- Accuracy in Media - https://www.aim.org -

Tweeting to the Left: Twitter’s Hidden Bias

Twitter has taken the social media world by a storm with its innovative medium for microblogging; however, it turns out, not even social media is free from bias.  Recently, Twitter announced that it would be eventually deleting its “suggested user” feature-because of its pro-Democratic bias.

The “suggested user” feature was originally meant to link new Twitter users to famous people, such as celebrities, athletes, and politicians.  However, as the Associated Press [1] reported, people on this feature were chosen by “company officials,” all of whom apparently were staunch Democrats who saw no need for ideological diversity on the list.  In California, only Democratic candidates for governor were placed on the list; Republican candidates, on the other hand, were only added after the Associated Press ran a story on the “perceived” bias three weeks ago.  The additions to the feature saw some notable results, as the AP reported on Monday:

Former eBay chief executive Meg Whitman, who led the Republican field with 4,160 Twitter followers, jumped to nearly 61,000 followers. Former Congressman Tom Campbell went from 1,660 followers to 57,500, while state Insurance Commissioner Steve Poizner’s nearly 2,600 followers increased to 56,500.

By comparison, Attorney General Jerry Brown, the presumed Democratic gubernatorial candidate, increased from 960,000 followers to 1 million during the same three-week period.

Clearly, this bias in the favor of Democrats has made a difference-three weeks on the “suggested user” list can get a candidate tens of thousands of followers, and up until recently, Republicans were de facto banned from the list.  Of course, having many followers doesn’t automatically translate into political success, but it is clear that Twitter’s “company officials” were hoping that it would.  If they wanted an even playing field for members of both parties, they clearly would have included Republican gubernatorial candidates instead of only Democratic ones.

These biased “company officials” at Twitter, though, had good reason to ignore Republican candidates. A higher percentage of Republicans use Twitter in the national political sphere [2]: 22.3 percent of congressional Republicans use Twitter, while only 8.5 percent of congressional Democrats do.  As of September 2009, Republicans in Congress were out-tweeting Democratic representatives [3]; 64 percent of congressional tweets came from Republicans, as opposed to the Democratic Party’s embarrassing 36 percent. The White House celebrated its one millionth follower on Twitter back on August 30-but John McCain had passed that point six weeks before. [4]  Top Conservatives on Twitter (#tcot) is widely popular among conservatives; but the Top Progressives group for tweeting libs (#topprog) has not done nearly as well (despite claims [5] of its “instantaneous” popularity).  Even now, months after Top Progressives’ Twitter infancy, a cursory glance at the #topprog page shows tweets from as far back as three hours-as compared to the #tcot page that tracks recent tweets by half-minutes.  While I was examining the two pages, I missed 87 new tweets from the Top Conservatives page, and zero (0) new tweets from the Top Progressives page.  And while President Obama may have over 2.6 million followers on Twitter, he recently announced that he has never actually used Twitter in his life [6].

So perhaps the “company officials” believed they were performing some sort of political affirmative action for wannabe Democratic governors, who according to statistics like that are comparatively bad at using Twitter.  

However, Twitter co-founder Biz Stone told a different story at a recent conference in Malaysia.  Huliq News [7] reports:

At the present time Stone and other company employees choose the people who are featured on the list. The list is based on users “who show that they provide value by posting often and engaging with their followers.”

This suggests that Stone believes something contrary to current statistics: that Democratic hopefuls “provide” more “value” by “posting often and engaging with their followers.”  But again, as of September, this did not hold true in the national political sphere, and the chances are slim that it holds true in California.  

This leaves two options for Stone.  First, he was seriously unaware that Republicans tend to tweet most often and engage followers most, and had simply never bothered to compare the Republican and Democratic candidates in California.  If Stone had bothered to compare them, he may have noticed that even in California, Republican candidates out-tweet Democratic ones.  For example, as of 10:03 AM on Tuesday morning, leading Republican Meg Whitman had tweeted 191 times; leading Democrat Jerry Brown had only tweeted 170 times.  It would appear that at least superficially and based on numbers alone, Meg Whitman deserves to be on the list more than Jerry Brown.

The second option is that Stone didn’t care that Republicans provide more “value” on Twitter when he was compiling the list of names to go on the feature.  Which do you feel is more likely?

At this Malaysian conference, Stone announced that the old feature would be deleted eventually, and replaced with “that is more programmatically chosen, something that actually delivers more relevant suggestions.”  This is good news for everyone who supports balance in social media.