How big data could change the social sciences

Human beings remain mysterious. Our social relationships and the societies we form are still difficult to analyze. All the more so as this mass of people, basically one and the same, changes and adopts different behaviors according to the times, living conditions, structures and so on. The social sciences attempt to decode and provide a rational analysis of the emotional beings we are.

However, with the advent of technology, the perception of these sciences has been profoundly altered. As the speed of information sharing has increased, sociologists have seemed out of touch. Young people are no longer interested in this type of knowledge. This is a distressing reality, since it means that these sciences can no longer act as a counterweight. However, the arrival of the massive data accumulated on the Internet gives many hope that the field of social sciences will soon change drastically.

A revolution...

All this data is potentially a gold mine for sociologists. Empirically, this is the largest database of information available since the early days of the social sciences. Now, almost any subject can be analyzed using the data collected. Methodologically, everything changes with Big Data. It becomes possible for scientists to examine an entire population much more easily. There's no longer any need for limited samples, which used to be open to criticism. Finally, theoretically, this mass of data can improve knowledge and answer questions that are more difficult to analyze with traditional sources, such as the evolution of a language.

The most passionate in the debate on Big Data in these sciences recognize one thing, however: the data collected by Google, Facebook and the other computer giants of this world are not neutral. So there's work to be done in terms of relativizing and criticizing its sources. But this was already the case with opinion polls and other tools used in social analysis before. So why not take advantage of this revolutionary windfall?

... or not

On the other hand, many sociologists criticize this approach. Firstly, because supposed revolutions in the social sciences have been announced periodically for ages and, in the end, they never last. For the time being, many researchers are not jumping through hoops when it comes to processing this massive amount of data. A lot of information is no indication of its quality. In fact, until now, relevant data has been scarce. This is only to be expected, given that most data banks come from departments whose primary aim is to build customer loyalty and increase profitability. A goal that has nothing to do with that of scientists.

What's more, big data methodology isn't all that interesting after all. We're back to Gallup, which, despite a smaller sample size (a few thousand) than The Literary Digest magazine (millions of readers), was able to predict the outcome of the presidential race in the United States. Having data from a larger population doesn't guarantee better results. On the contrary, it makes it easier to get hung up on insignificant details, make sloppy correlations and fall intoapophenia.

Noticeable changes

However, as this article published in the journal Sociologie points out, it would be dishonest to say that this massive data does not, in part, change the social sciences. Admittedly, they are clearly not the revolution announced, but they are making upheavals for scientists. They are forcing present and future researchers to acquire additional knowledge in computer science and even programming. After all, to find more precise information in a database, it sometimes only takes a line of code to sort through the clutter.

This has also changed statistical training, bringing with it the discipline of "machine learning", where researchers learn how to take advantage of large sample sizes and make decent, scientifically rigorous conclusions. The advent of such vast databanks and the Internet also means they can be stored and, potentially, accessed more easily. However, the question of privacy and data anonymization is far from settled, since clever computer scientists know that it is very easy to cross-reference information to find someone.

Big data is not the only future for the social sciences. But neither can it be totally ignored. In fact, some people are proposing the creation of university courses focusing precisely on this aspect of science, to train researchers who will be able to study these immense banks of information and draw conclusions from them to better understand human beings.

Illustration: a-image, shutterstock

References

Belot, Laure. ""Social sciences no longer play their role as a counter-power"." Le Monde.fr. Last updated: September 6, 2015. http://www.lemonde.fr/festival/article/2015/09/03/les-sciences-sociales-ne-jouent-plus-leur-role-de-contre-pouvoir_4745227_4415198.html.

Noyon, Rémi. "Et si c'était le Big Data qui nous disait qui est Charlie?" Rue89. Last updated: May 22, 2015. http://rue89.nouvelobs.com/2015/05/22/si-cetait-big-data-disait-est-charlie-259271.

Ollion, Étienne and Julien Boelaert. "Beyond Big Data." Sociology. Last updated: third quarter 2015. https://sociologie.revues.org/2613.