Political Science & Quants; Time to Bruise Some (Theoretical) Knuckles



Regressions HappyBig Data, the process of transforming different aspects of life in digital format has transformed the way we perceive the world. Google can predict market fluctuations and potential flu outbreaks just by analysing key words in searches.  Recently researchers from Cambridge produced a software that based on Facebook data can predict your personality with 88% accuracy. Social network analysis goes over tens of thousands of tweets predycting anything from riots to electoral results.  The list goes on. The abundance of data is staggering; currently 98% of all information is digitized.  Digital information doubles twice every 2-3 years.

This shift has affected not only scientists in labs and math geniuses that create algorithms worth millions.  It has also affected the less popular Humanities; this includes researchers looking at politics.  A modestly popular subject these days; globalization, human rights or the financial crisis are just some of the reasons many have moved into political science. 

Political scientists have followed the trend.  Statistical analysis and in particular inferential statistics' popularity has grown.  So we find out a lot about how things happened, what happened, and likely directions.  We get trajectories of likely electoral results, political party popularity, policy success and so on. 

Thus we examine correlation not causation.  How likely is one phenomenon to occur with respect to another.  There is nothing wrong with that as long as we recognise the weaknesses it includes.  Data is and will be useful in unravelling a great deal about the mysteries of the world whether in political science or elsewhere.  However, this has brought on a quantitative fixation and has reduced the value of qualitative analysis.  Regressions are ran nearly in autopilot just to increase the legitimacy of one's research. In some circles qualitative analysis is almost a dirty word; only members of the anthropology department dare mention it.

Too many studies today especially on the EU take a quantitative perspective without addressing the underlying theoretical implications of their research.  The classic motive for PS papers is something along the lines: "we analysed X [ridiculous] amount of data under an “original” model.  Statistical analyses fits model and vice versa; publish; break out the champagne".  We are not really advancing the level of theoretical argumentation. Theoretical frames allow for the deduction of arguments/ hypotheses.  Moving beyond methodological debates, just ensuring that your quants fit a sefl-created statistical model sounds and tastes like dry chicken. 

Time to Bruise the Knuckles

The result of this attitude has been a surge in political science academic articles that are not drawing any punches against anyone.  Instead a lot of analysis/ research is moving in parallel.  The theoretical argument is not moved forward while scholars demand further quantitative analyses; biggers Ns are gracefully aplauded. But little is said about the usefulness of arguments or more importantly which ones are not useful. 

Two things to bear in mind; first quantitative analyses will be an important part of the future research including political science.  However it should complement the “why” question instead of replacing it with the much more boring “how/ when”.  Second, researchers should step up the theoretical debate.  That means taking the gloves off, upping the critique and “going after” their colleagues.  No need for a brawl but perhaps tell off the next quants analysis that just gives you another correlation.  Boys and girls in the field should start pulling some theoretical jabs and punches.

Quantitative analysis is a great tool, it is supposed to test theory and move the debate forward.  It’s a shame we have been using to create statistical models on top of other models.  Healthy academic competition is nothing without costs.  The mass of data we are collecting is an opportunity to make theoretical leaps and complement qualitative research.

PICTURE: Practical Neurology

Related Stories

When reflecting on the Eurozone debt crisis, it is important to remember that while Portugal, Greece and...
6 years 31 weeks ago
Form Your Opinion is launching it's first online debate! With the support of the European Institute of the...
6 years 38 weeks ago