METHODOLOGY This report uses Burmese and English language material, including Ma Ba Tha statements, publications, videos of sermons, local and international news reporting, as well as an analysis of content and imagery disseminated on social media. We have supplemented this desk research with a field visit to Thailand and Myanmar in 2015, and have conducted over 100 remote and on the-ground extended interviews with editors and senior journalists of several leading Myanmar-based media organizations, anti-hate speech and human rights activists, and financial and real estate professionals. This report is primarily focused on examining online hate speech in Myanmar; as such, we have made significant use of social media to collect primary source material. Social media is extremely important in Myanmar’s media and communications landscape, where it is often one of the primary outlets for news and information and a key origin point for a significant amount of unique content that can later become the basis for news. Social media is particularly relevant to our research as a key disseminating channel for hate speech and hate speech related content, including material that can trigger large-scale mob violence. We began our analysis by constructing a dataset of 100 Facebook accounts. These included as many monks on the Ma Ba Tha Central Committee or in leading chapters as we could identify. From these, we expanded to other accounts, including those of monks, politicians, activists, government officials, and laypeople who serve as key disseminators or consumers and followers of the Ma Ba Tha monks’ content. Over a period of three months, we actively examined the public content on these accounts, as well as their affiliated websites and blogs. We paid particular attention to posts that contained hate speech and any available data on Ma Ba Tha and 969 activities. Of particular interest were images of activities such as conferences, rallies, and meetings and primary source documents, such as uploaded copies of Ma Ba Tha statements, sermons, donation receipts, and publications. We acknowledge some limitations in this course of research. Myanmar constitutes an inherently difficult data environment. The country ranks 144 out of 180 countries on the 2015 Press Freedoms Index and has only recently lifted draconian limitations on freedom of expression. Independent media is still in a nascent phase and the online space is rife with rumors and misinformation. This is particularly true on social media where data is additionally raw and un-vetted. To mitigate this, we have attempted several controls. First, we have only selected accounts that at a high degree of probability represent personal or official accounts of known personalities, or are highly relevant to our study as significant disseminators of hate speech. Second, we have sought to use a variety of sources of data to corroborate key facts. For example, imagery from Ma Ba Tha events found on accounts within our sample can provide examples of banners, signs, faces, and other identifying characteristics that can clarify the nature of the event, its location, attendees, and any other relevant information, and can be found or recreated in other forms of reporting. Finally, we acknowledge the difficulties in generalizing “narratives” and have tried as much as possible to avoid “cherry picking” information and to primarily use content that is representative of themes and references we have repeatedly observed, or that have been shared very widely. All of our social media data analysis has been manually hand-coded without the use of any automated data collection or “scraping” technology. 5