Serious Games para identificação automática dos traços de personalidade do jogador

Abstract

Serious Games (SG) are digital games that playfully use entertainment to achieve specific goals. The personality constructs directly influence the player’s profile and affect their choices. In addition, the Psychological Assessment (AP) presents the need for more reliable psychological instrumentation that allows reaching the trends of personality traits in a reliable way with less fatigue on the part of the respondent. No published studies present a SG tool to identify trends in personality traits using the Big Five Personality Factors model, also known as Big Five Factors (BFF), nor use the Item Response Theory (IRT). ) as the basis of its Intelligent Decision Model (IDM). Therefore, the objective was to develop and obtain evidence of the validity of an SG that automatically identifies trends in the player’s personality profile using an IDM based on the IRT. We carried out a methodological process to achieve the objective. The process considers, first of all, the design of a SG with specific content based on the BFF model. It was necessary to pursue convergent and divergent validity evidence. To achieve that, we performed chi-square and Spearman correlation coefficient analyses considering the SG, the Big Five Personality Factors and its Facets Assessment Instrument (ICGF-F), and the General Health Questionnaire ( GHQ-12). To find the IDM adequacy based on the calibration of the discrimination and difficulty parameters of the gradual two-parameter model of the IRT; the convergent validation of the scores generated by the ICGF-F and the SG. To achieve the goals, we counted a sample formed by people of any gender, over 18 years old, who accepted to participate in the study. During the process, we observed that the analyses conducted to find the evidence of convergent validity were inadequate. They presented unsatisfactory results, so we carried out network analyses that produced adequate results and confirmed the existence of convergent validity evidence, allowing new calibration and consequently the generation of more reliable SG scores. The work process developed in this thesis confirmed the argument that it is possible to identify trends in personality traits from an IDM applied to a SG developed for this purpose. Thus, we produced a SG that can be useful for application in psychology and guiding SG design processes.