The success of perovskite and dye-sensitized solar cells will depend on their stability over the whole life-time. Aging tests are of utmost importance to identify deficiencies and to suggest cell improvements. Here we analyzed the quality of 261 recent aging tests and found serious shortcomings in current practices. For example, in about 50% of the studies only one sample was considered, meaning that the sample size was too small for statistical significance. We propose a new procedure for aging tests based on careful planning and scientific reporting. This includes estimating the required sample size for an aging test and avoiding so-called nuisance factors, i.e. unintended variations always present in real world testing. The improved procedure can provide more reliable information on stability and lifetime, which could contribute to better understanding of degradation mechanisms important for improving these photovoltaic technologies.