General Instructions

• Only attempt HITs if you are a fluent speaker of English
• Your responses are confidential. Any publication based on these responses will not include your specific responses, only aggregate information from many individuals. We will not ask any information that can be used to identify who you are.
• The questions in this task involve reading English language tweets. Since the tweets used here are a sampling of those on Twitter, particularly from political conversations, some of them may be offensive and/or use coarse language.

I Understand

Task Instructions

Many emotions are expressed in online political conversations, including anger. The goal of this task is to determine how much anger is expressed collectively in a set of political tweets about a focal topic. Rather than judging how much anger is expressed by one user in a single tweet, we will ask you to judge how much anger is expressed together by multiple authors across multiple tweets. Your judgements will help us better understand how emotion is expressed around politically charged topics.

You will be shown 6 tweets at a time and asked to judge on a scale from 1 to 9 how much anger is expressed collectively in those tweets.

Important Notes

• You are judging how much anger is expressed across a whole collection of tweets, not how much anger is expressed in any individual tweet.
• You do not need to know how angry any individual author feels, just how much anger they are all together expressing in their tweets.

Broad Category of Anger

• Bitterness
• Fury
• Incense
• Annoyance
• Irritation
• Offence
• Wrath

That is, these are all feelings that can be reasonably assumed to involve some non-zero amount of anger. Sometimes the tweets may primarily express a different emotion (such as sadness or disappointment), but you will have to judge whether they also express some non-zero amount of anger.

Tweets Indicative of Anger

• Explicit expressions of anger
• Expressions of criticism, negative sentiment, disapproval, etc. of some entity (object, person, etc) from which one can infer that the tweets are likely expressing some non-zero amount of anger.
• Some information that allows one to infer that the tweets are likely to be expressing a non-zero amount of anger.

See an example