Although smoking rates have decreased in many parts of the globe, there are still an estimated 1.3 billion smokers globally, and tobacco use is still the leading cause of preventable death worldwide. To combat the ongoing problem of tobacco use, the WHO Member States negotiated and unanimously adopted the WHO Framework Convention on Tobacco Control (FCTC) which is an international treaty aimed at reducing tobacco use. About 87% of the 193 WHO member countries have ratified the FCTC (or its legal equivalent) as of April 2011. Understanding when, how, and why individual countries ratified the FCTC and subsequently enacted policies implementing the treaty obligations is a critical research question that has implications for understanding network effects, the science of diffusion of innovations, as well as prospects for enacting future international public health law. This study proposes to compile extensive network data from GLOBALink, an electronic forum for global tobacco advocacy, to construct multiple social networks that can be used to estimate network effects in a dynamic modeling framework. We study network influences on the diffusion and adoption of the FCTC and the strength of its implementation among the 193 member countries over a 9-year period. We also propose to augment the network data with an extensive database of country attributes such as total population;population distribution;size (area);income;income distribution;tobacco production;governmental form;region;spending on power, communications, and trade;existence of tobacco control NGOs;and so on. These attributes will be used to control for their possible associations with FCTC ratification and implementation to determine whether these characteristics are associated with infectivity or susceptibility to FCTC diffusion. We have also acquired data on the ratification of two other international treaties which diffused over a similar time-span. These treaties will act as """"""""comparison treaties"""""""" in which we do not expect GLOBALink participation to be associated with their ratification. At the same time, we will acquire other inter-country network data to test whether these other networks (trade and communication) were associated with the comparison treaties and not with FCTC adoption.
The final aim combines the traditional diffusion network approach with the newly developed actor-oriented MCMC approach created to address network dependencies. We have built a diffusion network stochastic actor oriented (SIENA) platform in R to test specific dynamic network influence hypotheses. This event history analysis will enable the testing of network influences over the 108 months of treaty adoption that incorporates parameters for reciprocity, transitivity, alternating k-stars, and the many other parameters built into RSIENA. These procedures enable the specification of complex network effects that dynamically consider the coevolution of networks and behavior. Finally, we will supplement our analyses with selected content analysis of messages exchanged between GLOBALink users. Our overall goal is to compile a rich dataset on the international diffusion of three treaties expected to be differentially influenced by at least 10 different networks. The significance of thi proposed research is derived from addressing an important global health topic, retrieving data of unprecedented richness and granularity, and employing a sophisticated yet new comprehensive modeling approach.
This study proposes to compile extensive network data from GLOBALink, an electronic forum for global tobacco advocacy, to construct multiple social networks that can be used to estimate network effects in a dynamic modeling framework. We propose to compile at least 10 different networks to study their influences on the diffusion and adoption of the Framework Convention for Tobacco Control (FCTC). In addition, we compile data on the diffusion of two comparison treaties, one environmental and one legal. We use a dynamic modeling approach combining the diffusion network perspective with the stochastic actor oriented model that enables the specification of a new class of dynamic network effects specified by a set of dynamic network-behavioral hypotheses.
