Serious Games for Teaching Conflict Resolution: Modeling Conflict Dynamics - Conflict and Multimodal Communication: Social Research and Machine Intelligence - page 456

In addition to the features shown in the table, two complementary factors
were computed from the recorded in-game actions and used in the analysis: the
detrimental factor ( D ) and the beneficial factor ( B ). The detrimental factor is the
summation of the counts of the harmful actions that the player was involved in
either as the giver or receiver (Eq. ( 21.1 )). The harmful actions in this study include
rejecting a trade, spreading a rumour, and theft

D D G D C G R C G S : (21.1)

The beneficial factor ( B ) is the sum of the counts of the positive actions that the
student was involved in, such as posting a positive message, accepting a trade, or
being involved in altruistic gifting, either as giver or receiver (Eq. ( 21.2 ))

B D G M C G T : (21.2)

The correlation analysis between conflict ratings, in-game actions, and other
reported features reveals some noteworthy effects. As is evident from Table 21.11 ,
in-game harmful actions ( D ) beneficial actions ( B ) were, respectively, positively
and negatively correlated with conflict ratings. Self-reported anger ( E A ) and
happiness ( E H ) were positively correlated whereas neutral feelings ( E N )were
negatively correlated with conflict ratings. In particular, the positive association
between happiness and conflict escalation is noteworthy. Since the emotion factor
is the summation of all the players' reports, we speculate that the happy emotion
reported during conflict escalation may represent the view of the player who is the
giver of the action rather than the view of the action's receiver. Imagine that a high
degree of conflict arises as a result of a theft. While the victim of the theft may
report an angry affective state (i.e., E A D 1), the other student (i.e., the thief) may
report happiness (i.e., E H D 1). In this case, the conflict escalation is associated with
both of the angry and happy emotions. The children's judgements ( S ) about other
players in the game were also found to be significantly (and positively) correlated
with conflict ratings. There was no significant relationship between the self-reported
emotion of being sad and conflict ratings.

Analysis and Discussion Strong correlations between conflict ratings and in-game
actions and self-reported responses were found. Not surprisingly, negative in-
game actions tended to escalate conflict and positive in-game actions tended to
de-escalate conflict. The student's own judgement about the other three players

Table 21.11 The result of Pearson's r analysis between reported conflict
intensity, children's reported features and in-game actions (number of reports:
2,126)
S D B E A E H E N E S

Social Detrimental Beneficial Angry Happy Neutral Sad

0.44*** 0.23*** 0.15*** 0.15*** 0.08*** 0.10*** 0.02

Significance notation: 0 “***”