Y a node lies around the shortest path in between all pairs of nodes; the moreOpen AccessFigure 1 Quantity of messages posted about e-cigarettes over time.variety of shortest paths it resides in, the larger the betweenness worth.23 Within this context, the larger blue nodes represent discussion threads that straight link many nations collectively when they otherwise could not be connected. We also calculate closeness centrality (not represented visually), which measures the distance any node will be to all other nodes. Generally, core nodes will have larger closeness, as they’ve shorter paths to all other nodes than those on the periphery. Together with the 2-mode network, we now have a clear picture of your pattern of interactions in the GLOBALink forums. We’ve got labelled numerous nodes of interest and have identified them. 1st, we consist of the top rated 5 nations as determined by degree centrality (ie, number of discussion threads they may be present in), which are the exact same five we had visually discovered in the country network’s core cluster. Subsequent, we label the top five discussion thread IDs, as determined by their betweenness centrality:8324, 6, 13 022, 6467 and 9236. These threads serve to mediate discussions in between lots of pairs of nations. Last, we collect the thread IDs for the discussions that happen to be connected to the isolates (not labelled).Sentiment analysis Table 1 provides a common description of the sentiment scores for each of the messages. Figure four shows the pattern of sentiment in every single message over time. To find out how e-cigarettes compared with other subjects in GLOBALink, an independent samples t test was carried out to compare the sentiment scores for the ecigarette messages against all other messages inside the similar time BCTC supplier period ( July 2005 pril 2012). There was a significant distinction inside the scores for e-cigarette messages (M=0.0103, SD=0.0244) and all other messages (M=0.0144, SD=0.0294); t (41 695)=-3.87, p0.001,Figure two GLOBALink network of country-country interactions.Chu K-H, et al. BMJ Open 2015;5:e007654. doi:ten.1136bmjopen-2015-Open AccessFigure 3 GLOBALink 2-mode network of country-thread interactions.indicating that e-cigarette postings had been significantly additional negative. A post hoc simple linear regression was carried out to examine if the difference in sentiment involving ecigarettes and also other PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/21331607 subjects may be predicted by closeness centrality. The results had been considerable, F(1,32) =8.67, p0.01, and accounted for 18.86 (adjusted R2) from the explained variability. The regression equation was: predicted difference=0.029.026closeness centrality). DISCUSSION The exploratory network evaluation provided information that helped inform the later content analysis. We can make quite a few observations determined by the country-country network graph (figure 2). The network shows a core periphery structure, with several nodes inside a closely connected dense centre surrounded by much more loosely connected nodes at the outskirts. We are able to clearly see the higher degree core nations, most notably the USA, Australia, Canada, Switzerland and the UK, indicating a very interactive group of countries that participated in a lot of discussion threads collectively. At the other finish ofTable 1 Description of messages and sentiment Observations Raw range of sentiment scores Imply sentiment score (SD) Mean sentiment score normalised by word count (SD) Messages with good scores Messages with adverse scores Messages neutral or unscored 853 -144 to 130 11.34584 (30.05033) 0.0103133 (0.0244054) 528 252the network, we also notice t.