What if we could experimentally manipulate characteristics of states, economies and policies and estimate their effects on citizens? This study presents evidence from Novaland, a virtual online democracy with characteristics drawn from German and Romanian welfare states (e.g., level of welfare and corruption). The study consists of an text- and image-based online platform where participants are randomly assigned to different experiences (e.g., income, corruption or unemployment), interact with each other and co-create decisions (e.g., elections and donations) that determine the course of Novaland. For this purpose, we recruited participants via social media networks (N = 347) and collected data in November 2022, which was funded by the European Research Council (ERC). Our results are: (1) The Novaland Experience worked technically well with over 300 participants interacting simultaneously. (2) Participants behaved in an internally valid manner, although there were signs of inattentiveness. (3) Volunteers’ political ideology was reflected in their Novaland behaviour, whereas socio-demographic characteristics showed no effects. Novaland offers the blueprint for innovative political science research on state characteristics and people’s political solidarities.