Tayyar dataset

What's included

• Field-level source citations. Each fact about each entity links back to where it came from — founding year cites a different source than current leader, which cites a different source than legal status. The coverage page tracks the rollup.
• 16 calibrated axes. Economic, social, state-religion, democracy, west-alignment, regional-stance, Palestinian question, civil liberties, regime stance, pan-Arab, federalism, modernization, gender, iran-posture, press-freedom, sectarianism. Each one comes with a scoring rubric and concrete MENA examples anchored along the scale.
• Richer status than yes/no. Parties carry a government role (lead, coalition major / minor, confidence-and-supply, opposition major / minor, extra-parliamentary, banned) and a legal status (legal, restricted, outlawed, dissolved, merged away). Opposition and independent flags sit alongside.
• Declared vs. behavioral on key cases. Parties like Hezbollah and Hamas read as more committed to democracy in what they declare than in what they do; 20 parties carry a rhetoric-vs-record gap that is itself the finding. The home compass has a lens toggle to switch between views.
• 425 primary-source documents and 101 verified quotes. The document corpus carries verbatim manifestos, charters, parliamentary speeches, and UN addresses with country / party / politician attribution. The quote corpus drives Who-said-it and the "On the record" sections on every party / politician page.
• A semi-live event feed. Pulse tracks recent political shifts with a confidence rating — confirmed, reported, rumored, speculative — so in-flux developments (party formations, merger talks) can sit alongside confirmed events without being conflated. Subscribable as RSS.
• Frozen, versioned release. The full dataset is published as v1.0: every table as CSV and JSON, with a README that explains the rounds and fields, CITATION.cff, and SHA-256 checksums. One deliberate exclusion: the r7-dvb study round ships with its own pre-registered paper, not here.

Cite as

Cite the frozen release for the data and the paper for the instrument. The release version pins a citation that cannot shift as the live dataset changes, and the DOI resolves to the archived copy on Zenodo.

Gara, T. (2026). Tayyar: a MENA political-position dataset scored by a language-model panel (Version v1.0) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.23704154

@misc{gara_tayyar_dataset_v1_0,
  author    = {Gara, Tarek},
  title     = {Tayyar: a MENA political-position dataset scored by a language-model panel},
  year      = {2026},
  version   = {v1.0},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23704154},
  url       = {https://doi.org/10.5281/zenodo.23704154},
  note      = {Frozen release v1.0 (2026-10-27); CC BY-NC-SA 4.0}
}

@misc{gara_tayyar_2026,
  author        = {Gara, Tarek},
  title         = {The Model as One Rater Among Several: Measuring Political Positions in Data-Sparse Regions with a Language-Model Panel},
  year          = {2026},
  eprint        = {2606.23042},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CY},
  doi           = {10.48550/arXiv.2606.23042},
  url           = {https://arxiv.org/abs/2606.23042}
}

Gara, T. (2026). The Model as One Rater Among Several: Measuring Political Positions in Data-Sparse Regions with a Language-Model Panel. arXiv:2606.23042. https://arxiv.org/abs/2606.23042

Where to read more

• Methodology — how the dataset got built and where it falls short
• Findings — the structural patterns the data shows
• Coverage — verification status, country breakdowns, special-status leaderboards
• Axes catalog — all 16 axes with correlations and per-axis stats

Data access

Release v1.0 is public under CC BY-NC-SA 4.0. The zip holds all 17 tables as CSV and JSON, the README that explains the rating rounds and every field, CITATION.cff, the licence, and SHA-256 checksums for every file (sha256sum -c SHA256SUMS.txt). One round is deliberately absent: r7-dvb belongs to a pre-registered study and is released with that paper.

Download tayyar-dataset-v1.0.zip 4.0 MB · SHA-256 022c4a54f2eedda3…

Before its first public release, v1.0 was corrected: 154 of one model's ratings had been attached to the wrong parties by an import error, and are now on the parties their written rationales name. The README lists every change. Single tables, as CSV:

• tayyar_aggregate_positions.csv 1,195 rows · 211 KB
• tayyar_analyses.csv 0 rows · 0 B
• tayyar_axes.csv 20 rows · 29 KB
• tayyar_bills.csv 61 rows · 75 KB
• tayyar_comparison_essays.csv 20 rows · 46 KB
• tayyar_countries.csv 20 rows · 9 KB
• tayyar_documents.csv 425 rows · 1.2 MB
• tayyar_events.csv 353 rows · 326 KB
• tayyar_human_codings.csv 0 rows · 0 B
• tayyar_irr_sample.csv 212 rows · 28 KB
• tayyar_method_runs.csv 0 rows · 0 B
• tayyar_parties.csv 98 rows · 91 KB
• tayyar_party_alliances.csv 4 rows · 754 B
• tayyar_politicians.csv 274 rows · 191 KB
• tayyar_quotes.csv 989 rows · 687 KB
• tayyar_ratings.csv 12,867 rows · 3.8 MB
• tayyar_second_rater_codings.csv 126 rows · 37 KB

Documentation: CITATION.cff · README.md · SHA256SUMS.txt · TAYYAR_LICENSE.md · manifest.json

What's next

• Independent human coding. A paid coder scoring blind from the same evidence. That is the validity check a model panel cannot run on itself, and the release already carries the empty tables it will fill.
• The declared-vs-behavioral study. Its ratings ship with its pre-registered paper.
• The next release. The live dataset has grown since the July freeze. A later version will carry those additions; v1.0 stays exactly as it is.