Political education strives to promote students` political competencies, which according to the model of Detjen et al. (2012) incorporate content knowledge, political judgement and political action abilities as well as motivational skills and attitudes. For achieving these goals, high hopes are placed on active learning environments such as simulation games (e.g., Raymond/Usherwood 2013; Guasti et al. 2015; Jones/Bursens 2015). Their anticipated advantages are of particular relevance with regard to problems identified for teaching about the European Union (EU) at school, such as perceived hyper-complexity and perceived distance from everyday life (Oberle/Forstmann 2015). However, there is a profound lack of systematically won empirical evidence regarding the effects of simulation games as a tool of political science teaching.
The empirical study presented here addresses this research deficit: It analyses the effects of a three-hour simulation game of the decision-making of the European Parliament in three different policy areas. It focuses on both the “subjective” evaluation of the simulation games by participants and the “objective” effects of the games on their EU-related political knowledge, motivations (e.g., political interest, internal and external efficacy) and attitudes (e.g., general and performance-oriented EU attitudes; perceived relevance of EU for everyday life), controlling for different background variables (gender, school type, cultural capital, pre-knowledge, pre-interest).
Data were collected in German secondary schools (intervention group N=308, control group N=108) via a pen-and-paper questionnaire in a pre-post design. In addition, face-to-face interviews with participating students were conducted. Political knowledge was assessed with 24 multiple-choice items (building on Oberle 2012; Oberle/Forstmann 2015). Using ConQuest, a one-dimensional Rasch model was applied. Students` evaluation of the game and their political motivations and attitudes (see, e.g., Deutsche Shell 2010; Kerr et al. 2010) were captured by items with a four-point Likert scale. Measurement models, structural equations and class analyses were calculated latently using Mplus 7.4.