To change the beliefs of an entire community, only 10 percent of the population needs to become convinced of a new or different opinion, suggests a new study done at the Social Cognitive Networks Academic Research Center (SCNARC) at Rensselaer Polytechnic Institute. At that tipping point, the idea can spread through social networks and alter behaviors on a large scale.
The research is still in its early stages, and it's uncertain if the results will apply to all kinds of beliefs, particularly in tense political situations.
But the findings do provide insight into how opinions spread through communities. The model may also help experts more effectively quell misconceptions and influence the choices people make about public health behaviors and related issues.
"This is really a starting point to understand how you can cause fast change in a population," said Sameet Sreenivasan, a statistical physicist and one of the co-authors of the study which included two graduate students and three senior faculty.
"The real world has a lot more complexity, obviously," he added. "But one of the things you can take away is that if you want to cause a fast change, there is an upper bound to how many people really need to commit."
Previous modeling studies have treated opinions like infections, Sreenivasan said. They assumed that beliefs would spread like epidemics, but that's not really an accurate depiction of the real world. He and his colleagues took a different approach.
The researchers, led by Prof. Boleslaw Szymanski, director of SCNARC, and Prof. Gyorgy Korniss, Associate Professor of Physics at RPI, started by creating three social network models each with groups of people with different links between them. In the first, everyone was connected to everyone. In another, everyone had the same average number of connections even though not every person was connected directly to everyone else. The third network included "opinion leaders," who had more connections than others did.
At the beginning of the experiment, every node in each social network was set to share one opinion. The researchers also set the model to follow a specific set of rules. For example, "listener" nodes adopted the ideas of "speaker" nodes. And, once listeners adopted a minority opinion, they became inflexible and unshakable. They wouldn't change their opinion again, but they could influence those they came in contact with.
When the researchers introduced a new, different opinion to a very tiny fraction of nodes, it took an exponentially long time for that belief to spread through the population. But when between seven and 10 percent of the nodes became convinced of that second opinion, the researchers reported in the journal Physical Review E,that there was a sharp drop in how long it took for the whole community to come to consensus. Rapidly, they all began to believe in that second opinion.
The exact proportion needed to reach the tipping point depended on the type of network involved. But all three types of network models produced the same conclusion, said Albert-László Barabási, director of the Center for Complex Network Research at Northeastern University in Boston.
"The main message is that each opinion has a critical point," Barabási said. "If the fraction of the population is under the predicted threshold, the minority can prevail, but it will take astronomically long times to do so. However, once the number of people who are unshakable in their opinion reaches the critical threshold, the time it takes for the majority to accept their point of view becomes rather short."
"It also shows that minorities can prevail in their opinion only if they strive to become less of a minority by turning a small fraction of the population to absorb their opinion," he added, "becoming unshakable supporters of their point of view."
The findings are not likely to apply to bipartisan debates about issues like debt ceilings and tax cuts. One reason, Sreenivasan said, is that those situations often involve unshakable opinions on both sides.
Instead, the study might be of use in public health campaigns, such as efforts to get people in Africa to start using bug nets to prevent disease.
Sreenivasan mentioned that Dr. Szymanski was contacted by people from Colorado who are trying to replace fogging of insecticides damaging the environment with the Integrated Pest Management but run against a long-held belief that smelling and seeing the toxic chemical was the only proof that something was being done about the bugs.
"We suggested," Sreenivasan said, "that instead of trying to convince everyone, it might make the most sense to target selectively the people who are open-minded enough to hear out the evidence and make up their minds rationally."