Measuring and quantifying the silent majority on the Internet

Increasingly, researchers are mining social media as a way of inferring public sentiment from the Internet. While ways to measure social media participation are beginning to be well established, there are little ways to measure the silent voice of a passive consumer. Passive consumers refrain from actively offering opinions, and as such cannot be completely quantified using social media crawls alone. Passive consumers account for a significant proportion of the casual Internet user, often amounting to over 90%. We argue that mining the Internet for public sentiments should quantify and consider both active and passive participation. This paper examines various ways of inferring passive participation by investigating methods to measure and quantify these silent voices using publicly available information on the web. Specifically, we take a closer look at two approaches: (i) measuring loading at popular web-portals, which can be used to estimate the number of online users at that web portal, and (ii) DNS caches, which can be probed to inspect browsing requests for a variety of domains. We present a case study using either approach, while identifying potential pitfalls and some ways to overcome them.