After two years of operating the Una Hakika information service, we are ready share our research findings is the “Una Hakika: Mapping and Countering the Flow of Misinformation in Kenya’s Tana Delta” released today. When our team first visited this part of Kenya that was deeply impacted by interethnic massacres in late 2012 and early 2013, we saw how rumours contributed to the atmosphere of fear, distrust, and hatred that enabled the conflict. This often false information circulating through the area ranged from seemingly mundane reports of Orma herders trespassing on Pokomo farms and damaging crops to extreme stories of outside actors supplying thousands of weapons to drive further violence. Ultimately, what mattered was not always the truth of the rumours but how much that people believed them. As we started developing possible solutions to the problem of misinformation in Tana’s conflict, it became clear that we had a lot to learn, which is why Una Hakika became not only an impact-focused project but also a research project with the support of the International Development Research Centre and the collaboration of iHub Research. Along the way, our work in the Tana Delta grew into the beginning of something more – the first phase of a much larger, longer-term effort which we aim to scale up into other parts of Kenya and beyond.
Overall, Una Hakika Phase 1 has been a great success that really highlights the potential of information and communications technologies (ICTs), especially mobile phones, to bridge information gaps in less-developed areas; monitor and counter the spread of incendiary rumours with the help of tools such as our purpose-built WikiRumours software; and contribute to stability, security, and peacebuilding in conflict areas. We encourage readers to read and distribute the full report but the following is a summary of key results for quick review.
The following numbers illustrate the story and success of Una Hakika Phase 1, especially how we built its reach in the Tana Delta, filling an important information gap left by the lack of local media, and became one of the most trusted information sources in the area.
- 71% of survey respondents said that misinformation contributed to the violence
- 369 unique rumour investigations and community interventions conducted
- 200 volunteer community ambassadors trained
- 1 in 15 adult mobile phone users in the project area subscribed to Una Hakika
- 1,591 direct subscribers
- 8,710 beneficiaries estimated by household size
- 45,000 indirect beneficiaries estimated by subscriber sharing behaviour
- 30 people on average receive shared updates from subscribers
- 1,200% improvement in women reporting good access to information
One of the most useful and interesting results of Una Hakika Phase 1 is an improved understanding of how rumours spread, especially the ways that various factors influence this. Geographical, demographic, cultural, informational, and technological characteristics of an area all shape where rumours spread, how quickly they get there, and how people react to them. Ultimately, one of our goals is to know enough to conduct predictive analysis of rumours such that their spread can be anticipated and misinformation proactively contained by misinformation managers. The following are just a few of the things we learned about these factors.
- Telecommunications infrastructure – Mobile communications have not replaced existing social mechanisms for information sharing and many Tana Delta communities still hold public meetings and their residents still have casual conversations or learn of events and news from truck and motorcycle drivers passing through their area. The Una Hakika team’s observations indicate that mobile communications actually graft onto these existing social mechanisms more often than replacing them. Social information relays are complemented by the rise of mobile networks, increasing the range of information made available through traditional methods and expediting its proliferation.
- Road networks – Several cases have shown how the spread of rumours follows roads, which obviously facilitate the movement of people and therefore information. Despite the near ubiquity of mobile phones in the Tana Delta, many rumours have closely followed the paths of main and secondary roads. It is also possible that people simply have stronger social connections with people in other villages that are physically more accessible from their own so that even when they do use ICTs for communications they do so mostly with people living along the same roads.
- Terrain features – Terrain features such as rivers and forests can act as either barriers or conduits in relation to information flows. For example, the Tana River generally acts similar to roads in that it serves as a transportation network for people and goods, thus facilitating the sharing of news by word-of-mouth between relatively distant communities that may not otherwise be in regular contact. However, in some locations the river can act as either a physical or symbolic barrier between different communities, thus hindering the flow of information, including misinformation.
- Seasons and climate – The Tana Delta experiences varying seasons of heavy rains and dry periods, which influence agricultural and animal herding activities (the primary occupations of most residents) as well as impacting physical mobility and telecommunications infrastructure. Together, these factors in turn influence the frequency of rumours and the general transmission of information. In addition to potentially discouraging actual violence, difficult conditions during wet seasons may also hinder the word-of-mouth transmission of information between distant locations since people stay closer to home. Another very significant influence upon rumour origination and transmission is the fact that during land preparation and harvest times the majority of people are generally busier working long days and do not have the spare time or energy to create or share the speculative information that becomes rumours. Quite simply, as one elder living in the Tana Delta explained to the Una Hakika team, “People are too busy and tired to gossip.” This means that while very alarming news of severe events (e.g. conflict or terrorist attacks) will still inspire rumours, less dramatic topics are generally ignored and less likely to be reported at such times.
- Cultural factors – Like much of Kenya, the Tana Delta is an ethnically and religiously diverse place and these factors influence how people acquire, think about, and share information. The ethnic factor was partially quantified during the baseline survey conducted in early 2014, which found a significant disparity in how much that respondents reported trusting information depending on the ethnicity of the person from whom they receive it. No specific ethnic group was distrusted more than another yet the majority of respondents cited less trust in ethnic groups which were not their own. By the time of the final survey in September 2015, the results indicated an improvement in the level of trust between ethnic groups in the Tana Delta. In addition to ethnic factors influencing information sharing between groups, different cultures in the Tana Delta also have different attitudes towards authority and questioning information so in some communities people who are skeptical of a given rumour may be less likely to share these doubts, thereby facilitating the spread of rumours.
In addition to the more academic findings of Una Hakika Phase 1, our team learned several important operational lessons through experience. The following are a few of the highlights.
- Misinformation management systems cannot be imposed from above. Instead, they must be implemented by entering into communities using culturally relevant introduction processes followed by cooperative efforts.
- Misinformation management efforts should not try to replace existing communication practices in a new environment but rather adapt to the variety of high tech and low tech methods already used in the area. Una Hakika anticipated obstacles which could prevent participation and devised solutions to maximize inclusiveness.
- Establishing and maintaining trust is one of the most important but also most difficult components of a misinformation management project. A project that is not transparent, honest, and fair will immediately lose the majority of its value to the community and the pursuit of peace. This requires maintaining clearly stated and dutifully followed objectives, actively working to include a diverse group of participants, and being aware of how the actions of the project or its staff might impact perceptions of neutrality.
- Some of the main beneficiaries of projects like Una Hakika are the most disadvantaged members of a population. Women, youth, and marginalized communities have the most to gain from a system which can circumvent many of the social, cultural, and structural barriers that prevent them from accessing reliable information.
One of the key components of Una Hakika is countermessaging, the dissemination of updates to subscribers in order to stop the spread of rumours and correct misinformation. This work requires a nuanced approach, especially when the subject of a rumour is particularly sensitive. For example, Una Hakika is committed to transparency and reporting the truth back to subscribers but there are cases where an inflammatory rumour (e.g. of a violent incident) is verified and found to be true, which is a situation that may actually contribute to conflict. Such situations must be treated with sensitivity and so the Una Hakika team has developed several tactics for dealing with them.
- Time delays – Subscribers reporting rumours are always given a time frame within which to expect a verified response but inflammatory cases may require the inclusion of a small delay in the verification and countermessaging turnaround time in order to allow for tensions to slightly subside, project staff to gather additional information, and to consult with relevant stakeholders on how best to approach the situation.
- Contextualization – An important aspect of the Una Hakika approach is for the counter-messaging which is ultimately sent to report back to subscribers not only verified factual information but also contextual details and possibly also instructions for relevant stakeholders on how to de-escalate tensions. For example, violence may have actually occurred in a particular village but was of a small-scale interpersonal nature. A common pattern is that by the time rumours of such incidents reach more distant locations the size of the conflict and its impact may have grown into reports of a battle with several casualties with its nature distorted into an intercommunal conflict.
- Geographical containment – It is important for Una Hakika to minimize the risk of inadvertently spreading rumours in the process of trying to contain and dispel them. For this reason, potentially incendiary verified information is only sent back to areas from where a given rumour was initially reported. For example, if there are three villages near each other and residents of village A report rumours that village B has been attacked, the relevant countermessaging will only be sent to Una Hakika subscribers in village A since subscribers in village C may not have heard the rumour and so there is a risk that some of them could pay attention only to the portion of a countermessage that reports a new rumour while ignoring the part that dispels it.
With the above tactics and examples in mind, Una Hakika countermessaging has been designed to always have the following characteristics.
- Responsive – Una Hakika provides information primarily in response to what subscribers are reporting and requesting, thus reacting to community needs rather than simply pushing out information in a one-way flow as a traditional information service might.
- Targeted – By only sending countermessaging back to specific locations where rumours have been reported Una Hakika minimizes unintentional spread while also ensuring responsiveness.
- Sensitive – Una Hakika takes into account local dynamics and engages relevant stakeholders and appropriate communication channels at all stages of its process but especially during verification and countermessaging.
- Timely – The Una Hakika team prioritizes and verifies rumours as quickly as possible (except for in cases where intentional delays may be introduced) since information can travel quickly through the project area, especially with the aid of information and communications technologies.
One of the main findings during Una Hakika Phase 1 has actually come from beyond the Tana Delta. This is the observation that misinformation poses a major problem in both developing and developed countries worldwide not only in driving conflict situations but also by hindering public health, governance, and other development efforts. For example, during the response to the recent Ebola outbreak in several West African countries, false rumours about the harmful intentions of public health workers and medical staff actually led to violence in some cases, thus hindering the public health effort and compounding it with a security dimension. With examples like this in mind, the Una Hakika team recommends that misinformation management mechanisms should also be integrated into any major international development, governance, public health, or peacebuilding effort. The work done during Una Hakika Phase 1 is an important first step towards identifying the knowledge, tools, and techniques required to do such work effectively. Other researchers and organizations should similarly devote resources to conducting further research and testing in a variety of contexts in order to collectively build up a community of practice around misinformation management.
Going forward, the Una Hakika team has ambitious plans to build upon the success of Phase 1 in the Tana Delta, especially in terms of geographical expansion. Kenya’s next general election will be in late 2017, which promises to be a time of heightened political and ethnic tensions, especially since the prevalence of misinformation will likely increase, a situation only worsened by the tendency of many Kenyan politicians to deliberately use false information in advancing their goals. With this situation in mind, the Una Hakika team is working to gather the resources needed for a scaled-up Phase 2 which will not only continue operating in the Tana Delta for at least another two years but also expand to cover Kenya’s entire Coast region as well as other key areas around the rest of the country. Outside of Kenya, preparatory work is already underway to set up a similar system in Mandalay, a major city in Burma (Myanmar), where misinformation has contributed to intercommunal violence, especially against the Muslim minority; this will serve as the foundation for further operations in other parts of Burma. In addition to their likely positive impact in terms of preventing violence, the concurrent operations in Kenya and Burma will enable further research looking at how the Una Hakika model works in sharply contrasting contexts, thus leading to the further testing and refinement of misinformation management knowledge and tools as well as a framework to guide this type of work. While concentrating on Kenya and Burma, the Una Hakika team will remain open to opportunities for expansion wherever else such a service may be required. Towards this end, the team is working on developing a toolkit which will enable others to deploy independent implementations of the Una Hakika model with the benefit of Sentinel Project guidance.