However, some of these seminal applications of recursive partitioning methods in psychology also reveal common misperceptions and pitfalls: For example, Luellen et al. (2005) suspect that ensemble methods could overfit (i.e., adapt too closely to random variations in the learning sample, as discussed in detail below) when too many trees are used to build the ensemble – even though recent theoretical results disprove this and indicate that other tuning parameters may be responsible for overfitting in random forests.
