Context

The image at the top of the page is a frame from today's (7/26/2016) Isis #TweetMovie from twitter, a "normal" day when two Isis operatives murdered a priest saying mass in a French church. (You can see this in the center left). A selection of data from this site is being made available here to Kaggle users.

UPDATE: An excellent study by Audrey Alexander titled Digital Decay? is now available which traces the "change over time among English-language Islamic State sympathizers on Twitter.

Intent

This data set is intended to be a counterpoise to the How Isis Uses Twitter data set. That data set contains 17k tweets alleged to originate with "100+ pro-ISIS fanboys". This new set contains 122k tweets collected on two separate days, 7/4/2016 and 7/11/2016, which contained any of the following terms, with no further editing or selection:

• isis
• isil
• daesh
• islamicstate
• raqqa
• Mosul
• "islamic state"

This is not a perfect counterpoise as it almost surely contains a small number of pro-Isis fanboy tweets. However, unless some entity, such as Kaggle, is willing to expend significant resources on a service something like an expert level Mechanical Turk or Zooniverse, a high quality counterpoise is out of reach.

A counterpoise provides a balance or backdrop against which to measure a primary object, in this case the original pro-Isis data. So if anyone wants to discriminate between pro-Isis tweets and other tweets concerning Isis you will need to model the original pro-Isis data or signal against the counterpoise which is signal + noise. Further background and some analysis can be found in this forum thread.

This data comes from postmodernnews.com/token-tv.aspx which daily collects about 25MB of Isis tweets for the purposes of graphical display. PLEASE NOTE: This server is not currently active.

Data Details

There are several differences between the format of this data set and the pro-ISIS fanboy dataset. 1. All the twitter t.co tags have been expanded where possible 2. There are no "description, location, followers, numberstatuses" data columns.

I have also included my version of the original pro-ISIS fanboy set. This version has all the t.co links expanded where possible.

Replication data set (STATA format) and R code to reproduce analyses and figures in the paper. Abstract: What citizens think about Muslim immigrants is of great importance for some of the most pressing challenges facing Western democracies. To advance our understanding of what “Islamophobia” really is – i.e. whether it is a dislike based on immigrants` ethnic background, their religious identity or their specific religious behaviour – we fielded a representative online survey experiment in the UK in the summer 2015. Our results suggest that Muslims are not per se viewed more negatively than Christian immigrants. Instead, we provide evidence that citizens’ uneasiness with Muslim immigration is first and foremost the result of a rejection of fundamentalist forms of religiosity. This suggests that com-mon explanations, which are based on simple dichotomies between liberal supporters and conservative critics of immigration need to be re-evaluated. While the politically left and culturally liberal have more positive attitudes towards immigrants than right leaning and conservatives, they are also far more critical towards religious groups. We conclude that a large part of the current political controver-sy over Muslim immigration has to do with this double opposition. Importantly, the current political conflict over Muslim immigration is not so much about immigrants versus natives or even Muslim versus Christians as it is about political liberalism versus religious fundamentalism.

This study, designed and carried out by the "http://www.asarb.org/" Target="_blank">Association of Statisticians of American Religious Bodies (ASARB), compiled data on 372 religious bodies by county in the United States. Of these, the ASARB was able to gather data on congregations and adherents for 217 religious bodies and on congregations only for 155. Participating bodies included 354 Christian denominations, associations, or communions (including Latter-day Saints, Messianic Jews, and Unitarian/Universalist groups); counts of Jain, Shinto, Sikh, Tao, Zoroastrian, American Ethical Union, and National Spiritualist Association congregations, and counts of congregations and adherents from Baha'i, three Buddhist groupings, two Hindu groupings, four Jewish groupings, and Muslims. The 372 groups reported a total of 356,642 congregations with 161,224,088 adherents, comprising 48.6 percent of the total U.S. population of 331,449,281. Membership totals were estimated for some religious groups.

In January 2024, the ARDA added 21 religious tradition (RELTRAD) variables to this dataset. These variables start at variable #12 (TOTCNG_2020). Categories were assigned based on pages 88-94 in the original "https://www.usreligioncensus.org/index.php/node/1638" Target="_blank">2020 U.S. Religion Census Report.

Visit the "https://www.thearda.com/us-religion/sources-for-religious-congregations-membership-data" Target="_blank">frequently asked questions page for more information about the ARDA's religious congregation and membership data sources.