This paper treats the problem of which type of units, individuals or whole societies, should be used when explaining societal phenomena. It is argued that factors operating at the individual level in principle form societies, and that societal phenomena therefore should ideally be explained at this level. However, it is also argued that many societal phenomena cannot in practice be analyzed at the individual level in a clear and strict way, but rather must be analyzed holistically, because it is not (...) known how they are related to individual factors due to the complexity of societies. For societies to be described and analyzed in a clear and strict way, they must as a rule be much simplified, and such simplification is often made easier by abandoning the individualistic perspective. If researchers still wish to describe a connection between individual factors and a societal phenomenon, they often must do so in vague terms. There are also other problems with the individualistic perspective that make it difficult to use. However, using the holistic perspective may be risky, as shown by two hypothetical examples. (shrink)