Economic and Social Research Council
 Printable version
|
Investigating Twitter abuse
Analysis of tweets posted during the 2014 Gaza conflict shows patterns of anti-Semitic abuse on social media.
It is now well known that each time there is an upsurge in the Israel-Palestine conflict there is a rise in violence against Jews around the world. So it was in 2014 with Israel's 'Operation Protective Edge' military action. According to the Tel Aviv University Kantor Centre Anti-Semitism Worldwide 2014 report, it was one of the worst years on record for anti-Semitic incidents globally.
With the growth of social media, an apparent upsurge of anti-Jewish abuse on social media networks such as Facebook and Twitter was also noticeable.
Insults against Jews are frequently hurled on the streets and other public places. But the sentiments expressed by offenders have not usually been accessible to researchers because of the fleeting nature of their occurrence. Insults slung on social media, by contrast, are preserved for scrutiny.
Criminologists at Lancaster University working in the ESRC Centre for Corpus Approaches to Social Science undertook an analysis of tweets sent during the Gaza conflict that took place in July and August 2014. They were commissioned to provide a rapid response analysis to inform the 2015 report of the All-Party Parliamentary Inquiry into Anti-Semitism.
Working with many millions of tweets, they carried out a detailed analysis, using the core techniques of corpus linguistics – computer-aided linguistic analysis – on a sub-sample of 38,460 tweets containing the words 'Israel' or 'Gaza', along with the words 'Jew', 'Jews' or 'Jewish'. The results from the studying selection were telling.
One technique from corpus linguistics is keyword analysis – looking for words with unusually high frequencies. In this data, a keyword approach showed the spectre of Nazism looming large in the data, with words such as 'Hitler', 'Holocaust', 'Nazi' and 'Nazis' in the top keywords.
Negative sentiments
Unlike other approaches to the analysis of social media data, which rely largely on automated techniques, the corpus linguistics approach blends expert human analysis and computer-assisted analysis because the linguistic contexts in which keywords and hashtags are located matter.
By exploring the data in this way we better understand the highly negative sentiments of the tweets. Some tweets contained explicit anti-Jewish invective, which, if shouted out on the streets, would clearly amount to criminal offences. Some wished violence upon Jews as proxies for Israelis, or simply just as Jews. Some expressed the sentiment that 'Hitler should have finished the job'. Shockingly, the use of gas chambers on Jews was invoked.
The impact of the analysis was underlined by the Report of the All-Party Parliamentary Inquiry into Anti-Semitism. It stated that the 'importance of this research should not be underestimated. It helps identify some of the themes in discourse and, with time, could help to detect patterns of anti-Semitism and therefore better direct resources to combat it'. It also called for further research, which is being undertaken by the Lancaster University team.
Further information
This article was originally published in our Britain in 2016 magazine.
