Warmth and Competence (Sentence-Level Social Perception)

# Warmth and Competence — Sentence-Level Social Perception
# Based on the W&C-Sent dataset:
#   Ayesh, Mohammad, and Ousidhoum (2026), "Annotating Dimensions of Social
#   Perception in Text: A Sentence-Level Dataset of Warmth and Competence."
#   ACL 2026 (Volume 1: Long Papers), pp. 6374-6412.
#   Paper: https://aclanthology.org/2026.acl-long.289/
#   Data:  https://github.com/nedjmaou/W_C_Sent
#
# Warmth and competence are the two fundamental dimensions along which people
# evaluate individuals and social groups. Each item pairs a sentence with a
# target entity; annotators rate the sentence, with respect to that target, on
# three independent 7-point (-3 to +3) dimensions:
#   - Trust        (a component of warmth)
#   - Sociability  (a component of warmth)
#   - Competence
# In the original study each sentence-target pair was rated by 4-7 annotators,
# with each dimension presented independently to reduce cognitive load.

annotation_task_name: "Warmth and Competence - Social Perception"
task_dir: "."

data_files:
  - sample-data.json

item_properties:
  id_key: "id"
  text_key: "text"

output_annotation_dir: "annotation_output/"
export_annotation_format: "jsonl"

port: 8000

annotation_schemes:
  - annotation_type: likert
    name: trust
    description: "TRUST (a component of warmth): how much does the sentence express trust vs. distrust toward the target?"
    labels:
      -3: "-3 high distrust"
      -2: "-2 moderate distrust"
      -1: "-1 slight distrust"
      0: "0 neutral / not applicable / not expressed"
      1: "+1 slight trust"
      2: "+2 moderate trust"
      3: "+3 high trust"

  - annotation_type: likert
    name: sociability
    description: "SOCIABILITY (a component of warmth): how much does the sentence express sociability vs. unsociability toward the target?"
    labels:
      -3: "-3 high unsociability"
      -2: "-2 moderate unsociability"
      -1: "-1 slight unsociability"
      0: "0 neutral / not applicable / not expressed"
      1: "+1 slight sociability"
      2: "+2 moderate sociability"
      3: "+3 high sociability"

  - annotation_type: likert
    name: competence
    description: "COMPETENCE: how much does the sentence express competence vs. incompetence toward the target?"
    labels:
      -3: "-3 high incompetence"
      -2: "-2 moderate incompetence"
      -1: "-1 slight incompetence"
      0: "0 neutral / not applicable / not expressed"
      1: "+1 slight competence"
      2: "+2 moderate competence"
      3: "+3 high competence"

annotation_instructions: |
  You are shown a social-media post together with a TARGET entity (an individual
  or a social group). Read the post and judge what it expresses about the target
  along three dimensions, each on a scale from -3 to +3:

    - Trust:        does it portray the target as trustworthy or untrustworthy?
    - Sociability:  does it portray the target as warm/friendly or cold/hostile?
    - Competence:   does it portray the target as capable or incapable?

  Rate each dimension independently. Use 0 when the dimension is neutral, not
  applicable, or not expressed. Judge only what the sentence conveys about the
  named target, not your own opinion of that target.

user_config:
  allow_all_users: true
max_annotations_per_user: 50
num_annotators_per_item: 5