The LinkedIn Algorithm: How it Works
By Joshua BoydDec 13th
Published July 20th 2016
Political discourse in America has become increasingly divisive over the past few years.
This is nowhere more apparent than on Twitter, where several groups have emerged as uniting forces for Conservative and Progressive, including #tcot (Top Conservatives on Twitter) and #UniteBlue.
#tcot has emerged in the last decade as a uniting hashtag for Conservatives on Twitter, a way to identify like-minded individuals via a simple search.
According to this International Business Times article from 2013, the hashtag first appeared in 2008. The hashtag #tcot appeared in more than 3,500 Twitter bios.
#UniteBlue serves much the same purpose for those on the left, allowing Progressive users to find their brethren on the site.
Although the cause was the subject of some controversy in the past, including allegations of fraud and shady backing, the cause is still fairly strong, with the hashtag appearing in more than 2,100 Twitter bios.
It is worth noting that both groups also have unique Twitter handles, @tcot and @UniteBlue – @tcot is an all but dead handle, with two posts in 2008, while @UniteBlue is a fairly inactive handle, with sporadic activity going back to 2011.
In this article, we’ll try to get at the heart of these groups, particularly how they interact with leading American journalists and who is seen as a leading voice for the causes.
Later, we’ll also take a look at how these audiences interact with the leading sources of media in America. Along the way, we’ll find some surprising data about these groups and their shared habits. All research is based on Brandwatch data from May 1 2015 – April 2016.
By using our Audiences tool, we have found a way to get at the heart of the matter, the ‘hashtags in the sand,’ if you will, that divide passionate Twitter users on either side of the political spectrum.
This research was accomplished with a few simple steps.
First: we wanted to see who the top Journalists are in America according to Twitter followers. This was done by searching for Individual users in America with the self-identified profession ‘Journalist’.
We then exported this list and wrote a Query based on @mentions of the top 300 journalists in America, again, sorted by followers.
Second: we wanted to see who identifies with the #tcot and #UniteBlue Twitter groups in order to understand their interactions (through at_mentions) with the top journalists in America.
This was done in a similar way, by searching for Individual users within the United States with either key hashtag in their bio, thereby going beyond keyword searches for authors strongly identifying with the #tcot and #UniteBlue causes.
These authors (3,697 for #tcot, 2,178 for #UniteBlue) were then uploaded to the Brandwatch platform as author lists, one for the #tcot audience and another for the #UniteBlue audience.
Our date range for this project was one calendar year, from May 2015 – April 2016.
As an introduction to the two groups, let’s take a look at the gender breakdown of people with the hashtags in their bios compared to those that use #tcot and #uniteblue in their tweets.
As shown, users of either hashtag skew decidedly male, with 61% of users of either hashtag identifying male.
Among users in the #tcot audience, this divide is slightly less pronounced, with 41% of authors self-identifying as female. Among the #UniteBlue audience, approximately 58% of users were female.
Now, let’s take a look at the weekly trends in @mentions by the #tcot and #UniteBlue Twitter groups.
Reflecting the fact that the group has fewer authors, #UniteBlue’s weekly trend is well below that of #tcot week over week. As shown, each group retained unique peaks in @mention-driven conversation.
Among the #tcot Twitter audience, the GOP Debate drove conversation during its highest-volume week, August 3rd. Within the #UniteBlue Twitter audience, two similar peaks are apparent: during the weeks of September 14th and February 8th.
During the week of September 14th, the CNN Debate drove @mention conversation.
This shows that, despite the lack of alignment between the groups, election politics are vital and salient to each hashtag-united group, as we’ll see later when we take a look at the most-mentioned authors for each group.
As another way of looking at the data, we’ve also calculated the mentions per user per week for each group.
As with any large social cause, certain authors will be more active than others. This is reflected in the below chart, with the @mentions per user by week being quite low for each group.
However, this gives us a great idea of not only the distribution of @mentions but also shows just how disparate these two audiences are, with the trends in conversation at times mirroring each other and rarely lining up.
In terms of weekly trend, these two groups have a barely positive correlation of .08, suggesting that the way they are sharing information from and interacting with leading journalists journalists is not on the same schedule over the course of the year.
From the start, we wanted to see how these two groups are interacting with the leading American journalists on Twitter and how that compares to the general population.
Again, these are @mentions of the leading journalists. The leading journalists below have been sorted by Twitter followers.
We found that among the general population, no one journalist made up a significant portion of @mentions.
However, the most-mentioned author is #tcot favorite Fox News personality @BretBaier (approximately 734K followers), suggesting a broad mainstream appeal for the anchor.
As shown, each audience @mentions about three of the leading journalists, with minimal overlap.
Among the #tcot audience, @mentions of Daily Wire Editor @benshapiro, Fox News anchor @BretBaier, and Conservative Review Editor @michellemalkin make up a significant portion of the audience’s @mentions.
Between the two groups’ interaction trends, there actually exists a negative correlation of -.24, suggesting a divergence. #tcot and #UniteBlue are unlikely to speak to the same journalists.
On the other hand, there exists a positive correlation between the general audience and each opposing audience (#tcot: .52; #uniteBlue: .37).
The stronger correlation shown between #tcot and the general population suggests an alignment between the two samples when interacting with these leading journalists.
As part of our research, we found the leading hashtags for each audience as well.
Perhaps not surprisingly, among the #tcot Twitter group, the #tcot hashtag appeared the most frequently, in more than 5% of all the audience’s mentions over the year.
On the other hand, the use of hashtags was far less prominent among users on the other side, with no hashtag cracking the 1% mark.
Leading hashtag #demdebate was used in only 0.7% of the audience’s @mentions. Despite #UniteBlue being the uniting hashtag among this particular progressive audience, it was only the ninth-leading hashtag, suggesting a weaker uniting force among the group.
Although they were certainly used in differing contexts by both groups, three hashtags (#trump, #gopdebate, and #blacklivesmatter) were within the top 10 hashtags for #tcot and #UniteBlue.
This shows a strong overlap in issues important to the opposing parties.
We also wanted to examine the leading (most mentioned) authors among each group, not necessarily limited to the top 300 authors on which we based our Query.
However, @mentions of leading journalists were prominent within the top ten for each group.
Among both audiences, @mentions of presidential candidates @hillaryclinton (3.56% of #UniteBlue), @realdonaldtrump (7.29% of #tcot), and @tedcruz (5.28% of #tcot) made up significant portions of audience @mentions of leading journalists, suggesting a trend toward co-mentions of politicians and journalists among both groups.
It is worth noting that @berniesanders was the 21st-leading author among the #UniteBlue audience, suggesting a slightly weaker association between the candidate and journalists.
Before moving on to the similarities between these two groups, let’s examine the broader media landscape in America. With a simple Query, we have collected all @mentions and interactions with the leading media sources in America and measured them as a percentage of the total sample (see below).
As we can see, the chart reinforces just how different these groups are.
Fox News is the most @mentioned news source by far for the right-leaning #tcot audience, comprising nearly 60% of the group’s media interactions.
Approximately the same portion is split between three media sources for the #UniteBlue audience: The New York Times, CNN, and MSNBC. Again, this shows that the two groups get their news from very divergent sources.
Despite these striking differences between the two groups, the social habits of both groups when interacting with the top American journalists are remarkably similar.
First, let’s take a look at the daily and weekly habits of #tcot and #UniteBlue, with volumes represented as a percentage of the group’s total @mentions.
The daily trend chart shown (based on EST) gives us a great idea of just how similar these two groups are when interacting with top journalists.
In fact, the two groups have a daily correlative value of 1, the highest possible value.
As shown, the two groups’ interactions with journalists increase throughout the day, with a pronounced peak at 6 pm followed by a sharp drop during the evening.
Although the correlation value for the two groups’ weekly trend is a bit weaker (0.57), the trends are certainly aligned, particularly on Wednesday and Friday.
Finally, we wanted to take a look at how often these mentions of journalists are organic (non-retweet) mentions or retweets.
We found something surprising, that for both groups, the majority of these @mentions are actually organic mentions. For the #tcot Twitter audience, approximately 98% of these interactions were organic. Approximately 97% of #UniteBlue’s @mentions were organic.
This suggests that when interacting with top American journalists, #tcot and #UniteBlue are similar in the way they mention and share stories, despite the fact that they are interacting with hardly any journalists in common.
All this shows that alongside Brandwatch Analytics, Audiences has a wide utility, helping to build not only queries, rules, and categories, but also author lists.
No matter your industry, the Audiences tool can help you build out your target market and understand your customers better.