Estimating Terrorist Recruitment Using Social Media Data – A Perilous Task

Although recruitment and general building of support are widely believed to be key motivators of terrorist attacks, the illegal nature of terrorism leaves researchers with little opportunity to test this assumption. They lack the easily available organizational records and extensive annual surveys, which characterize the study of most legal enterprises. Recently, some researchers (e.g. Barcelo and Labzina 2018) have turned to social media – specifically Twitter - to gauge fluctuations in levels of support in response to terrorist attacks. Although commendable for its ingenuity, in this research note I explore the pitfalls of this approach. Besides the questionable assumption that online mirrors offline support, issues also arise in relation to the impact of social media rules of engagement with violent or illegal content. When the timing of twitter take-downs of terrorist related accounts is taken into consideration, the claimed significant effects of terrorist attacks on number of supporters disappear. In addition, the original study also suffers from a range of questionable research practices regarding the statistical modeling of effects. In summation, though a valuable new source of information on a type of organization inherently difficult to study, I recommend caution when using social media data in quantitative research on terrorism.