Information Transmission of a Policy Program: Models for the Optimal Detection of Key Players

Abstract

The underlying motivation for this work stems from the observation that there was weak participation in a policy program to modernize the commerce of a city center. This was due in part to a poor performance from the trade association board in the transmission of information. Using the tools of social network analysis and combinatorial optimization, we search for new sets of key players that are better positioned to disseminate information in the collective. We detect 2 new sets of key players and compare them with the trade association board. The comparison shows that social network analysis and combinatorial optimization can be useful tools in making policy implementation processes more effective.

Keywords:

• combinatorial optimization
• commerce
• information transmission
• key players
• policy
• social networks

ACKNOWLEDGMENTS

Early versions or portions of this article were presented at the Seminar “Networks, Markets, and Organizations,” QMSS 2 sponsored by ESF, at the University of Groningen in August 2009 and at the 1st ICC Conference on Network Modelling and Economic Systems (NMES 2008), ISEG/Lisbon, October 9–11, 2008. We are grateful for comments by participants. We also thank Raquel Rego and the anonymous reviewers for comments on an earlier version of this work, and Ann Henshall for the English revision.

Notes

¹This specific program was called Urban Commercial Project and was part of a larger program called PROCOM. This case study took place from 1999 until 2001.

²This particular policy program, unlike most, even sponsored a diagnostic study prior to intervention, but its low quality impeded a good knowledge of the concrete social system of the collective studied.

³We differentiate key players from opinion leaders as these are not necessarily optimally positioned.

⁴Obviously, selectivity requires information about how interested each person is in the collective good, as well as their personal level of resources as an indicator of how much each is able or likely to give. This also suggests that the organizers need to know how to craft messages that focus on maximizing benefits for contributors while minimizing their costs (Marwell & Oliver, 1993).

⁵In the larger work (Varanda, 2005), of which this article is part, we show that relational measures built based on indirect links did not significantly explain behaviors like participation in the program, while measures using direct links, such as degree or cliques, did.

⁶These two board members were not eligible for the program because in one case the board member's business was located outside the city center and in the other case the member's field of business was not allowed in this specific program. These two members were not included in the sociomatrix, which was not considered problematic as they were not cited by any of the eligible traders.

⁷For more information on the homophily tendency, see Lazarsfeld and Merton (1954). For a more recent approach and literature review concerning this concept, see for instance Mcpherson, Smith-Lovin, and Cook (2001).

^a Since one shop had opened very recently, there are no data for the evolution of performance.

⁸Through field work we found out that he was not involved in the city's commerce and his participation on the board was also diminishing. For instance, he was frequently absent from meetings.

^a Since one shop had opened very recently, there are no data for the evolution of performance.

^a Since one shop had opened very recently, there are no data for the evolution of performance.

^a Since one shop had opened very recently, there are no data for the evolution of performance.

^b Two board members have their shops outside the program intervention area.

^c Information not available for three board members.