I have been experimenting this morning running Linguistic Inquiry and Word Count (LIWC) over students’ peer feedback responses. So far, all I’ve done is analyse the last 50,000 responses in aggregate, and only on a few dimensions. Here is what I found:
The ‘personal’ and ‘formal’ columns are comparison averages generated by the system. I assume those are from bodies of text written by adults.
What stands out to me is how much more ‘cognitive’ students’ texts are. I would have expected quite the opposite. It is also worth noting that the use of longer words is right up there with the adult cohort.
The peer feedback responses are excellent subjects for this kinds of analysis. The writing and opinions are nearly all the student’s own, unlike longer works where there is a lot of quoting and plain plagiarism. It is also sampled on many occasions – whenever a student takes a quiz – which smooths out momentary effects of mood and circumstance.
I have been speculating what else we might do with this technology. LIWC claim to analyse over 70 ‘language dimensions’. Here are three ideas off the top of my head:
- Reading age over time, perhaps added to the existing Flightline chart.
- Progress against Bloom’s taxonomy. That is something I prototyped 10 years ago, but at that time could not harvest the text for.
- Individual or class mood over time. Are they excited by your teaching? Dare you find out?
The Peer Feedback responses are a gold mine for anyone wanting to develop an understanding of their students using digital tools. If you see possibilities in this that excite you, or that would interest your SLT, let me know! Productionising this for Yacapaca would be expensive, but it can be done if there is demand.
Yacapaca 