1) The document analyzes whether YouTube users have a sense of community and whether this is reflected in their interactions and connections on the platform.
2) Qualitative analysis of 100 videos and 32 in-depth analyses found that users develop strong personal connections and feelings of friendship through communication and support in video comments.
3) Structural network analysis found YouTube networks have higher average degrees of separation and clustering than other social networks, indicating stronger communities.
The YouTube Community: Sense of Structure and Rapport
1. The Community is Where the Rapport is: On Sense and Structure in the YouTube Community Dana Rotman, Jennifer Golbeck, Jennifer Preece
2.
3.
4.
5.
6. Sense of community membership Influence fulfillment Shared emotional connection
7.
8.
9.
10.
11.
12.
13. Statistics for several web-based social networks and the YouTube networks 0.05 8.7 2.1 1,832 YouTube Subscribers 0.08 5.7 1.9 1,512 YouTube friends 0.34 3.3 10.0 114,639 Tribe 0.45 3.5 8.3 1,326 Hamsterster 0.25 2.4 18.2 4,584 Fotothing 0.39 4.0 3.2 687 FilmTrust 0.43 2.1 35.7 8,143 Academy Clustering coefficient APL Average Degree Size Network
14.
15.
16.
17. Dana Rotman – drotman@umd.edu Dr. Jennifer Golbeck - golbeck@cs.umd.edu Prof. Jennifer Preece – preece@umd.edu
Editor's Notes
YouTube needs little introduction, here are some numbers: Youtube is the largest online video sharing site. In 2009 it has reached 100 million viewers, and holds a market share of more than 90% of the video sharing market. The latest (May 09) data, released by YouTube, claims that 20 hours of video are uploaded every minute. Uploading videos and sharing them (embedding them in other online formats such as blogs, social networks, etc.) is extremely easy, and contributes to the popularity of the site.
Many people think of YouTube as a broadcasting service and question if there is, indeed a community of YouTube users. In fact, the clear aim of YouTube is to create a “community that is highly motivated to watch and share videos”. The social networking tools that the community is given are varied Social network features: befriending, comments, subscriptions to get updates on the user’s activities, contact details (email, blogs, webpages) Users can calibrate the level of their participation – from passive viewing of videos to active participation in uploading videos, vlogging, creating artistic endavors, and socializing through comments, reciprocal videos and communication with other users.
Web based social networks and online communities exist on the same continuum. In effect, social networks are a form of the online community, or at least provide a platform for the creation on such community. Online communities are characterized by being an open arena for communication – fostering close, intimate relationships between members brought together by a shared interest. The ongoing rapport creates densely aggregated groups, which have many routes of communication – reciprocal (dialogues) or peripheral (sporadic comments with no reciprocation), leading to a feeling of connectedness. Structurally - online communities are considered to be a hub or a matrix of personal ties and continuous communication. The relationships are not nodal, but cluster-like.
McMillan and Charvis (1986), defined group members’ “sense of community” as the feelings group members develop towards their group when they actually believe they belong to this group. A sense of community is built upon 4 points: Membership -> the subjective feeling of belonging or relating to other group members Influence -> the group’s influence on its members, and the reciprocal influence of the group members on the group Fulfillment -> of some or all the member’s needs through the community. These may be knowledge, security concerns or status wishes, that the community can bestow upon the members Shared emotional connection among members – joint history, repertoire, closeness which could be found among community members. When applying a “sense of community” principle to a community, we can assume that the livelier and animated interaction is, in a way that involves as many members of the community as possible, the stronger the sense of community is.
The largest graph we obtained had 31,727 users – two factors limited the information we could gather: first, several users made their friendship lists private and inaccessible. Second, at random points the API would cut off the crawler, and some adjacency lists were thus unavailable. Therefore we ignored users for whom we couldn’t get complete adjacency lists.
Users were almost unanimous on their strong feelings about YouTube being an established community. Concepts of attachment, shared emotional connection, membership and fulfillment were often addressed. Many times they were discussed in reciprocal video and textual dialogue, through comments. YouTube was seen as a social space that offers the conditions needed to cultivate a community. The rapport established among YouTube members created recognition of fellow “YouTubers” and emotional attachment among them, as well as a joint culture of words, rituals and collaborations.
While users explicitly described a hub-like interaction pattern (like the video at the beginning), a close examination of the actual depiction of the interaction patterns proved that their behavior was less cohesive. Most interactions involved a small number of users, usually in pairs, who personally interact with changing degrees of frequency. Intimate meetings and artistic collaborations are highly regarded, but do not contribute much to a cohesive community. Reports of group activities were rare, and usually included solitary, individual, efforts towards a common goal (fundraising, awareness creating), or meetings based on geographical proximity.
The nodal nature of interaction on YouTube is best exemplified by the way users use comments. It must be noted that pattern of ongoing discussions and complex conversations, which usually portray a community, were not found in our analysis. Also, comments were not used to initiate friendship offers, and most commentators did not follow on their comments by subscribing to the user’s channel. Most personal communication was done by back-channels, such as email and through personal websites.
Our aim in this part of the study was to try and find corroboration to the sense of community users expressed or to their individualistic pattern of interaction, which we observed. To determine if, indeed, YouTube is a social network, we used small world theory. Small world networks are characterized by 2 structural features: the have a low average shortest path length (APL), similar to a random graph; and a high clustering coefficient similar to a regular connected graph. We generated a regular graph with the properties expected in a SWN, with the same no. of nodes and average degree, as the YouTube networks. In the random graph we generated, which has small world features – that are detailed in the box, the network properties are an APL of approximately .53, and a clustering coefficient of .44 In the friendship network, on the other hand, the APL was 5.7, which is within the range of what could be expected from a small world network, but the clustering coefficient is only .08, which is dramatically lower than the .44 that is expected. The same can be said for the subscription network APL 8.7, clustering coefficient .05, even lower than in the friendship network.
The implication of such a low clustering coefficient is that the network is almost random , we only see these values when approximately 95% of the edges have been randomly rewired in the simulation.
We found a discernable gap between the way users view their interaction patterns and the ways in which these interactions actually unfolded. While YouTube offers users the ability to communicate with each other, the tools it provides are geared towards broadcasting and less equipped for building a cohesive community. The personal channels (profiles), personal bulletin boards, and personal comment sections enhance users’ individual presence, but there are no “commons” on the YouTube community, which could potentially bring large groups of users together. The solitary channels create a random structure of the network, and lack community-wide rapport. Despite that users reported attachment and association with the YouTube community. That suggest that these feelings may be the outcome of a small number of personal relationships or small scale hubs. We also think that visual recognition plays a significant part on creating closeness and intimacy which enhance the feeling of familiarity to the other community members.
We believe that researchers grasp only a partial understanding of online communities or social networks when looking at them through a specific methodological perspective. In the case of YouTube, using only qualitative analysis of users’ accounts and their narrative, would have caused us to believe that YouTube is a cohesive community, but would not have provided us with the structural form of YouTube. Relying only on structural analysis would have presented us with a random network, at best, and would not have revealed the thriving interaction behind it. Only the compilation of the two offers us a wider perspective which gives a more complete picture of the community. Online communities and social networks are about the interplay of structure and users. Both are needed to create a successful community, and both need to be acknowledged in research, and in the method used for this research.
YouTube’s primary purpose is video sharing, and creating a repository for publishing videos. The social practices related to video sharing may be peripheral to this main purpose. Most social interaction may be but an occasional nod or thumbs up (or down) to an artistic or personal effort, and not a continuous complex interaction. That said, users do not care about the “actual” structure of a social site. They look for companionship, the want the “sense of community”, and they are willing to work around the site’s formal interaction mechanisms and objective constraints in looking for their community. They will utilize whatever tools or back-channels they can find, even on large scale networks such as YouTube. For the community members, the community is created where the rapport grows, however limited this rapport is.