Employment Status Driven Interactions in a Bank Wiring Room
Introduction
In the following study, we analyzed the Bank Wiring Room data set, an observational study conducted in 1939 by F. Roethlisberger and W. Dickson at the Western Electric plant. This data documented the networks of fourteen employees: two inspectors, three solderers and nine wiremen. For our analysis, we interpreted the roles of each employee type as follows: inspectors regulate and manage the plant’s division, solderers join or mend metal materials, and wiremen maintain electrical equipment. The researchers collected data on the network interactions between these individuals for six different behaviors: horseplay, arguments over open windows, antagonistic behavior generally, giving help with work tasks, trading work tasks, and friendship. Additionally, the data was considered in an asymmetric manner, indicating a possible lack of mutuality between two actors.
The data was originally gathered by Roethlisberger and Dickson in order to measure the “Hawthorne Effect,” or the way in which individuals alter (most typically by improving) their behavior when they know that they are being watched. At the time, the discipline of social network analysis was not nearly as established as it is now, and they did not have the computer power to conduct the kind of analyses that are now possible. For this reason, we return to this dataset from 1939 to ask new social network analysis questions that are now answerable with current computer power. We propose the following questions:
1. Are the network structures for pro-social behaviors (horseplay, friendship, helping one another with tasks) consistent, and do they significantly diverge from the network structures for anti-social behaviors (arguments over windows and antagonistic behavior generally)?
2. Does employment level influence tie formation within pro-social and anti-social networks?
For our first research question, we predict that the network structures for pro-social behaviors will be consistent with one another and will diverge from the network structures for anti-social behaviors. For our second research question, we predict that actors of higher status will be less involved in the behavioral interactions of the workplace than actors of lower job status. Otherwise stated, job status is inversely related to network centrality and network interaction. This indicates that inspectors will be removed from the behavior networks while employees of a lower status will be central figures within the behavior networks.
