Evidence behind media patterns.
Anmat conducts structured content research using computational analysis, documented datasets, and human interpretation to investigate framing, recurring narrative patterns, source representation, and coverage gaps.
The Problem
We examine large media corpora to identify recurring linguistic associations, represented voices, and coverage gaps that require contextual interpretation.
Methodology
Research Infrastructure
Documented corpus decisions, computational methods, human review, contextual interpretation, and clear analytical limitations.
Published research evidence
Model choice changes how Arabic media discourse is interpreted
Anmat and InfoTimes compared Arabic BERT models and general-purpose language models across 10,990 Arabic news headlines related to the 2023 Gaza war.
The study found that model selection is not a neutral technical decision: different models produced materially different sentiment interpretations.
10,990
Arabic headlines analyzed
Investigations
Featured Projects
Research Method
EKa'a | إيقاع
EKa'a | إيقاع
Analyzing contemporary Arabic song lyrics to identify societal portrayals & recurring themes. Expanding methodology across the Global South.
Analysis
Framing Gaza.
Framing Gaza.
An analysis of 25,000+ English-language articles across 16 outlets in the United States, United Kingdom, France, and Germany, covering the Gaza war's first year (October 2023 to October 2024), using Association Rule Mining (Apriori) to identify recurring linguistic associations.
Dataset
Women Framed.
Women Framed.
Building multi-regional datasets that help journalists and researchers understand how women are framed in violent events—whether as victims, perpetrators, or bystanders.
Collaborate With Us.
Have a research question about framing, source representation, recurring language, or coverage gaps across a defined media corpus?
Tell us what you are studying
We work with researchers, NGOs, journalism schools, think tanks, and media analysts on datasets, methods, workshops, and custom media ecosystem analysis.
Share the topic, outlets, region, or narrative question you want to investigate.