WASSA 2021 shared task

Download
Request access

About WASSA 2021 shared task

Task description

Participants are given an extended release of the empathic reactions to news stories dataset which contains essays and Batson empathic concern and personal distress scores in reaction to news articles where there is harm to a person, group, or other (for more details see the paper). The essays are between 300 and 800 characters in length. The extension of this dataset also includes person-level demographic information (age, gender, ethnicity, income, education level) as well as personality information. Additionally, we include emotion labels for the essays at both the document and sentence level, these emotion labels have been predicted automatically.

Track 1: Predicting empathy

The formulation of this task is to predict the Batson empathic concern (“feeling for someone”) and personal distress (“suffering with someone”) using the essay and any of the additional information, i.e. personality information, demographic information as well as the emotion labels.

The evaluation metric for Track I is Pearson correlation with the gold ratings (overall, empathic concern, personal distress). Below is an example of essays and labels.

Empathy

Distress

Essay

age

income

emotion

personality_openess

4.8

3.1

it is really diheartening to read about these immigrants from this article who drowned.  it makes me feel anxious and upset how the whole ordeal happened.  it is a terrible occurrence that this had to happen at the mediterranean sea.  thankfully there were some survivors.  the fact that babies were lost makes it that much more emotional to read all of this

33

5000

sadness

5

3.3

3.5

I think almost everyone has an opinion on Hilary Clinton over Donald Trump. Many didn't expect Donald to win the election but here we are. So many political scandals unfolded and more continuing to unravel. Many didn't agree with Hilary from the Bengazi investigation and many didn't agree with Trump, but in the end that's just politics.

25

7500

anger

7


Important dates

  • Development phase:
    • December 22: initial training data release
    • Before mid January 2021: launch codalab website
  • Test phase:
    • February 1: test data release
    • February 3: deadline submission final result
  • February 10: deadline system description paper (max. 4p)
  • February 22: notification of acceptance
  • March 1: Camera-ready papers due

Development data

You can access the development data by filling in your credentials at the top of this page. Please note that by downloading the data you agree to the following terms and conditions:

  • The organizers and their affiliated institutions makes no warranties regarding the datasets provided, including but not limited to being correct or complete. They cannot be held liable for providing access to the datasets or the usage of the datasets.
  • The dataset should only be used for scientific or research purposes. Any other use is explicitly prohibited.
  • The datasets must not be redistributed or shared in part or full with any third party. Redirect interested parties to WASSA2021
  • If you use any of the datasets provided in the shared task, you agree to cite the associated papers.
More information will be provided later.