Research Annotator — Qualitative and Mixed Methods Analysis of Digital Hate

Job Description:

AddressHate is hiring Research Annotators to join Decoding Hate, a field-building research hub housed at NYU's Center for the Study of Antisemitism that examines how hate circulates, adapts, and normalizes in digital environments. The position offers the opportunity to contribute to rigorous academic research, policy-relevant insights, and the development of context-aware AI detection systems.

We're looking for annotators with high sensitivity to the layered, implicit, and multimodal forms hate takes online — particularly where meaning is carried through irony, coding, visual reference, intertextual play, and shifting register across platforms. Rather than mechanical tagging, the position requires theory-informed interpretive work to form the empirical backbone of interdisciplinary hate speech research.

Our work demonstrates that a large portion of hate content in mainstream digital spaces operates through implicit, coded, or multimodal forms — spread through irony, political framing, and shared references that are largely undetectable by automated systems. Understanding this requires human expertise grounded in linguistics, multimodal analysis, and discourse studies.

The role is US-based (Eastern Time preferred; NYC-based ideal) and reports to the project's Lead Researcher.

Key Responsibilities

Qualitative Analysis & Annotation

• Conduct continuous, systematic annotation of social media content across platforms including TikTok, X, Instagram, and Reddit
• Analyze textual, visual, and multimodal content, including memes, image–text combinations, and visual rhetoric
• Identify implicit meanings, indirect references, metaphors, narrative frames, and visual-symbolic cues
• Identify shared rhetorical and discursive mechanisms across hate domains — including dehumanization, conspiracy narratives, victim–perpetrator reversal, moral inversion, and normalization strategies
• Analyze intersecting and overlapping hate discourses, including cases where antisemitism, misogyny, and racism co-occur or mutually reinforce one another
• Distinguish between legitimate political expression and harmful discourse, with close attention to context, pragmatics, and discursive positioning

Research Collaboration & Method Development

• Participate in regular team meetings and inter-annotator calibration sessions
• Contribute to the iterative refinement of coding schemes, category definitions, and annotation guidelines based on empirical findings
• Work closely with a team of data scientists in the modelling step, translating qualitative analytical insights into structured input for computational analysis
• Produce analytical memos and structured documentation that make interpretive decisions transparent and reproducible

Qualifications

• Native or near-native English proficiency with strong language sensitivity
• MA or PhD student (or completed degree) in Linguistics, Discourse Studies, Semiotics, Visual Studies, Media Studies, Hate Studies, History, Political Science, or related field. Candidates with a background in linguistics and multimodal/visual analysis are particularly valued.
• Demonstrated experience with qualitative content analysis, discourse-analytical methods, and mixed methods research
• High awareness of the different layers of hate speech and digital hate as they appear in text and imagery — including implicit, coded, ambiguous, and context-dependent forms
• Fluency with contemporary social media platforms, particularly TikTok, X, Instagram, and Reddit
• Openness and resilience to engage with hate speech content, including antisemitism, different forms of racism, and misogyny
• Curiosity, collegiality, and flexibility. Decoding Hate is a collaborative environment where coding schemes evolve, new platforms and case studies enter the pipeline, and analytical questions sharpen over time. We're looking for people who thrive in that kind of evolving research environment.
• High motivation and capacity for autonomous work. The role involves sustained, continuous engagement with difficult online content. Strong candidates set their own analytical goals within the project framework, structure their time independently, and maintain rigor and momentum across long annotation cycles.

Desired Backgrounds

• Research experience across more than one hate ideology — antisemitism, misogyny and gender-based hate, anti-Black racism, anti-Asian racism, or intersectional hate research. We particularly value candidates who can think comparatively across these domains rather than specialists in only one.
• Experience with multimodal analysis (memes, visual rhetoric, image–text relations)
• Experience with MAXQDA or comparable qualitative analysis software
• Strong written documentation skills

Analytical Orientation

Decoding Hate prioritizes depth of interpretation over surface-level labeling. Annotators are expected to approach hate communication as a discursive, contextual, and socially embedded phenomenon, recognizing that meaning often emerges indirectly and relationally across target groups and platforms. High-quality annotation is understood not as mechanical tagging, but as theory-informed interpretive work that forms the empirical backbone of rigorous academic research, policy-relevant insights, and the development of context-aware computational models.

Terms

20–30 hours per week · 1–2 years (initial trial period) · $25 USD per hour · Hired through AddressHate

About AddressHate and Decoding Hate

AddressHate is the organizational home of Decoding Hate, a research initiative based at NYU's Center for the Study of Antisemitism. The Center provides the academic anchor for Decoding Hate's research infrastructure, supporting rigorous, interdisciplinary scholarship on antisemitism and other forms of group-based hate in their historical and contemporary forms.

Why This Work Matters

The analytical tools we develop inform education, policy, and platform governance. By the time hate's effects are visible in the world, the discourse that produced it has long since moved on. We're building the infrastructure to see it as it happens — and to understand what it means.