March 21, 2018 by Kate Bowles
In week 4, you’re at a great stage in a new research project: thinking about the questions that really need answering, before you have to confront the reality of what you can and can’t find answers to.
As it happens, this question of how we can find stuff out is also the one that’s behind this week’s big global news item: the work that was done in 2014 by a coalition of quite shady actors including Facebook to figure out if online personality profiling could be used to influence voter behaviour.
Remember all those personality quizzes and social games that were around a few years ago that asked you to log in with Facebook? Yes, those.
It turns out that researchers who wanted to model means of influencing human behaviour had come across the gift that kept on giving: cheap or even free user data at unimaginable scale.
There’s so much to be appalled by in this story, not least the behaviour of academics. But let’s stop and think about this, as our goal is to understand research practice, even when it’s confrontingly awful.
Cambridge Analytica started out with quizzes. And this is where many of you are starting out: snap polls. They had to confront the risk of small sample size, and you can have some sympathy there.
So they must have been pretty happy to have access to Mechanical Turk, a well-established source for getting surveys done for micropayment (or “a marketplace for work that requires human intelligence”, as they put it) owned by another giant corporation, Amazon.*
And then they applied a big multiplier: Facebook. If Turkers would both complete surveys and hand over their Facebook details, researchers would potentially gain access to unimaginably huge troves of user data, thanks to corporate standards of consent that hadn’t anticipated what could possibly go wrong.
This breaking news story is one that anyone interested in research practice should be following closely. It’s a cautionary tale with far reaching consequences, and it gives you a sharp lens on how research is actually being conducted. We’ll be tracking it over the next couple of weeks, as we think about big data visualisation, and research ethics.
Appreciated this week
So it’s a relief to get back to BCM212 projects, that are being conducted by actual humans working at small scale, working reflexively, and working with an ethic of relational care for the people who might answer their questions. This stuff we’ve been looking at so far isn’t just a matter of words — and we see this most clearly when these values go missing.
Here’s a tiny sample of good things happening.
Ashleigh is looking hard at the prospect for creative employment in Australia, given the cost of a degree in this sector. Cost-awareness is an issue that affects many projects: are students conscious of how much degrees cost?
Cooper is one of quite a few of you looking at transport impacts. The detail that caught my eye in Cooper’s post is that the physical and mental cost of travel is significant for commuter students: arriving at uni too lethargic to function. Meanwhile Eirist is looking at a specific and notorious bus route and asking whether students from the Macarthur area have good enough transport options.
With quite a few of you considering similar projects, April’s post caught my eye as this was the only one I saw on something the university really invests in: O-week. Is O-week just a festival for commencing first years, or could it be something that has value for returning students? This is a great question as it asks us to think about what universities are up to in their O-week practices.
But as you all know, uni turns out to be not all jumping castles and showbags. Matt’s post is full of practical details about the cost of living in university accommodation. For students from regional areas, living somewhere out of home is essential. Uni accommodation is costly–but what are the alternatives? and who pays? And what about homeless students? And Brock is asking where Centrelink and the university interact in relation to students having to work full-time while studying full-time.
There are some important new research initiatives in the United States looking at the impact on students of insecure housing and food insecurity, asking colleges to be much more attentive to how much paid employment students are having to take on to get by. If you’re researching work or housing, follow researcher Sara Goldrick-Rab on Twitter, as she’s been leading a huge national effort in the US to change the way colleges think about who their students are, and the lives they lead.
Thanks to all the bloggers we’re reading this week, for putting such care into why you’ve chosen your project, why it’s important to you, and what we all have to learn from it.
More next week.
*(For insights on the experience of turking in the global gig economy, follow Rochelle on Twitter. And the Mechanical Turk itself has a pretty interesting history as a model for fraud — check it out.)
An earlier version of this said CA started out with surveys — important to clarify that the research that’s the focus of critique used a quiz to gather data. It’s not quite the same.