions were used: reflective followed by directive feedback (A) and vice-versa (B). We found that both forms of feedback were read multiple times. Students required longer times to respond to reflective, compared to directive, feedback. The type of feedback, however, made little difference to the types of revisions that students performed. Overall, our findings point to the difficulty of encouraging students to reconsider and revise what they have already written.