Basic experimental research in psychology is based on the assumption that law-like behavior can be observed if the complexity of the human psyche is reduced by the creation of experimental settings in which simple psychical phenomena occur which reflect the effect of an isolated psychological mechanism. However, we show that this assumption does not hold for many phenomena studied in basic experimental psychology because even phenomena that are regarded as simple and fully controllable often fluctuate unpredictably as a function of unintentionally chosen details of the experimental setting. The reason is that in a complex system like the human psyche, even minimal, and from the perspective of the investigated research question irrelevant, differences in the experimental setting can build up to large unsystematic effects. Law-like behavior in experiments could only occur if truly low-level mechanisms were studied in a truly isolated way. However, this is often not the case in current experimental research. One problem is that often fuzzy theoretical terms are used which only give the impression that low-level mechanisms are being investigated, although in reality the complexity of the human psyche is unintentionally brought on board. Another problem is that, unlike in the natural sciences, the mechanisms of the human psyche can only be isolated from each other to a limited extent because the human psyche always reacts as a whole system. If such problems could be overcome, meaningful knowledge could be gained through experimental psychological research. However, the knowledge gained is very limited in terms of its explanatory power for human behavior, as it is only helpful for understanding a very specific aspect of behavior, namely the mechanistic functioning of isolated low-level mechanisms. When it comes to understanding motivated behavior in real life, knowledge about the non-mechanistic functioning of the higher levels of the human psyche is necessary, but this knowledge cannot be gained through the experimental method.