These data are based on our paper:
Eitan Muller and Stefan Stremersch that appeared in Marketing Science, 2010. As discussed in the paper, we constructed two (linked) data-sets: Authors Database and Links Database. Please note - the following are links to large zipped files (16 MB and 28 MB respectively) each containing 37 Excel files for the years 1973-2009. When clicking on the underlined links you will be given the option to save the files on your hard drive.
Authors Database includes author (with his/her ID); Average separation; Cumulative number of papers; Cumulative number of journals s/he published in; First year of publication and last year of publication. Links Database includes two authors (with their IDs); Cumulative number of joint papers; List of journals at which they published their joint papers.
Main component and average separation
In order to get familiar with the database, and to allow for some quick yet interesting observations, observe the following figure that shows the evolution of the collaboration network over time, for two focal concepts: (1) the growth of the main component over time as percent of the main component size in 2008; and (2) average separation.
The numbers that are presented are from the years 1979-2008, as the size of the main component before that was very small, and may not be a good indication of the network parameters that we examine. The line labeled “main component” presents the size of the main component through time, as a percentage of the size in 2008. We see that the size of the main component grows in a nearly linear trend from the mid 1980’s. Our data also show a consistent growth in the percentage of the main component out of all researchers in our database, from about 23% in 1979 to about 69% in 2008. Thus, the collaboration network in marketing seems to show increasing marketing research coherence with time. An interesting result regards the declining average separation, also shown in this figure. This finding is surprising given the increasing size of the system, as network models generally predict that average separation should increase with network size. In this case, in 2008, network size is more than 25 times that of 1979; yet average separation declined more than 35%. This intriguing phenomenon which may be related to the changing collaborating patterns over time is yet to be explored.
We find the closeness centrality of all players in this network going back to 1970's and calculate the researcher with the lowest centrality of each year: Roland Frank, Paul Green, Vijay Mahajan, Wayne Desarbo, and Donald Lehmann. An interesting observation regarding the central egos relates to their entrance and exit from that role. One way to gauge this phenomenon is to examine the centrality of the five central egos over time. However, if the average centrality across the entire network changes over the years (which happened in the marketing network), then changes in the centrality of those egos might reflect network players, and not necessarily personal status vis a vis the network.