Data Science & Psychology Data Science applied to Values, Morals, Politics, & things that matter.

This weekend, I submitted a talk which considers how semantic technology will help us answer future questions for consideration at SXSW as well as helped edit a chapter on Moral Foundations Theory that contains a section on the future of moral psychology.  I have a lot of thoughts on the future of moral psychology, many of which relate to future semantic technology, that probably aren't quite right for the academic audience of that chapter, so I thought I would list them here while I was thinking of them.

What will the future of moral psychology look like?  Here are a few trends I see defining the next 10 years.

- Thinking outside academics: I hope that moral psychology, as studied by social psychologists, will start to think of itself as a part of the world rather than as a largely academic exercise.  There is a kind of epistemological arrogance/insecurity that exists in any academic discipline (this paper by Paul Rozin points it out in psychology best), which leads one to believe that your methods will point out the truth while others' won't.  Philosophers, neuroscientists, anthropologists, and sociologists all study issues of values.  Facebook, Google, novelists, and human resources departments all collect data of relevance to moral psychologists and answer questions about the relationship between values and various dependent variables every day.  My hope is that these methods will become more compatible, which leads me to my next point.

- Leveraging semantic technology to crowdsource findings: Right now, most findings in psychology exist as text that is impenetrable to machines.  The model is one where a small minority from a single institution (academic psychology) are supposed to do all the work, in the form of these papers that are supposed to be definitive, as opposed to more modern models of collaboration made possible with technology.  This TED talk by Clay Shirky illustrates it best, where he shows how questions can be answered by the "long tail" of contributors.  Rather than a single researcher figuring something out, we need to figure out a way for any individual to be able to contribute data toward an answer and then to be able to aggregate/coordinate that data towards a more robust answer.  That is the promise of semantic technologies, which allow study results and data to be combined, in the same way that meta-analyses (the current labor intensive gold standard in psychology) do now, but in a way that anyone can contribute data to an answer and the meta-answer is updated in real-time.

- Open access results: Of course, semantic standards will make little difference as long as results are accessible only to those with expensive institutional access to research.  Leveraging the power of collaboration that Shirky talks about requires a level of openness that academics are moving towards, though slowly.  Brian Nosek and Yoav Bar-Anan have a great paper on openness that will likely convince many people (also see our forthcoming commentary on that paper).  We are publishing our libertarians paper in PLOS One, an open access journal, in part because, all things being equal, we'd rather not lose the copyright to our work and would like others to be able to read it freely.  One of my hopes is that it starts a trend where, all things being equal, people choose open access journals.

- Answering real world questions:  Openness allows non-academics the chance to contribute and merge their data into datasets that combine variables from moral psychology, business, web analytics, economists, politicians, etc.  With this data, we can answer questions that seem to interest me, but are deemed too "applied" for pure academic study such as psychological differences between owners and renters, cell-phone users and landline users, or people who like or dislike public transportation.  These are the kinds of real-world variables that (in my opinion) real people care about and where we can make a real contribution to society.

- Creating real-world value: How do we know that physics creates value?  Everytime we turn on a light or start our car engine, we prove principles of physics.  People use physics.  If we want people to really believe in the power of moral psychology, we need to get people to use moral psychology.  There is no shortage of organizations out there that want to impact variables of interest to moral psychologists and fail or succeed, we'll learn something.  People will really believe in the power of psychology when they see that people are using it to achieve real world outcomes, such as getting people to vote via descriptive, rather than injunctive norms on Facebook.  Some of that exists already, but the opening of psychology to non-academics who can contribute their own data and their own variables will vastly speed up this process.  Imagine a world where Facebook shows that it's "I Voted" avatar accounted for 3% of the variance in turnout in the Ohio election, tipping the presidential election, and updating the data commons on social influence effect sizes as they pertain to voting, as opposed to other domains.  That future isn't that far off.

- Ravi Iyer