Positive Chat in Online Games Can Be Hard to Read

The play gives the words a setting

Positive chat in online games can carry more than its words reveal. Imagine a team arena where a player types “That was sharp” just after an ally blocks a scoring attempt. The same line appears in another round after an ally mistimes the block. A reader who saw the play has a clue about what the line refers to.

Even the second line leaves room for encouragement. The sender could be praising an earlier move or trying to steady the team after a mistake. A short chat record leaves those possibilities open. The event the sender had in mind and the event the reader remembers may differ.

A survey presented two sets of chat

Poeller and colleagues' 2023 survey analyzed 959 League of Legends players. Random assignment determined whether participants saw ten positive or ten hostile chat screenshots. They rated how representative the set seemed and could leave written responses.

In the invented arena, a reader may recognize friendly wording and still hesitate to say that a whole set of friendly exchanges resembles their matches. Someone who usually plays with friends might bring a different reference point from someone who queues alone. A study can ask about the examples and the reader's reference point separately.

Typicality depends on the reader's reference point

Mean representativeness ratings were 5.1 out of seven for the hostile set and 2.6 for the positive set, a statistically significant difference.

A participant could remember one match full of cheerful messages and still rate a collection of them as unusual. Those responses describe perceived typicality. Measuring how often each kind of line appears would require a sample of actual match chat with defined rules for counting it. The result could depend on whether the unit is a message, a player or a match.

Written doubts and sarcasm ratings capture different responses

Among 348 people who commented after seeing positive examples, researchers coded responses from 303 as downplaying those examples. That category included descriptions of the chat as rare. A combined category for suspected fabrication or sarcasm included 66 commenters. Separate sarcasm ratings showed no significant group difference.

Someone can describe a message as unusual without deciding whether it was sincere. In the arena example, a reader might recognize the praise and still wonder which move prompted it. Their uncertainty could be useful on its own. A question that demands a verdict about intent would lose that information unless readers could also explain what they needed to know.

The survey's setting matters

Participants responded in German and mostly played on the Europe West server. Researchers selected the screenshots and capped each written answer at 100 characters.

A player reading the line while defending a goal would also have to divide attention between chat and play. A survey answer gives the reader a different task, with space to consider how the line sounds. The two situations could yield different reactions. Actual changes in trust or in how long someone keeps playing would require observations suited to those outcomes.

Testing how the previous play affects a reading

A follow-up could show the two fictional rounds with identical wording and different preceding plays. Keep the sender relationship the same. Ask readers which action they think the line describes, what they believe the sender intended and how sure they are. Those answers would show where interpretations begin to diverge.

If readers choose different actions as the subject of the compliment, the next comparison could show a line that names the action. If they agree on the action but differ on intent, the study could instead vary the conversation immediately before it. Each comparison has a specific uncertainty to investigate.

Keep exploring

FOLLOW THE SOURCE

Sources & further reading

Have a correction? Tell our editorial team.

✦ NICE. YOU LEARNED SOMETHING.