X-Rotary (8): Meeting Networking
This is the first installment of a series of essays that attempt to apply Social Network Analysis (SNA) methods to the study of the Shanghai Rotary Club, with a view to detect relational patterns within the club. Previous essays proposed statistical and spatial analyses of membership based on rosters of members. Yet this approach remains static and disembodied. It tells nothing about how members actually participated in the club’s activities and how they interacted with each other. Moreover, relying solely on rosters leads us to exclude non-members who potentially attended club meetings.
In this new series, we supplemented rosters by newspapers in order to bring the club to life and to shed light on members in action. The Shanghai Rotary Club hold regular meetings every week on Thursday at lunch time. These “tiffins” usually consisted in luncheons and lectures covering a wide range of topics (from business to science, education or international relations). Occasionally, the club also organized special events (farewell dinner, Christmas party, recreational outing). These meetings were widely advertised in the local press. Newspapers reports usually list the participants and the roles their performed (a minima). At best, some articles provide very detailed information on attendees’’ behaviors.
Since regular attendance to weekly meetings was a duty for every member, we focus on meeting attendance as a starting point for this research. We argue that club meetings offer the best sites to identify club members and to observe concretely how they participated in club activities. Many Rotarians, however, did not appear in newspapers. Only 36% of members listed in the roster issued in 1930 appeared in newspapers, and 44% in 1940. Missing members either did not attend or were just omitted in press reports. In both cases, their absence suggests weak membership. On the opposite, newspapers provide information on participants who were not members of the club, but occasional visitors, special guests or invited speakers. These persons offered the Rotary an entry point to the outside world.
In this series, we propose a multimodal approach to persons in networks, in which persons are conceptualized as club members or meeting attendees.. In this approach, persons are treated alternatively as nodes and edges. We pursue two main purposes. Substantially, how did participants behave during meetings? What roles did they perform? How did they (re)shape club rules and conventions? How did they interact with each other? How did roles and behaviors translate into positions in networks? Conversely, can we infer participants’ roles and status from their positions in networks? Did personal attributes (status, nationality, gender, profession) affect social behaviors locally (members’ positions) and the structure of the network globally (i.e. Rotary’s organization) over time? Methodologically, what can we learn from a multimodal approach to persons in networks? Can affiliation (two-mode) networks bring a new light on members’ status and personal behaviors in social life? What can we gain from, and what are the challenges of applying Social Network Analysis (SNA) techniques to historical newspapers – a source that has remained neglected in the current SNA-driven historical scholarship to date?
Methodologically, we proceeded in several successive steps. We first constructed three sample datasets extracted at different point in time. Each sample covers a one-year period which corresponds to a presidential term selected for its historical and documentary significance. The first sample under Petit’s presidency covers the club’s first year of existence (1919-20). It was a decisive period for setting club rules and members’ roles. The second sample (Fitch, 1930-1) marks a period of intense activity and a turning point in term of membership, with an increasing number of Chinese members joining the club after the establishment of the Nationalist regime in Nanjing (1927-1937). It also the best documented period in both newspapers and rosters. The last sample (Harkson, 1938-9) is situated during the Sino-Japanese war (1937-1945). It is a critical juncture in the local history of the club and of Shanghai more generally, as many foreigners began to leave the city and as Japanese membership became problematic. This sample will serve to address the issue of whether the war disrupted club’s life and members’ roles. While newspapers reports became scant, we can rely on a richly detailed roster of members for this term. For each sample, we extracted the following information. From newspapers, we extracted the dates of meetings, the names of participants and the roles they performed during meetings (used as basic information for identifying nodes and qualifying edges). We also coded the type of events (regular meetings, special events, outing, etc.) and topics of discussion (used as events attributes). From the rosters, we extracted additional information on members’ status, nationality, and social-professional occupations (used as persons’ attributes). The attached table compares the data available for each period.
Drawing on these data samples, we built four sets of networks. Starting from a bimodal network of meeting attendance (persons-events joined by roles), we first propose a direct-method analysis aimed at identifying the most important persons and events (nodes) and at examining the specific roles participants performed during events (edges). The second step consists in projecting the initial two-mode network into two one-mode networks of persons-persons joined by events (co-attendance) and events-events joined by persons (multiple attendance). In order to supplement the bimodal analysis, we offer a dual projection approach, which consists in jointly analyzing the two projections with a view to offset the loss of data resulting from the projection process (Borgatti 1997, 2013). In the next essay, we examine the co-attendance network in order to refine our analysis of personal profiles, to identify most frequent co-attendees and communities based on co-attendance. The next essay focuses on the networks of multiple attendance. The main objective is to investigate how persons (participants) could serve as links between events. We eventually built a network of roles joined by persons in order to examine more closely how participants could perform multiple roles, to detect specific patterns of role combination and to how these patterns changed over time.
To be continued….
Alonzo Bland Calder papers (1911-1956), [Box no.1, no.19], Hoover Institution Archives, Stanford, California. https://oac.cdlib.org/findaid/ark:/13030/tf0w10012w/entire_text/ [Last accessed August 29, 2019]
ProQuest Historical Newspapers™ – Chinese Newspapers Collection https://www.proquest.com/products-services/hnp_cnc.html [Last accessed August 29, 2019]
Borgatti, Stephen P, and Martin G Everett. “Network Analysis of 2-Mode Data.” SON Social Networks 19, no. 3 (1997): 243–69.
Borgatti, Stephen P, Martin G Everett, and J. C Johnson. Analyzing Social Networks. Los Angeles, Calif.: Sage, 2013 (chap.13).
Doran, Derek. Network Role Mining and Analysis. Cham: Springer, 2017.
Graham, Shawn, Ian Milligan, and Scott Weingart. Exploring Big Historical Data: The Historian’s Macroscope. London: Imperial College Press, 2016. http://www.themacroscope.org/ [Last accessed August 29, 2019]
Newman, M. E. J. “Analysis of Weighted Networks.” Physical Review E 70, no. 5 (November 24, 2004): 056131.
Opsahl, Tore. “Projection.” Tore Opsahl (blog), June 21, 2011. https://toreopsahl.com/tnet/two-mode-networks/projection/ [Last accessed August 29, 2019]
Sauer, Nils Christian, and Simone Kauffeld. “The Ties of Meeting Leaders: A Social Network Analysis.” Psychology 6, no. 4 (March 3, 2015): 415–34.