| Main features | Reduction approach evidences |
| Binary representation of human actions (failure/success) | An ontological reduction is made when human actions are reduced to two types of actions. |
| Dichotomy of errors (omission and commission) | Tasks are broken into procedural steps through hierarchical task analyses (HTA) |
| Attention on human actions’ observable characteristics and their classification (human error taxonomy) | Taxonomy approach comes from biology and tries to classify entities. Methodological reductionism assimilates a problem to a known conceptual scheme and is expected to make intelligible the basic structure of the present problem. |
| Low attention to cognitive actions (lack of a cognitive model) | Human actions are explained at their basic level (methodological reductionism), observed behavior, or physical actions. Reduction does not imply, in this case, an appeal to other scientific concepts but, rather, to choose a few ideas and try to explain with them all reality or a vast sector of it. “Mind is understood as a box-like construction with a mechanistic trade-in internal representation” |
| Emphasis on quantifying errors by statistical methods | HRA is described in terms of hardware reliability, i.e., human behavior is addressed in terms of mechanical components (mechanistic reductionist assumptions Statistical methods are based on binary logic where membership degrees or degrees of truth are not allowed. Ontological reductionism opposes any form of dualism |
| Indirect treatment of context with PSF | PSFs are basic characteristics (entities) of context. Most theories are based on implicit functions that relate PSFs with error probabilities. Gnosiological reductionism argues that the best strategy is to search for explanations in terms of the basic constituent entities of objects or processes under study. |