What's in a Name?

http://www.sciencedirect.com/science/article/pii/S0747563215301655

We analysed players' self-chosen user names to estimate both age and ANT. Many of these user names contained information that informed us about these parameters. Specifically, players often embed their birth date in their user names (e.g. goodplayer1996) and in a separate analysis we show that these dates are highly-correlated with the self-reported ages of the players in the registration procedure. In addition, many usernames contain explicit or lightly obfuscated expletives, racial slurs and boasts that are clearly designed to attract attention (e.g.‘g0ats3x’). Players must invest some time in generating these ANT names as multi-player online games typically have simple filters in place to block straightforward examples of offensive language.

Once we had identified user names that appeared to contain either age or ANT information, we asked if there was a relationship between ANT or age and the average valence of reports that each player sent or received within the game. We found that both age and ANT are predictive of in-game interaction valences as measured by honors and reports. Importantly, we find this effect for both incoming and outgoing ratings (in other words, ratings generated by a player and directed towards other teammates or, alternatively, ratings generated by teammates directed to a player).