Teaming up for big data

Big data offers exciting new opportunities for social research. Group processes that could only be described sketchily, because so much was happening at the same time, can now be analysed in detail, since a lot of data can be captured simultaneously.

Recently, Wi-Fi tags were introduced in a Dutch health care centre for the elderly with psychiatric problems. Both the 65 patients and staff wore them continuously for three weeks. Receivers in all rooms tracked their movements. The aim was to acquire detailed information about interactions between employees as well as between employees and patients in the health care centre, with the ultimate goal of providing better care.

‘Of course, the Wi-Fi tags only say something about proximity’, says assistant professor Maaike Endedijk, whose interest is in complex learning settings. ‘We cannot be sure if people have been interacting, but when they were in the same room together for some time, it is very unlikely they just ignored each other. Also, we don’t know the nature of the interaction.’

Still, this information is so much richer than traditional methods for social research like observation and questionnaires, according to Endedijk’s colleague Elze Ufkes, whose background is in social psychology. ‘In this field much research takes place in lab settings just because that’s an ideal way to closely observe important, but often subtle, group processes. But how does this knowledge translate to real world situations? That’s why it is so exciting to have new tools at our disposal.’

Potential

Endedijk and Ufkes specifically teamed up to explore the potential of new technologies. The Wi-Fi tags are easy to use, but still fairly limited in what they can achieve. There are more extensive tags that also record speech and the direction people are facing, but these are bulky and hence awkward to use. Smartphones might be an alternative, but these have their own drawbacks, such as mutually incompatible technologies. Even more serious are the privacy issues raised by following “research objects” 24 hours per day.

‘One of the things we encounter is that our traditional methods of analysis, which work fine if you have a couple of hundred questionnaires, simply cannot cope with millions of data points’, says Endedijk. ‘How do you extract information from that? We are lucky to work at a university where lots of expertise on big data is available, which we can call upon when searching for new analytical methods.’

With automated data collection and analysis new ways of doing research also come within reach. For instance, one could automatically send a questionnaire to participants for extra information about an incident as an interesting data pattern turns up, instead of finding out (or not) about it through a questionnaire weeks later, when recollection may already have become hazy. Also, it is easier to follow developments through time, as they emerge – until now, with interviews at regular intervals, researchers are likely to miss the crucial moments.

Currently, Endedijk and Ufkes focus on methodology, which brought them together from research areas that traditionally have very little to do with each other. ‘There is still so much to discover at the basic level, with implications for all social sciences’, Ufkes explains. ‘For instance, how do you read data from devices, how do you clean up this data, how do you perform a quick scan to see if the data is any good? We are developing protocols to deal with that, so others won’t any longer have to draw their own rules. This really is pioneering a new field.’

Application

So, there is no question that a plethora of social data is about to open up. According to Endedijk and Ufkes it will certainly prove to be useful in numerous settings where people work together. Take, for example, the flexible office plans that are popular nowadays. What do they mean for interaction between employees? Or, the emergence of self-managing teams. What actually happens in these teams, which interactions at which moments appear to be vital for them to function efficiently?

Endedijk: ‘Many organisations strive to be learning organisations. Many want to be innovative. We do know something about what makes them learning or innovative, but not a lot. Closely following the interactions in these type of settings may tell us a lot more about the factors that lead to success. We may be in the pioneering phase right now, but we are convinced that this development will have a large impact.’

Geschreven voor een brochure van de Universiteit Twente