In the social sciences, I use statistical techniques to uncover data trails of human behaviour. I look for some really hidden behaviours – ones that no-one is going to admit to in a survey, like ‘Did you cheat or guess the answers in your exam?’ We use probability based measures and look at error terms to make our judgments. It doesn’t sound wildly different from the work of people who prefer to use the title ‘data scientist’. Yet it’s taken me the best part of 4 years full time study to make sense of what my computer science colleagues are up to. Educational assessment data analysis is, in principle, a similar process, whether you came to it via years of classroom teaching or after studying COMPSCI 501, but there’s one huge difference. We don’t have the same rule book, and that’s the wonderful thing that Web Science does.
What does the rule book look like in my subject of educational assessment? Well, for a start, I would be very, very careful what I put in the data collection process. That rule book goes on forever. Some are obvious:
‘Make sure the question is clear. I mean really make sure.’
Others take a while to process:
‘Don’t ever, ever mention granny*.’
I would trial the data collection process, experiment with different assumptions, about the distribution of the trait in the population, agonise over how I’ve modelled missing data, and rigorously remove anything from the data collection process that looked a little suspect. I’d still be deeply worried at the end of that process that perhaps I hadn’t done enough. And, most importantly, it would never, ever occur to me to suggest that a solely statistical decision was adequate. Reading computer science papers, I can never tell really what assumptions they’ve made, but there seem to be an awful lot.
The GDPR is a universal rule book for all, when it comes to data matters. We all know now, at a basic legal level, what rules we are meant to be following. Nearing the end of my studies with the Web Science CDT, I’m still looking forward to attending the Web Science Summer School as though it was the first week. There’s still so much we need to learn from each other. And for me, that is what Web Science has really been about. It’s about working together to create one rule book for all.
Clare Walsh is a final year student on the iPhD programme and will be attending the Web Science Summer School the week of the 8th – 13th September.
*The reason for this is rather sad, but obvious once you’ve seen it. Around the same time that we ask children to take stressful major school leaving exams, many will be experiencing the loss of one or more grandparent. They’re at that age, and exam rooms are already highly emotionally charged environments. It would be grossly unfair to remind them of their recent loss and sadness.
Written by Clare Walsh