See Andy's other stuff:

Contact Me >>

Notes on the History of Social Network Analysis

For you word of mouth / viral methodology geeks.  Let’s remember that we may think this is a new field, but it’s been studied for decades.  For more academic/historical perspectives, I also recommend Linked by Albert-Laszlo Barabasi. 

Compliments of my Northwestern colleague Tom Collinger, Integrated Marketing Communications Chair, Associate Dean, Medill. 

——— 

11 April 1996. Drawn from John Scott Social Network Analysis: A Handbook. Sage Publications. 

Although some of the ideas of network analysis are found in the writings of scholars going back to the ancient Greeks, the main development of the field occurred in the 1930’s by several groups in different traditional fields working independently. 

Psychology

The gestalt tradition associated with Kohler (1925) is about how the mind works. They stressed the organized patterns that structure thoughts and perceptions. These organized patters were regarded as "wholes" which have properties distinct from those of their parts. In part, we see objects like plants and people the way we do because we have preconceived conceptual schemes built into the mind. Perception is a kind of pattern-matching process. Three scientists working in the Gestalt tradition fled to US around that time: Kurt Lewin, Jacob Moreno, and Fritz Heider. 

Moreno developed sociometry. He started asking people who their friends were and explored the ways in which their relations with others served as both limitations and opportunities for action and for their psychological behavior. He founded the journal sociometry which investigated the relationship between psychological well-being and "social configurations". He believed that large scale social phenomena, such as the economy and state, were sustained and reproduced over time by the small scale configurations formed by peoples patterns of friendship, dislike and other relations. 

Moreno invented the sociogram — a diagram of points and lines used to represent relations among persons. Before that, ideas like the "social fabric" or "social network" were just vague ideas. Moreno used sociograms to identify social leaders and isolates, to uncover asymmetry and reciprocity in friendship choices, and to map chains of indirect connection. One of the configurations he observed was the sociometric star: an individual chosen by many others as a friend. 

Lewin studied group behavior, which he said was a function of conflicting social forces. He thought of the group as existing in a social space or field consisting of the group and it’s perceived environment. The group and its environment interact and the meaning of these interactions is constructed by the group members on the basis of their perceptions and experiences. Lewin argued that the structural properties of this social space could be investigated mathematically using vector theory and topology. 

Again, the field is seen as consisting of points connected by lines. The points are individuals, their goals, or their actions and the paths represent the interactional or causal sequences that connect them. The field model is all about representing interactional interdependencies.  Fields are divided into regions, separated by an absence of paths between them. The opportunities which individuals are determined by these boundaries. 

Heider worked in the area of social perception and attitudes. He developed what is known as balance theory. He said the mind seeks balance (an absence of tension) by trying to hold ideas that are not in conflict with one another. This also applies to attitudes towards other people. He was especially concerned about what happens when a person is emotionally close to two people who start becoming hostile to each other. If A likes B, then A wants to like and dislike all the things that B likes and dislikes. If B dislikes C, then A wants to dislike C, but what if A and C are friends? There is a tension that must be resolved. One solution is to choose sides. A can dislike C.  In actual groups, imbalances develop because not everyone is interacting equally with everyone else at the same time. But once the imbalances make themselves felt, they exert force to resolve themselves, leading to changes in the group structure. 

Cartwright and Harary showed mathematically that the outcome of this process is necessarily a group subdivided into cliques within which all ties are positive and between which all ties are negative. All groups in which there is any imbalance are in a state of slow transition towards cliques. 

Anthropology

One of the biggest emphases in social anthropology this century was on social relations. It was understood by Radcliffe-Brown and Nadel that what social structure was was concrete relations among individuals. It had long been understood that in pre-industrial societies kinship relations were extraordinarily complex and important. But other relations, such as friendships, were equally important in industrialized societies. These people wrote at a theoretical level about the web of relations comprising society. 

At the empirical level, the work of Warner, Mayo, Roethlisberger and Dickson on the Hawthorne plant of the Western Electric Company in chicago in the 20s was a milestone. Came from a series of worker efficiency studies in the tradition of Taylor. They tried to figure out how alterations in the physical conditions of work (heating, lighting, rest periods) affected productivity. They found that productivity seem to increase with any change they made. So they called in Mayo who decided that the very fact of participation in the research made workers pleased with management and more integrated and this motivated higher productivity. 

So they began to study anthropologically. They simply watched people work all day, with special attention to the relationships among the workers. In the process, they discovered the "informal organization" of the organization — the hidden social structure which seemed to have as much effect on worker productivity as anything in the company manual. 

Later on in the 50s, the dept of social anthropology at Manchester University in England began looking at conflict in groups. People like Barnes and Bott and Mitchell began investigating how the structure of relations among people affected not only the individuals but the society as a whole (e.g. its cohesiveness). This group paid as much attention to the structure of social relations as to the content of those relations. 

Based on their work, a group at Harvard led by Harrison White in the 60’s and ’70s further developed the mathematical side of social network analysis, translating many important concepts from the social sciences, such as the notion of "social role" into mathematical form that allowed them to be measured and modeled. 

One of White’s students was Mark Granovetter, who wrote a book called Getting a Job in 1974 in which he asked people how they got the jobs they hold. Most obtained them through accidental contacts with others rather than purposeful search via official means (e.g. newspaper advertisements). And of those who learned of the opportunities through contacts with others, few got it from family members and close friends. 

Most were through acquaintances. This phenomenon was explained through a theory about information diffuses through social networks. This was a seminal work in social network analysis. 

A similar study was done by Lee in 1969 called The Search for an Abortionist. Abortion was illegal, and doctors who performed them could not advertise nor operate in clinics. To find them, women asked their friends and acquaintances. She found that on average, the abortionist was 4 links away from the patient in social space. (Woman — person — person — person — Doctor). 

Mathematics

Euler (1736) settled the Konigsberg Bridge problem. By translating into a mathematical notation involving points and lines, and then deriving some proofs. This idea was rediscovered many times in different areas of math and applied sciences. In statistics they developed the notion of markov probability chains. In physics they were used to understand molecules adjacent to each other in euclidean space. In operations research, graphs were used to map out the location of goods and channels of transmission. 

Today, graph theory is a well-developed area located at the intersection of combinatorics and topology. 

[contact-form-7 id="27185" title="contact-form 3 TellAFriend-Post"]