Social Research Themes for Computer Scientists
Just a week ago, I presented a research article at an international conference on digital governance and digital politics. This conference was organized by St. Joseph's college. While most of the papers were by political scientists and mostly qualitative in approach, a couple of them used sophisticated tools from econometrics. Mine was the only paper that used the computational approach.
This was not the first time I presented at a social science conference. Experiences were new, however. During my earlier presentations, I was working at a think-tank. This time I am a teacher at an engineering institute. Now, one of my primal duties is to motivate my students to read beyond the textbooks, to do research. A thought was frequently haunting me; how could I inspire my students to pursue research in the areas I am interested in, the social sciences?
During the breaks, I sketched the links between the social sciences and the computational. There are indeed several approaches or themes which computer science students could follow to produce a significant impact in social science research.
The straightforward way through which a computer scientist could enter the domain of social science is to apply or adapt existing computational methods to a sociological problem of interest. This is exactly what I did in the conference paper. My paper analyzes a large number of ICT policy documents using topic modeling and n-gram analysis.
Computer science students could identify several other research questions where they can use computational methods. For example, they could use machine learning to infer inequality patterns from large scale household survey data, like that from the national census. They could also use graph theory to understand how information and misinformation flows in social media. There could be many such research questions. However, I need to put a cautionary note here. A mere application of computational methods to social data, without a good understanding of the research domain, would not be fruitful. In my case, for example, I have been involved in policy research for the past several years. This is why I could interpret the results of the topic modeling algorithms, and could convert the findings into a conference paper.
Rather than applying an existing method, a computer science student could dig into the depth of the method to understand how it behaves in a sociological context; this type of research could be interesting even a hardcore computer scientist. For example, consider big data analytics. It is usually argued that big data gathered through the Internet of Things (IOT) could be utilized to build better policies by the government. But because of the fact that there is huge inequality in internet use across the individuals, socio-ethnic groups, and geographies, inferences from IOT generated big data is likely to be biased in the favor of the rich and the influential. A computer science student could study the extent of and the ways to mitigate such biases.
The problem I mentioned above is a specific question of the broad and newly emerging academic field known as Algorithmic Governance, which studies the pros and cons of using computing technologies, mainly AI and blockchain, in everyday social life. Researchers from various disciplines like political, social and computational sciences are contributing to understand Algorithmic Governance and its consequences. For a student of computer science, Algorithmic Governance could be an enthralling research theme.
Finally, I mention one more research area for computer scientists, computational sociology, which aims to understand various sociological phenomena through simulation modeling. Important topics in computational sociology are mainly the evolution of wealth, caste and other inequalities, and the evolution of social norms. Computational sociology is already a mature subject with extensive research literature.
The conclusion is that a computer scientist's research does not need to confine within the decades long definitions of what computation is. In other words, one need not limit his or her study to the traditional areas like databases, data structures, or software engineering. One could be a serious computer scientist and yet research in the sociological domain.