Published March 14, 2025 | Version v1
Dataset Open

Antisemitism on X: A Dataset Tracking Trends in Counter-Speech and Israel-Related Discourse Before and After October 7

  • 1. ROR icon Indiana University
  • 2. ROR icon Goethe University Frankfurt

Description

# Social Media & Hate Research Lab, Institute for the Study of Contemporary Antisemitism (ISCA) at Indiana University

 

Dataset Description

This dataset comprises 292 Top Tweets containing the keyword “Jews,” spanning various topics related to Jews, Israel, and antisemitism from September 2023 to May 2024. The annotation process was carried out using our annotation portal (https://annotate.osome.iu.edu/), which allows for the annotation of live data in its full context—including images and threads. The dataset includes tweet ID numbers, usernames, creation dates, tweet text, and classifications that indicate whether a tweet is antisemitic and/or calls out antisemitism.

 

 

Sampling

The dataset consists of four subsamples from different time periods. The samples were manually compiled using X’s Advanced Search tool with two newly registered accounts to ensure that no prior search history influenced the results. The Advanced Search feature allowed keyword searches within specific time frames, and the keyword “Jews” was used across four distinct sampling periods: Period 1: September 7–13, 2023; Period 2: October 7–13, 2023; Period 3: May 1–6, 2024; Period 4: September 7–13, 2024. For each period, we collected the first 75 Top Tweets, totaling 300 tweets, of which 292 remained live after annotation. To capture a more complete dataset, searches were adjusted to two-day intervals, starting with Top Tweets before switching to Latest Tweets with a minimum threshold of 200 views per tweet. However, due to platform inconsistencies, some tweets with fewer than 200 views were included while others with significantly higher view counts appeared lower in the ranking.[1]

Annotation Process

Each sample was annotated by three independent annotators. The annotators were students of the class “Israel on Social Media. Conceptions, Misconceptions, and Controversies” at Indiana University in the fall of 2024. The course included training on recognizing online antisemitism.

Annotations included:

·        Classification of tweets as antisemitic based on the Working Definition of Antisemitism by the International Holocaust Remembrance Alliance (IHRA).

·        Identification of tweets that denounce or call out antisemitism.

·        Thematic analysis and categorization of accusations against Israel.

Discrepancies in classification of antisemitism and calling out antisemitism were resolved through discussion among annotators, which increases the annotation quality.[2]               

File Description

The dataset is provided in CSV format, with each row representing a single tweet, including replies, quotes, and retweets. The file contains the following columns:

·        TweetID: Unique identifier for the tweet.

·        Username: Account that published the tweet.

·        Text: Full, unprocessed tweet text.

·        CreateDate: Date the tweet was posted.

·        Biased: Label indicating whether the tweet is antisemitic or non-antisemitic.

·        Keyword: The keyword used in the query, which may appear in the tweet text, a mention, or the username.

·        CallingOut: Tweets represent instances of "calling out" antisemitism

 

Observations

1. Trends in Antisemitic and Counter-Speech Content

The dataset shows fluctuations in the prevalence of antisemitism and counter-speech in top tweets about Jews:[3]

 

·        Before October 7, 2023, antisemitism in top tweets was already high, with over a quarter of tweets (25%+) classified as antisemitic. However, tweets calling out antisemitism were equally prevalent.

·        During the week after the Hamas attacks in Israel (October 7–13, 2023), the level of antisemitic tweets remained stable, but counter-speech increased to nearly 40% of tweets.

·        By May 2024, more than half of the tweets were antisemitic, while tweets denouncing antisemitism dropped to 12%.

·        In September 2024, antisemitism in top tweets had increased significantly compared to a year earlier (43%), yet counter-speech engagement also remained relatively high, making up one-third of all top tweets containing the word “Jews”.

 

2. Thematic Analysis

The topics covered in top tweets about Jews varied significantly across different periods:

·        Before October 7, 2023, discussions about Jews focused twice as much on general themes (e.g., religion, U.S. politics) compared to the Middle East conflict.

·        During the week following October 7, 75% of all top tweets were related to wars involving Israel, predominantly the war with Hamas.

·        One year later (May 2024), the war in Gaza remained the dominant theme but fell below 40% of tweets.

·        Notably, tweets covering anti-Israel protests increased significantly after October 7. While they accounted for only 6% of tweets before 10/7, in all subsequent sampled periods, their share more than doubled.

A key discussion in early May 2024 centered around the Antisemitism Awareness Act, which was passed by the U.S. House of Representatives on May 1, 2024. Many tweets referenced the bill, including a widely circulated post by Republican Representative Marjorie Taylor Greene, which amassed nearly 5 million views. [4] In her tweet, she explained her reasons for voting against the bill and made a statement that could be interpreted as echoing the antisemitic accusation that Jews were responsible for the death of Jesus.

 

Acknowledgements 

We wish to acknowledge the Advanced Cyberinfrastructure Curriculum Fellows Program for their exceptional support, with particular appreciation to Yu Ma and Tony Walker for the outstanding resources and assistance they provided.

This work used Jetestream2 at Indiana University through allocation HUM200003 from the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, which is supported by U.S. National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296.[5]


[1] Due to platform limitations, the Advanced Search option stopped functioning after 40–45 minutes, even when alternating between two newly registered accounts. After approximately 20 minutes of use, search functionality frequently ceased, requiring several hours of waiting before resuming data collection. This issue, combined with inconsistencies in the “Top Tweets” ranking, may have influenced tweet selection.

[2] Gunther Jikeli et al., “Differences between Antisemitic and Non-Antisemitic English Language Tweets,” Computational and Mathematical Organization Theory, September 9, 2022, https://doi.org/10.1007/s10588-022-09363-2.

[3] Please find more details in the research report Myles Behar et al., “The Image of Israel on X and TikTok After 10/7. Research Report, December 2024” (Research Lab Social Media & Hate at the Institute for the Study of Contemporary Antisemitism at Indiana University, December 2024), https://isca.indiana.edu/publication-research/social-media-project/documents-soc-media-proj/research-report-2024-the-image-of-israel-on-social-media-after-10-7-1.pdf.

[4] Antisemitism Awareness Act. Roll Call 172, Bill Number: H. R. 6090, 118th Congress, 2nd Session,” Office of the Clerk, U.S. House of Representatives, May 1, 2024, https://clerk.house.gov/Votes/2024172.

[5] Timothy J. Boerner et al., “ACCESS: Advancing Innovation: NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support,” in Practice and Experience in Advanced Research Computing (PEARC ’23: Practice and Experience in Advanced Research Computing, Portland OR USA: ACM, 2023), 173–76, https://doi.org/10.1145/3569951.3597559.

Files

dataset_14_03_25.csv

Files (93.8 kB)

Name Size Download all
md5:c24595353113e56b91b4c4023ea34b4e
93.8 kB Preview Download