Serious Games for Teaching Conflict Resolution: Modeling Conflict Dynamics - Conflict and Multimodal Communication: Social Research and Machine Intelligence - page 446

Table 21.2 Categories and types of data collected during the game (online) and just before

(offline)

Category Data type

Offline Demographic data Age, gender, years in Portugal

Cultural tendency individualist index, collectivist index

Aggression tendency victimization index, aggression index

TKI conflict resolution style accommodation, competition, compromise,

collaboration, avoidance

Online In-game actions Trading accept and reject

rumour and positive message posting
stealing

Self-reported Conflict intensity (11-point Likert scale)

affective state (angry, sad, happy,neutral)
social relationship with others (5-point Likert scale)

as no in-game actions have happened yet. Subsequently, the conflict intensity was
updated whenever any student reported his or her own rating about conflict, on the
completion of a quest, and at the end of the game.
Additional data gathered from the participants included their demographic
background data, a psychological profile, self-reported conflict resolution strategies,
and behavioural data from the game session. The demographic data consisted of
information about participant age, gender, and whether they had been residents
of Portugal for a minimum of 5 years. This information was collected from the
participants' school teachers.
Profile information consisted of a number of psychometric measures obtained
from the following questionnaires (see Table 21.2 ): The Aggression/Victimization
Scale (Dahlberg et al. 2005 ; Orpinas and Frankowski 2001 ), an adapted version
of the Thomas-Kilman Conflict Mode Instrument (Thomas and Kilmann 1974 )
modified to facilitate comprehension by children, and a questionnaire assessing
individualist/collectivist orientation (Triandis 1995 ). The children completed the
questionnaires before interacting with the game (offline).
Children's self-reports were collected through the in-game interface during
gameplay (see online data types in Table 21.2 and the reporting schemes/interfaces
in Fig. 21.4 ). The reports include ratings (0-10) of experienced levels of conflict,
ratings of conflict with other individual players (a variant of a 5-point self-
assessment manikin), as well as indications of the player's immediate emotion (four
affective states to pick from: angry, sad, neutral, happy). In-game behavioural data
was logged automatically including all core game actions (i.e. trading, spreading
rumours, message posting, and stealing).
Finally, sessions were video recorded in order to ensure supporting qualitative
material for the subsequent interpretation of the game data, as necessary.
For the user study, the following protocol was followed: Background data was
collected several days before the actual user study, in order to allow the children