Social network processes are thought to be key to understanding diffusion of innovation and adoption of new behaviors with regard to health. The current project will establish a unique source of individual level social network data, addressing fundamental problems in conventional social network research designs. It will have wide ramifications for epidemiological, public health and social analyses. We will field a longitudinal panel survey in two waves in a rural population in Senegal, West Africa. After extensive qualitative analysis, survey data on social networks and beliefs concerning health, illness and medical treatment will be collected from 2500 respondents in two survey waves, separated by 2 years. Experimental tests for the accuracy of the network data collected will be conducted between waves of the main survey. These will yield insight into the extent of biases in analyses using conventional social network research designs. The fundamental innovation of this project is in linking network members cited by survey respondents to longitudinally collected data concerning their health behaviors in a surveillance system (which is a small scale population registration system). This will allow us to gather data on up to 5 times more network members for each survey respondent than in conventional research designs, which usually constrain the size of social networks they collect data on to a few individuals for logistical reasons. It also gives us more accurate information on wider range of network members'health behaviors (such as their use of maternity clinics, anti-malarial prophylaxes and medical care for children among many other things) over time than is conventionally collected. For these reasons, this data will make possible better estimates of the effects of the structure of network ties, social learning and social influence on health than were previously possible. An added benefit of the present research design is that it also allows us to spend valuable survey time and resources collecting data concerning the strength of ties between respondents and their social network members as well as ideational elements such as attitudes and beliefs that are critical to the diffusion of information and adoption of new health technologies and behaviors. A sample of 1500 respondents will be taken from a census of 3 contiguous villages, enabling us to capture extensive inter-relationships of social network ties. The additional 1000 respondents will be drawn randomly from the larger population. This latter sample will help us refine estimates of effects of network structure on health that are generalizable to the population as a whole.
Social network processes have been recently been seen as key to understanding a wide range of health and health related behaviors. By linking individuals'social network members to extensive health information already collected on them in an ongoing health surveillance system, we redress fundamental problems in previous social network research designs that may lead to erroneous conclusions concerning the causal effects of social networks on health. This project will produce the highest quality and most comprehensive social network data related to health outcomes to date, allowing the scientific community to test for effects on a wide range of health and biomedical phenomena of social network structure and interpersonal influence.
|Delaunay, Valerie; Douillot, Laetitia; Diallo, Aldiouma et al. (2013) Profile: the Niakhar Health and Demographic Surveillance System. Int J Epidemiol 42:1002-11|
|Sandberg, John; Rytina, Steven; Delaunay, Valerie et al. (2012) Social learning about levels of perinatal and infant mortality in Niakhar, Senegal. Soc Networks 34:264-274|