Hellisotherpeople/OpenDebateEvidence-Annotated-Anonymized

OpenDebateEvidence-Annotated (Anonymized)

An LLM-annotated subset of OpenDebateEvidence debate evidence, with all debater-identifying columns removed.

This is an anonymized, Parquet-converted redistribution of Hellisotherpeople/OpenDebateEvidence-Annotated. 85,600 rows, 44 columns: 25 annotation fields plus the 19 retained OpenDebateEvidence evidence fields. The original pipe-delimited CSV had 70 columns, of which 26 were dropped. See Anonymization.

Companion datasets: OpenDebateEvidence-Anonymized (full corpus) and OpenDebateEvidence-Deduplicated-Anonymized.

Usage

from datasets import load_dataset

ds = load_dataset("Hellisotherpeople/OpenDebateEvidence-Annotated-Anonymized", split="train")

Anonymization

The upstream dataset truncated debater first and last names to their first two characters, but several other columns re-exposed the full identities. Most significantly, filePath and opensourcePath contained the uploaded document filenames, which embed full surnames and school names:

./documents/ndtceda15/Army/JoSo/Army-Johnson-Sowatzke-Aff-Navy-Round1.docx

Team codes such as JoSo are themselves built from debater surnames, and roundId / fileId join back to the public caselist wiki where the names remain. Because many of these debaters were minors at the time, the following 26 columns were dropped entirely:

Category	Columns dropped
Direct names	debater1First, debater1Last, debater2First, debater2Last
Team codes derived from surnames	teamId, teamName, teamDisplayName
School and geography	schoolId, schoolName, schoolDisplayName, state, chapterId
Paths embedding names and schools	filePath, opensourcePath
Named third parties	opponent, judge
Free text that may name people	notes, report
Join keys back to the source wiki	fileId, roundId
Round context that joins to the public Tabroom record	tournament, round
Caselist identifiers that narrow a row to a league and season	caselistId, caselistName, caselistUpdatedAt
Team composition	teamSize

The last three categories were removed in a second pass. Tabroom.com publishes tournament entries and results by name, so a tournament plus a round number is a lookup key against a public record even with every other identifier stripped. caselistDisplayName is retained because it names a league and season (NDT/CEDA 2015-16) and identifies nobody.

All 25 annotation columns are retained. None of them contain personal information.

Residual risk

If you find a record that identifies someone and you would like it removed, open a discussion on this repo.

Overview

The OpenDebateEvidence-annotated dataset is an extension of the base OpenDebateEvidence dataset, designed to provide a more comprehensive and detailed analysis of debate evidence. It includes additional columns that offer deeper insights into the quality, relevance, and strategic value of the evidence used in debates.

Annotations were produced by an LLM against the EvidenceAnalysis schema at the bottom of this card. Multi-valued fields are stored as stringified Python lists (e.g. "['empirical', 'expert_opinion']"). Parse them with ast.literal_eval.

Annotation Columns and Rationale

Core Analysis

• debate_relevance_score: How relevant the evidence is to common debate topics. Helps assess applicability across debate scenarios.
• author_credibility_rating: Overall source credibility rating, crucial for evaluating trustworthiness.
• evidence_type: Primary type of evidence presented (empirical, analytical, expert opinion, ...).

Methodological Assessment

• methodology_quality: Quality of the research methodology if applicable. Indicates robustness.
• sample_size_adequacy: Adequacy of sample size if empirical. Important for generalizability.
• data_recency: How recent the data or analysis is.

Argumentative Properties

• warrant_strength: Strength of the logical connection between evidence and claims.
• counter_argument_resistance: How well the evidence stands up to common counter-arguments.
• scope_of_impact: Scope of impact discussed (individual, community, national, ...).

Argument Classification

• argument_type: Types of debate argument the evidence supports (kritik, impact, link, ...).
• ism: If kritik evidence, the category (capitalism, colonialism, ...).
• impact_type: Types of impact discussed (extinction, war, economic, ...).

Strategic Value

• offensive_value: Utility for making affirmative arguments.
• defensive_value: Utility for defending against opposing arguments.
• uniqueness: How unique or novel the evidence is compared to common arguments.

Technical Details

• technical_complexity: Level of technical or academic complexity.
• jargon_density: Amount of field-specific terminology used.
• citation_quality: Quality level and type of citation source.

Topic Tags

• policy_domains: Policy areas discussed (foreign policy, economic, social, ...).
• theoretical_frameworks: Theoretical frameworks employed (realism, liberalism, ...).
• geographic_focus: Geographic regions discussed.

Debate Application

• best_use_timing: Best debate speech in which to use the evidence.
• side_applicability: Which side can best utilize the evidence.
• speech_position: Specific speech positions where the evidence is most useful.

Overall Quality

• overall_evidence_quality_rating: Overall quality rating of the evidence (1-10), stored as an integer.

Retained Evidence Columns

Column Name	Description
id	Unique identifier for the evidence
tag	Abstract summary of the evidence/argument made by the debater with evidence
cite	Short citation of the source used for the evidence
fullcite	Full citation of the source used for the evidence
summary	Longer extractive summary of the evidence
spoken	Shorter extractive summary of the evidence
fulltext	The full text of the evidence
textLength	The length of the text in the evidence in characters
markup	The full text of the evidence with HTML markup for parsing/visualization purposes
pocket	Top level heading, usually the speech name
hat	Medium level heading, usually the broad type of argument
block	Low level heading, usually the specific type of argument
bucketId	Unique identifier for the deduplication bucket of the evidence
duplicateCount	The number of duplicates of the evidence
side	The debate side on which the evidence was used (Affirmative or Negative)
caselistDisplayName	The league and season the evidence was uploaded under, e.g. NDT/CEDA 2015-16
year	The year in which the debate round took place
event	The event in which the debate round took place
level	The level of the debate (college, high school)

Annotation Schema

class EvidenceAnalysis(BaseModel):
    # Core Analysis
    debate_relevance_score: Literal["very_low", "low", "medium", "high", "very_high"]
    author_credibility_rating: Literal["very_low", "low", "medium", "high", "very_high"]
    evidence_type: List[Literal["empirical", "analytical", "expert_opinion", "statistical",
        "case_study", "theoretical", "experimental", "observational", "survey", "interview",
        "archival", "meta_analysis", "systematic_review", "longitudinal", "cross_sectional",
        "correlational", "qualitative", "quantitative", "mixed_methods", "ethnographic",
        "phenomenological", "grounded_theory", "narrative", "historical", "comparative",
        "descriptive", "exploratory", "explanatory", "predictive", "evaluative", ...]]

    # Methodological Assessment
    methodology_quality: Literal["very_low", "low", "medium", "high", "very_high"]
    sample_size_adequacy: Literal["very_low", "low", "medium", "high", "very_high"]
    data_recency: Literal["very_old", "old", "moderate", "recent", "very_recent"]

    # Argumentative Properties
    warrant_strength: Literal["very_weak", "weak", "moderate", "strong", "very_strong"]
    counter_argument_resistance: Literal["very_weak", "weak", "moderate", "strong", "very_strong"]
    scope_of_impact: Literal["individual", "community", "national", "international", "global"]

    # Argument Classification
    argument_type: List[Literal["kritik", "impact", "link", "turn", "alt", "perm", "framework",
        "theory", "topicality", "case", "counterplan", "disadvantage", "solvency", ...]]
    ism: List[Literal["capitalism", "colonialism", "feminism", "anthropocentrism",
        "securitization", "ableism", "racism", "sexism", "classism", ...]]
    impact_type: List[Literal["extinction", "war", "economic", "environmental", "social",
        "political", "moral", "structural", "nuclear_war", ...]]

    # Strategic Value
    offensive_value: Literal["very_low", "low", "medium", "high", "very_high"]
    defensive_value: Literal["very_low", "low", "medium", "high", "very_high"]
    uniqueness: Literal["very_low", "low", "medium", "high", "very_high"]

    # Technical Details
    technical_complexity: Literal["very_low", "low", "medium", "high", "very_high"]
    jargon_density: Literal["very_low", "low", "medium", "high", "very_high"]
    citation_quality: List[Literal["primary_peer_reviewed", "primary_government",
        "secondary_academic", "secondary_meta_analysis", "grey_literature", "preprint", ...]]

    # Topic Tags
    policy_domains: List[Literal["foreign_policy", "domestic_policy", "economic", "social",
        "environmental", "security", "technology", "education", "healthcare", ...]]
    theoretical_frameworks: List[Literal["realism", "liberalism", "constructivism",
        "critical_theory", "postmodernism", "feminism", "marxism", ...]]
    geographic_focus: List[Literal["north_america", "south_america", "europe", "asia",
        "africa", "oceania", "global", "space"]]

    # Debate Application
    best_use_timing: Literal["opening", "constructive", "rebuttal", "closing"]
    side_applicability: Literal["affirmative", "negative", "both"]
    speech_position: List[Literal["1ac", "1nc", "2ac", "2nc/1nr", "1ar"]]

    overall_evidence_quality_rating: Literal["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"]

The full enum lists are available on the original repo.

Citation

@inproceedings{
    roush2024opendebateevidence,
    title={OpenDebateEvidence: A Massive-Scale Argument Mining and Summarization Dataset},
    author={Allen G Roush and Yusuf Shabazz and Arvind Balaji and Peter Zhang and Stefano Mezza and Markus Zhang and Sanjay Basu and Sriram Vishwanath and Ravid Shwartz-Ziv},
    booktitle={The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track},
    year={2024},
    url={https://openreview.net/forum?id=43s8hgGTOX}
}