Page - (000157) - in Knowledge and Networks

150 Analytical Strategy Classifying Trajectories My first concern was to gain an understanding of different trajectory types by exam- ining relevant factors. I selected three aspects: lifespan, average betweenness cen- trality, and the number of periods of membership in the network’s core. In this section I explore each of these dimensions in turn. Probably the most critical indication of an artist’s success is the ability to earn an adequate income in the chosen artistic field. Zuckerman, Kim, Ukanwa, and Rittman (2003) show how novice film actors frequently submit themselves to typecasting in order to establish a foothold in the industry. Even if an artist is able to enter the industry, survival is never entirely guaranteed, because fads and shifts in fashion can impact an artist’s fate (Hirsch, 1972/2011). Hence, one’s lifespan is a measure of success in a networked industry. Better connections with other players increase the likelihood of an artist being able to exploit resources. On the other hand, higher coupling to the industry struc- ture can constrain the artist’s action and creativity. Nonetheless, more resources and information generally increase the odds of an artist being employed in the most attractive opportunities (Granovetter, 1973/2011). The method I chose to measure the resources available to an individual was to calculate the average betweenness centrality across his or her lifespan in the field because it indicates the extent to which an actor functions as a broker for a network’s paths (Hanneman, 2001). Kadushin (2004) observes that the stability of a field can be grasped by the com- position of a network’s core. A polarized core can trigger social changes. Core membership empowers its holders to influence a field’s rules. Core members are able to establish a consensus among different actors. Conversely, membership in the core provides the opportunity to disrupt the current consensus and to fight to estab- lish new rules. It can be asked whether membership in a network’s core does not automatically entail a high betweenness centrality, making dimensions such as betweenness and core membership redundant. This correlation would be high for star networks in which the center is occupied by a single member and all peripheral actors are connected only to the center. In more decentralized networks, peripheral actors might control critical resources despite their positions. Conversely, novice actors might have a very early opportunity to develop a project with a core actor. In doing so, they would become central actors as well. However, if their only connec- tion is to their “godfather” at the core, they have relatively less freedom of action than better connected peripheral actors. I performed a K-means cluster analysis using the variables average betweenness centrality, lifespan (periods), and number of periods at the core to obtain six clus- ters of musicians (see Table 8.2). The limitation of the period of analysis to the years 1930–1969 resulted in left and right censoring effects, given that it therefore did not encompass the entire history of jazz. For instance, the trajectory of a musician that began in 1925 would have a longer lifespan than the trajectory captured in my anal- ysis. In order to mitigate left and right censoring effects, I excluded from the trajec- C. Kirschbaum