Interventions use data as close to what actually happened as possible, such as tape-recorded meetings and conversation, rather than inferences or retold stories, because such data connect with theories-in-use and make it hard for people to disown their causal responsibility for what occurred. Argyris's research-intervention goals are to discover how far participants' theories-in-use fit Model I, how they reason defensively, what rules keep them unaware of the discrepancies, and what organizational defensive routines exist, while any bewilderment or frustration produced is treated as further data.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.