We adapt the experimental setting in Milinski et al (2009) on a “collective-risk social dilemma” to be run with non-monetary incentives with adolescents. Rather than financial endowments, adolescents are endowed with tokens to play a videogame (at the rate 1 token - 1 minute) which is executed from a central computer. The collective risk social dilemma that the group of adolescents faces is the risk of failure of the central computer if a certain level of investment in maintenance (in terms of contribution of tokens to the collective account) is not reached. The dangerous event the group of adolescent faces is not playing with the videogame. We replicate the main findings in Milinsky et al (2009): the inverse relationship between the risk of simulated computer failure and the probability of achieving the target threshold. However, we get that the percentage of groups of adolescents that achieve the maintenance threshold is larger in our experiment that in Milinski et al (2009). There is hope…