(Sociology 596) Computational Social Science: Social Research in the Digital Age

In the last decade we have witnessed the birth and rapid growth of Wikipedia, Google, Facebook, iPhones, Wi-Fi, YouTube, Twitter, and numerous other marvels of the digital age. In addition to changing the way we live, these tools---and the technological revolution they are a part of---have fundamentally changed the way that we can learn about the social world. We can now collect data about human behavior on a scale never before possible and with tremendous granularity and precision. The ability to collect and process "big data" enables researchers to address core questions in the social sciences in new ways and opens up new areas of inquiry.

This course on computational social science will emphasize social science rather than computation. We will focus on how traditional concepts of research design in the social sciences can inform our understanding of new data sources, and how these new data sources might require us to update our thinking on research design.

Now a little about mechanics. Each three hour class will consist of a general discussion based on several readings. Then, students will take turns presenting specific papers that apply the ideas from the general discussion. Students are expected to come to class prepared for the general discussion as well as present a few articles during the course of the semester. There will be no exam.

Your grade will be based on the following components:

• Class participation and in-class presentations: 25%

Each student will be expected to present a few articles during the course of the year. Each presentation should begin with a 30-second summary of the article and then move to a more elaborate discussion of the key issues in the paper. The student presenter will be expected to answer any questions that come up from the class.

• Response papers: 75%

Each student will be expected to write a short response paper (2-3 pages) every week, except the first week. Students should view them as a chance to play with the ideas in the readings: look for contradictions, establish connections to your own research, develop empirical tests, etc. The response papers should not be simple summaries of the readings. All response papers should be sent to me by Monday at midnight on the day preceding the class.

There are no official prerequisites for the course, and students from all departments are welcome. Undergraduates interested in taking the course should contact the instructor for permission.

Meeting 1 (11/6/12) Introduction and Ethics

In this first class we will cover a broad overview of web-based research, focusing on both strengths and weaknesses. We will also discuss ethical issues that will arise throughout the course.

For general discussion

Meeting 2 (11/13/12) Individual experiments

The web offers numerous advantages over the traditional laboratory for the conduct of social science experiments. First, the web allows researchers to conduct experiments on a completely different scale; lab experiments are limited to hundreds of participants, but web-based experiments involving tens of thousands of participants have already been conducted and larger experiments are becoming increasingly practical. The web also allows researchers access to a much broader pool of participants and allows researchers to study decision making in a more natural environment. But, conducting experiments on the web also includes some drawbacks including unknown participant pools and limited control over participants. In this meeting we will discuss four types of web-based experiments where the unit of analysis is an individual: A/B tests on existing sites, overlayed experiments on existing sites, quasi-experiments, and experiments using micro-payment platforms (e.g. Amazon's Mechanical Turk). The strengths and weaknesses of the various approaches will be compared.

For general discussion

For presentation

Meeting 3 (11/20/12) Collective experiments

The web also allows for collective experiments, where the unit of analysis is a group, not an individual. These collective experiments introduce numerious logistical complications, but can be used to address questions that are otherwise extremely difficult to study.

For general discussion

For presentation

Meeting 4 (11/27/12) Mobile phones and wearable sensors

There are approximately four billion mobile phones in the world. While these devices are often thought of as "phones," the newest wave of "smart phones" that are increasingly dominant in developed countries are actually sophisticated mobile computers that offer amazing opportunities for researchers. In this class we will discuss the two main forms of research using mobile phones and wearable sensors: research that uses data collected from individual devices and research that uses aggregate data collected by mobile phone companies. Within the category of research that users individual devices, we will distinguish between research that uses phones and research that uses custom-build devices. We will also distinguish between active and possive data collection.

For general discussion

For presentation

Meeting 5 (12/4/12) Digital traces

Human behavior in the digital age often leaves behind traces, and these traces are being aggregated on a scale that is difficult to comprehend. In this meeting we will discuss the strengths and weaknesses of using these traces for social research.

For general discussion

For presentation

Meeting 6 (12/11/12) Crowdsourcing, Citizen Science, and Conclusions

Anyone who has used Wikipedia understands the power of large-scale social collaboration. How can we harness this collective power for other intellectual challenges?

For general discussion

• [play ESP game]

For presentation