Islamophobia Study


Media Categorization

To identify Islamophobic websites, we first compiled a list of 24 sites categorized by the Southern Poverty Law Center as presenting extreme anti-immigrant or anti-Muslim views or being associated with individuals or organizations known for extreme anti-immigrant or anti-Muslim views (“Hatewatch,” 2019). We then cross-checked this against sites identified as Islamophobic in the Fear, Inc. (Ali, Duss, Fang, Keyes, & Shakir, 2011) and Fear, Inc. 2.0 (Duss, Taeb, Gude, & Sofer, 2015) reports from the Center for American Progress, A Fairness & Accuracy in Reporting study of Islamophobia (Rendall, 2008), and the work of two scholars of the alt-right, Benjamin Lee (Lee, 2017) and Thomas Main (Main, 2018). Sites were only included on our final “Islamophobic” list if they had been identified as such by at least two of the above sources. We then confirmed that all the remaining sites were categorized by Media Bias/Fact Check as “Right Bias,” “Questionable Sources” or “Conspiracy-Pseudo-Science” (“Media Bias/Fact Check,”). There were three exceptions to the “two source” rule: CreepingSharia.blogspot was included because its self-described mission is to stop “the slow, deliberate, and methodical advance of Islamic law (sharia) in non-Muslim countries” (2019); was named as an alt-right site by Lee (Lee, 2017) and the authors confirmed the presence of Islamophobic content on the site; and was included because it is categorized as “questionable” by Media Bias/Fact Check and contains content published by other Islamophobic sites (Guandolo, 2016). This became our master “Islamophobic Websites” list.

The list of “Conservative media” was compiled from the Media Bias/Fact Check list of websites it categorizes as “Right Bias” (“Media Bias/Fact Check,”). We then cross-checked the above lists against four other sources that categorize media, The Righting (“An A-Z Guide To Right Wing Media,” 2019), (Hawkins, 2018) , a Pew Research Center report on media polarization (Mitchell, Matsa, Gottfried, & Kiley, 2014), and the crowd-sourced ratings tool AllSides (“Media Bias Ratings,”). Finally, using the website traffic tool SimilarWeb, we captured information on the national ranking and monthly visitor data for each website (“Website Performance,” 2019) to produce a list of the top 20 largest conservative news and information websites.

The list of “Far Right Blogs” includes a range from individuals espousing extreme conservative ideas or identifying with the Tea Party or Libertarian Party to white supremists. The categorization is drawn from the definition of the term to as “the group of people whose political views are the most conservative” (“The far right,” 2019). The core of the list comes from NewWhy, a web marketing firm that developed the list to allow advertisers to prevent their ads from appearing on alt-right, racist or sexist websites (Winslow, 2018). We eliminated corporate or think tank sites from the NewWhy list. As we reviewed and categorized the articles in each candidates’ Lexis-Nexis search, publications/sites that were not on our lists were individually reviewed and categorized by checking whether they were categorized by Media Bias/Fact Check, and, if not, by reviewing each site’s stated mission and content. Sites that reposted stories from publications on the “Islamophobic” list were automatically included in the Alt-Right website list or the “Far Right” list.

The “Islamophobic,” “Far Right,” and “Conservative” lists were used when categorizing news stories from Lexis-Nexis database (“NewsDesk,” 2019). We did not compile a “Liberal” list because that is outside the scope of this study.

Because Lexis-Nexis does not comprehensively monitor the sites on our master Islamophobic list, we had a team of graduate students do Google searches for coverage on these sites (e.g. “Rashida Tlaib”

The individual site searches also provided a glimpse into the echo chamber of the anti-Muslim online media. We selectively pasted headlines from Breitbart and other sites into the Google search window to obtain an anecdotal picture of the degree to which these sites repost articles. The Breitbart article about Omar Qudrat, headlined “Local News Blasts Democrat Scott Peters for Avoiding Debate with Republican Omar Qudrat,” was reposted by eight other websites.

Survey Methodology

We identified 23 candidates for Congress and 50 candidates for state legislatures and five candidates for statewide offices. Through a combination of emails, telephone calls and Twitter direct messages, we were able to reach 66 of the 78 candidates, requesting that they complete the online survey housed on SurveyMonkey. Forty did so, for a response rate of 61 percent.

Social Media Methodology

Facebook posts and comments were captured back to three months before the respective primaries. For candidates who lost the primary, capture ended on primary day. Because the state primaries are staggered, this means that the actual num

Real-time capture of Twitter posts began Sept. 21 as the final primaries were taking place. Historic Twitter data is not readily available. Therefore, Twitter analysis was confined to candidates who won their primary. Nov. 4 was set as the cutoff date for analysis of social media posts. This was to avoid distorting the data with the flurry of get-out-the-vote posts on the 5th and congratulatory or sympathy posts after the results were in. For context, Ilhan Omar, the most active candidate on social media, had 90,202 Tweets that tagged her between Sept. 30 and Nov. 4 (inclusive). During the period Nov. 5 – 7, the account had 12,564 Tweets.

The ability of researchers to track social media trolling is affected by several factors. On Facebook, some campaigns delete offensive comments, so an after-the-fact review of comments can present a distorted picture. On Twitter, individuals are powerless to control posts about them or aimed at them. However, Twitter itself is relatively aggressive in taking down offensive or threatening tweets and in the case of bots will sometimes detect and stop them even before they are posted.

Charts comparing the level of trolling among candidates are based on the percentage of overall Tweets or Facebook comments, rather than raw numbers because total social media traffic among candidates varied wildly. For example, Omar, Tlaib and Qudrat had a combined total of more than 110,000 Tweets by or about them. All the other Congressional candidates combined had only 68,000.

Rep. Andre Carson was not included in the analysis. He was the only sitting member of Congress in the group. As such, he was the outlier; media coverage of him was heavily skewed by hundreds of articles by the Targeted News Service and the Federal News Service, which cover Washington. Likewise, his social media feeds were full of discussion of issues he was working on in Congress, most notably the conflict in Cameroon, which overwhelmingly dominated his social media feeds.

All tweets were individually coded. We identified two key categories of tweets, those that were Islamophobic and those that were pro-Israel. Islamophobic tweets were those that contained language that was overtly anti-Muslim or immigrant. “Pro-Israel tweets were those that accused the candidate of anti-Semitism. In many cases, the latter also contained high levels of Islamophobia or anti-immigrant sentiment. In addition to those key categorizations, we also coded tweets with categories that were unique to particular candidates. For Republican Omar Qudrat, we coded 1) Supportive tweets from known conservative media commentators; and 2) Tweets that made positive mention of his Muslim faith. For Muslim candidates, we also created a category for tweets trolling them for not being supportive enough of Muslim/Palestinian causes.

We also coded based on whether the tweet was original or retweeted. Those allowed is to identify those trolls who were the most active against the candidates. We did not code for generic hate speech. Posts had to include overtly anti-Muslim or pro-Israel language in order to be classified as such. Even where a particular tweet was clearly responding positively to an Islamophobic or pro-Israel tweet, we did not code it unless it contained language that clearly identified it as such. We also did not code a tweet based on the handles included. Finally, we did not open the links included in a tweet, even if we suspected it was a link to an Islamophobic or pro-Israel tweet.

Finally, we coded all tweets that were posted by trolls, whether or not the tweet itself contained overtly negative language. To do this, we categorized all tweets by individuals who had posted at least one tweet that was coded according to one of the categories above.

We also ran the social media posts through a lexicon based primarily on Hatebase, giving us raw output of numbers of mentions of each word in the lexicon, coded as Islamophobic, hate speech, misogynist, etc.





  1. About. Retrieved from

Ali, W., Duss, M., Fang, L., Keyes, S., & Shakir, F. (2011). Fear, Inc.: The Roots of the Islamophobia Network in America. Retrieved from Washington, D.C.:

Duss, M., Taeb, J., Gude, K., & Sofer, K. (2015). Fear, Inc. 2.0: The Islamophobia Network’s Efforts to Manufacture Hate in America. Retrieved from Washington, D.C.:

The far right. (2019). Merriam-Webster Dictionary Online.  Retrieved from

Guandolo, J. (2016). Islamic Movement in U.S. Preparing for Battle. Right Side News. Retrieved from

Hatewatch. (2019). Retrieved from website:

Hawkins, M. (2018). The Top 10 Conservative News and Opinion Websites. Retrieved from website:

Lee, B. J. (2017). It’s not paranoia if they’re really out to get you. Behavioral Sciences of Terrorism and Political Aggression, 9(1), 4-20.

Main, T. J. (2018). The Rise of the Alt-Right. Washington, D.C.: Brookings Institution Press.

Media Bias Ratings.   Retrieved from

Media Bias/Fact Check.   Retrieved from

Mitchell, A., Matsa, K. E., Gottfried, J., & Kiley, J. (2014). Political Polarization & Media Habits. Retrieved from Washington, D.C.:

NewsDesk. (2019).   Retrieved from

Rendall, S. (2008). The Dirty Dozen. Smearcasting: How Islamophobes Spread Fear, Bigotry and Misinformation. Retrieved from

Website Performance. (2019). SimilarWeb.  Retrieved from

Winslow, N. (2018, Feb. 10). Excluding Alt-Right, Racist & Sexist Websites on Google Adwords & Facebook’s Audience Network.   Retrieved from

An A-Z Guide To Right Wing Media. (2019). The Righting. Retrieved from

%d bloggers like this: