For about eight decades, research instruments in the social sciences have been orbiting around Likert’s proposal for his famous response scale. Before him, and also after he managed to impose it, many researchers tried to find a better solution. This, however, has proven difficult. While solving methodological problems for measuring concepts, concentrating all responses into only five categories brings major disadvantages as well: it has extremely low variation, it does not produce metric scores unless combined with similar items, and it cannot be used as such for advanced statistical analysis. In this article, we propose using a continuous response scale as a solution to each of these problems. In our opinion, the possible application of this solution has an extremely high potential to advance social science research methodology.