The Changed Conversation About Surveillance Online
The revelations about sustained and pervasive NSA surveillance that started in June and are still ongoing are having a sustained effect on the global conversation about censorship and surveillance on the Internet. Using 'big data' analysis of public sentiment we can illustrate the change in this conversation. The increase in online posts on the topic important because it indicates that the global conflict over Internet norms is occurring in real-time and is not fading out of the global consciousness or being consumed by a Huxleyan dystopia. If anything, the conversation has increased and has remained significantly higher than it was in the months prior to the first wave of revelations.
Our analysis, using Crimson Hexagon's media monitoring platform, comprised more than 2.5 million online posts from April 6 until October 1 from almost every country in the world. The conversation was and remains dominated by the expected players (largely due to an English language bias or our survey), United States, United Kingdom, Canada, and Australia. Even though we conducted our survey for English language posts only, the global geographic location of posts accounts for nearly all nations around the world. The below images illustrates the post density and posts per capita globally.
To analyze the relative volume of conversations globally on the topic of surveillance and censorship we established June 6th, 2013 as the start point of our analysis. This was the date of the first publication of NSA leaks by Edward Snowden in the Guardian newspaper. First, we established a two-month baseline prior to the leaks to assess the volume of online content related to censorship and surveillance. Second, we collected data for the four months following the June 6th publication. Third, we created a dummy variable indicating pre- and post-NSA revelations. Lastly, we ran the data as a Newey-West time-series analysis using days as the unit of time comparing the volume of posts pre- and post-leaks.
What we find is a significant increase in the volume of conversation at an alpha of .01. This indicates that if we were to repeat this test we would achieve the same result 99 times out of 100. The image below illustrates the conversation associated with surveillance and censorship globally. The primary line (blue) represents changes in posts from day to day. While the lower two dashed horizontal lines represent the mean (blue dashed) and median (red dashed) associated with the conversation prior to the first publication of leaked information and the two upper dashed lines represent the mean (blue dashed) and median (red dashed). The vertical red-line illustrates the date and point in time in which the story first broke on the Guardian’s website.
There are several caveats that should be noted in our analysis. First, while the initial leaks through the Guardian and the Washington post were the bombshells that clearly instigated a global online conversation, subsequent revelations and publication of information about activities of the NSA and that of other country's intelligence agencies have also occurred that are likely causing spikes in content. This has had the effect of maintaining the story in the public consciousness and bolsters the mean and median of content volume. However, absent these subsequent revelations, the evidence suggests there would still be a significant and sustained bump in the volume of content postings associated with censorship and surveillance. Second, regardless of the content volume bump, global public attitudes towards surveillance and censorship on the Internet are highly negatively skewed as the image below illustrates.
The main takeaways of our analysis should be that there is a global conversation happening that is affecting how people around the world think about topics surveillance and censorship and that the sentiment towards these issues is highly negative.