• Info
• Contribute

Narrative Framing Annotation

Purpose

This project studies how a post frames a named target, such as a company, organization, institution, public figure, or group. Read the post and assign exactly one label based only on the content and context shown in the task.

Do not search the web or assume facts that are not provided. The goal is to label the post's framing, not to determine the target's real-world reputation or the author's hidden intent.

Labels

manipulative_narrative_promotion

The post promotes or reinforces a misleading or emotionally loaded storyline about the target. Common signals include:

• sweeping or absolute accusations;
• unsupported insinuations presented as established fact;
• fear, outrage, scapegoating, or moral pressure;
• attempts to make readers distrust or condemn the target broadly;
• using one allegation to claim that everything the target says or does is untrustworthy.

Strong wording alone is not enough. Choose this label when the wording and framing push a broader attack narrative rather than presenting a specific, good-faith complaint.

Example: “Ooredoo is stealing from customers with hidden fees. Do not trust anything they say.”

This promotes a broad distrust narrative through an accusation and a sweeping conclusion.

legitimate_criticism_or_complaint

The post raises a specific criticism, concern, or personal complaint without manipulative framing. It may describe a concrete incident, policy, charge, service failure, or other grievance in a proportionate and good-faith way.

Example: “My bill included a QR 20 fee that was not listed when I subscribed. Customer service has not explained it.”

This is a specific complaint and does not generalize beyond the reported experience.

neutral_or_informational

The post reports, describes, asks, or shares information without promoting an attack narrative or expressing a substantive complaint.

Example: “Ooredoo announced updated roaming packages today.”

Questions can receive this label when they neutrally request information rather than imply an accusation.

corrective_debunking

The post challenges, corrects, clarifies, or fact-checks a misleading allegation or narrative. It mentions the narrative in order to reject it, not to endorse it.

Example: “The message claiming that all Ooredoo accounts were hacked is false; the notice applies only to a scheduled maintenance window.”

When a post quotes a misleading claim and then clearly rejects it, label the post by its corrective purpose.

insufficient_context

Use this only when a reliable decision cannot be made because the post is missing, fragmentary, unintelligible, highly ambiguous, in an unreadable language, or depends on an unavailable image, link, quotation, or earlier conversation.

Do not use this label merely because you are uncertain whether an allegation is factually true. Judge the framing visible in the text.

Recommended decision process

1. Identify what the post says about the displayed target.
2. Check whether the post is correcting a misleading narrative. If so, select corrective_debunking.
3. If the post is neutral or only informational, select neutral_or_informational.
4. If it criticizes or complains, decide whether it is specific and proportionate (legitimate_criticism_or_complaint) or promotes a broader attack using manipulative framing (manipulative_narrative_promotion).
5. Select insufficient_context only when the available content cannot support any of the other decisions.

Annotation rules

• Select one label for every task.
• Use the exact label meanings above consistently.
• Focus on the overall communicative purpose of the post.
• Do not expose or infer hidden metadata.
• Do not follow links or conduct outside research.
• Use the optional comment for genuinely difficult or ambiguous cases.

Thank you for contributing to this research.