Loading

Messages Created

The first metric is the number of messages created. It counts all messages created by a particular user or group including all private and public messages. Since it can be calculated for individuals and groups, it is of ego-centric and global scope. It is proposed by Berger et al. (2014), Hacker et al. (2015) and Smith et al. (2009). Viol et al. (2016) suggest a variant of this metric called "public messages" created. Because of the metrics' similarity, both of them are discussed in the following.

For the calculation of the message count $m$ of the user with the id $uid$ two steps are necessary. First, I select the data rows whose source attribute is equal to the given user id from the relationships table. Second, I run the $count$ aggregate function that returns the total number of messages of the particular user:

m(uid) := SELECT COUNT(source) FROM relationships WHERE source = uid; 

For the calculation of the message count $m$ of a group with the id $gid$, the first step selects the data rows with the given group id. The second step stays the same and returns the total number of messages of the particular group:

m(gid) := SELECT COUNT(groupid) FROM relationships WHERE groupid = gid; 

Riemer et al. (2015) strongly relate the number of posts to the level of participation of a particular user. They theorise that "participation in ESN constitutes social capital" (p. 9) and active users have a higher job performance than non-active users.

Berger et al. (2014) operationalise user activity via the number of posts. They state the activity is an important factor for the identification of key users. Key users are responsible for a significant number of messages. This is reasonable, because if users contribute a lot, they share their knowledge and help other users. This assumes that the posts are related to professional services of the organisation, which is the case in Enterprise Social Networks (Riemer et al., 2015).

Viol et al. (2016) determine that a high number of created messages characterises a power user. Power uses are well connected (c.f. degree- and closeness centrality) and highly visible. They are always online and among the most popular users of the network. Contrary, a low number of messages indicates either a temporary user or a niche expert. The former is not very active in the network and does not participate in discussion as the user usually takes a long time to reply in a conversation. Such a user receives a lot of likes and thanks messages and thus is popular and helpful towards other users.

The number of posts overlaps with the degree- and closeness centrality, which is shown by findings from Hacker et al. (2015). They correlate a high number of posts to a high engagement, but narrow focus. Smith et al. (2009) agree and find that a high number of messages characterises users, who are very active and engaging in discussions. They spawn new threads, new ideas and contribute to other threads, potentially helping other users. Angeletou et al. (2011) call such users elitists and popular users, who communicate a lot with their peers and are drivers of activity.

Because the message count is not a structural aspect of the network, it is not per se an indicator of Social Capital. However, due to the common factor and high similarity with the degree- and closeness centrality, I want to apply the internal perspective of Social Capital to the message count. Steinfield et al. (2009) state that a high number of posts and activity leads to bonding relationships and strong ties. These social relationships facilitate Bonding Social Capital, which enables a common understanding and effective collaboration. Contrary, a low number of posts indicates a lack of cohesion and therefore implies a low level of Bonding Social Capital. This interpretation is applicable to the ego-centric and the global version of this metric.