Chris Williams is a visiting fellow at University of London (IoE) and a visiting professor at Chubu University Japan. You can contact him at email@example.com
Over the past 2-3 years, crowdsourcing research has become significant in diverse areas - astronomy, archaeology, epidemiology, environmental activism, state crime. Bird watching is one of the oldest examples. The RSPB uses thousands of volunteers to map sightings on a website.
Syria Tracker is a very pertinent current example. Evidence from within Syria - mobile phone photos, videos, emails - is sent to a Facebook site. Volunteers around the world compare this, with satellite images of Syria. Signs of state military action - like square objects which have a projection at one end, and are probably tanks - can be compared, for example, with the on-the-ground reports of hospital causalities. The comparisons are the key to ensuring the reliability of the evidence.
But there is no systematic understanding of this emerging methodology, so I'm investigating it through web searches, email contacts with activists, and interviews with technical experts.
I need to find out the underlying structures of this new approach to research, which means finding the exciting new e-enabled initiatives, but also identifying the historical precursors. For example, few people know that the Oxford English Dictionary, and museums like Pitt Rivers in Oxford, were products of crowdsourced research.
Understanding manual and e enabled strategies can suggest ways to develop future e-enabled methods, perhaps combined with manual methods.
For example, Google earth is being used as a basis for people in remote African communities to create detailed maps of their region. The data from numerous individuals, who are doing micro maps of their locality, can be compiled into a big regional map.
One of the in-progress findings is about sampling, and this is being discussed avidly by epidemiologists, who would normally use very robust random samples. The argument is that if they can collect very large amounts of crowdsourced data, can they then take a random sample from that mass of crowd data, and get robust results, even though the respondents are self-selecting?
Another outcome is that, generally for international research, we have to rely on state data - dubious statistics that are endlessly copied and rehashed. Crowdsourcing can get data about states that completely by-passes the state, as in the Syria example.
One significant difficulty is that relevant initiatives - old and new - are not all termed and tagged as 'crowdsourcing'. It took me ages to discover interesting initiatives, beyond a Google search on ‘crowdsourcing’ which mainly throws up boring commercial endeavours using the Amazon Mechanical Turk platform.
In Japan, since 1985, a high school teacher, Masatoshi Yamashita, has been collecting data from elderly fishermen who were affected by radiation from the Bikini Athol nuclear tests. He has used schoolchildren to identify and talk to them. Of course, the Japanese government has colluded with America to cover-up this problem, so it requires an effective ‘crowd’ endeavour to get the evidence before these victims die.
The reason this relates to the ‘termed and tagged’ problem, is that I know Mr Yamashita. A few years ago I sat in his house and saw him work on his maps. But I didn’t make the link that this was a form of crowdsourcing. And…Mr Yamashita is the uncle of my wife!
So I need to create constantly evolving typologies of crowdsourcing research designs and data collection methods, and then revisit the data, and my memory, to see if something reminds me of other hidden examples, like my uncle-in-law. So the key qualitative ingredient is actually a sort of reflexive personal brainstorming - matching what is happening in the world with what might be happening, hopefully, in my memory.
Williams, C. (2013) Crowdsourcing research: a methodology for investigating state crime, State Crime, 2:1.
Williams, C. (2-12) Researching power, elites, and leadership. London: Sage, p140.
Williams, C. (2014) Doing international research. London: Sage (In-progress)