The foundational finding

Leonid Rozenblit and Frank Keil’s research on the illusion of explanatory depth established the core experimental finding. Participants asked to rate their understanding of everyday mechanisms, including bicycles, toilets, and zippers, produced moderate to high self-ratings. They were then asked to provide a detailed causal explanation of how the mechanism worked. The explanations were dramatically impoverished relative to the initial confidence ratings; participants who believed they understood how a bicycle worked could not accurately describe the causal mechanism that makes a bicycle steer or remain upright. Following the explanation attempt, they revised their understanding ratings sharply downward.

The bicycle example is particularly useful because it strips the IOED of the excuse of technical complexity. Most adults have extensive experience with bicycles, have seen them explained, and operate in social environments where bicycle mechanics are accessible. The failure to explain is not a failure of access to information. It is the product of a specific cognitive pattern in which familiarity with a thing, exposure to its explanation, and the social availability of expertise about it generate a subjective experience of understanding that is not accompanied by genuine explanatory depth.

The distributed cognition mechanism

Frank Keil’s extended research on the IOED documented the specific mechanism that produces it in social environments: people systematically confuse their access to explanations in their social and information environment with their own possession of those explanations. The founder who is surrounded by advisors who understand the unit economics, investors who have modelled the market size, and team members who understand the technical architecture does not independently possess that explanatory depth. The social availability of the explanation, its accessibility in conversation and in the documents and pitch materials that summarise it, generates the subjective experience of understanding that the community’s explanatory depth provides rather than the founder’s own.

The distributed cognition mechanism has a specific commercial implication: the confidence the founder has in their understanding of their own business is partly a function of the quality of the team and advisors they have assembled, not only of their own genuine comprehension. The well-advised founder surrounded by deep expertise in every relevant domain may have the highest IOED in the room, because the social accessibility of that expertise is maximally available to generate the subjective experience of understanding without the requirement to produce the explanation independently.

The Dunning-Kruger profile

Justin Kruger and David Dunning’s research on metacognitive accuracy predicts the specific confidence profile of the IOED: the partial understander of a complex domain generates higher confidence than either someone with no understanding or someone with genuine expertise, because the partial understander has the vocabulary and surface-level familiarity of the domain without the awareness of its complexity that deep expertise reveals. The expert knows what they do not know; the IOED sufferer does not know what they do not know, and the absence of that awareness is itself the source of the confidence.

For the founder who has spent months in their domain, the Dunning-Kruger profile predicts the specific moment when the IOED is most active: not at the beginning, when genuine ignorance is obvious, and not after years of operational experience have revealed the system’s genuine complexity, but in the early-to-middle period when surface-level familiarity has been acquired and the domain’s vocabulary is available but the genuine causal complexity has not yet been encountered in enough detail to register as unknown. This is the period when most of the significant strategic decisions are being made.

You might also like
Working with your thinking patterns rather than against them and what the research supports as sustainable cognitive practice
The attempt to eliminate unwanted thinking patterns is not just difficult. The research on thought suppression suggests it reliably makes them worse. The alternative is more effective and considerably less exhausting.
The psychological impact of individualism versus collectivism on team dynamics and entrepreneurial motivation
The individualism-collectivism dimension is the most thoroughly research-documented cultural difference in psychology. It predicts not only the broad social values of different national cultures but the specific, commercially consequential ways that team members work together, what motivates them, how they understand success and failure, and what happens when the entrepreneurial assumptions of one self-construal system are applied to people whose entire psychological organisation is built on the other. For the cross-cultural entrepreneur, understanding the depth of this difference — and recognising that it reaches below conscious preference to the most fundamental level of how the self is psychologically constituted — is the prerequisite for managing across it.
How culture is created — the mechanisms through which founder behaviour becomes team norm
Culture is not installed through values statements or culture decks. It is transmitted continuously, automatically, and primarily through what the founder does — and every person in the organisation is receiving and processing that transmission whether the founder intends it or not.
The psychology of good “enough decisions” versus “perfect ones”
What the pursuit of perfectionism costs.

The customer behaviour domain

The IOED’s most commercially consequential domain for the founder is customer behaviour. Research applying the IOED framework to entrepreneurial assumptions about customer motivation consistently documents that founders’ confidence in their understanding of why customers buy, what drives their decision process, and what would change their behaviour is substantially higher than the explanatory depth those confidence ratings correspond to. The founder who believes they understand their customer cannot, in the Rozenblit and Keil explanatory depth test, produce an accurate mechanistic account of the causal sequence from the customer’s problem recognition through to their purchase decision.

The commercial consequence is the specific class of product, pricing, and positioning errors that the IOED produces in the customer domain: features built for an assumed motivation that is not the actual motivation, pricing set at a level that reflects the founder’s model of the customer’s willingness to pay rather than the customer’s actual decision process, and messaging constructed for the customer behaviour the founder’s model predicts rather than the customer behaviour that the genuine causal model would reveal. These are not failures of effort or intelligence; they are the predictable commercial output of the IOED in the domain where it is most consequential.

Competitive dynamics and operational dependencies

The IOED operates with equal force in the competitive dynamics and operational dependency domains, where the genuine causal complexity is higher and the social availability of simplifying explanations is correspondingly greater. The founder’s confidence that they understand why their competitors are positioned as they are, what their competitive response to a market change would be, and which operational dependencies in the current business model would fail under growth pressure is generated by the same distributed cognition mechanism that produces the customer behaviour overconfidence. The explanatory depth is available in the environment; the independent possession of it is not.

The operational dependency version of the IOED is specifically relevant to the growth contexts that the founding process generates. The business model that appears to the founder to be well-understood in its causal structure reveals its genuine complexity at the point of scaling, when the dependencies that were stable at small scale become the constraints and failure points that the founder’s model had not incorporated. The IOED does not prevent operational problems; it prevents the recognition of operational dependencies before they become operational problems.

The accountability correction

Philip Tetlock’s accountability research predicts the most reliable available correction for the IOED: the requirement to explain one’s reasoning to an audience that will genuinely evaluate it forces the surface-level understanding into the explanatory depth test that reveals its limits. The founder who is required to explain their customer model, competitive advantage, or operational causal structure in sufficient detail for a sophisticated questioner to probe it is performing the Rozenblit and Keil test in its most commercially available form. The board meeting, the investor due diligence conversation, and the team strategy discussion are all accountability mechanisms; their quality as IOED corrections depends on whether the questioning genuinely requires mechanistic explanation or accepts the surface-level familiarity that the IOED provides.

The practical implication is that the founder who can explain their business to a sophisticated questioner who asks “why?” at each level of the explanation until the genuine causal mechanism is reached, or until the genuine limit of the founder’s understanding is revealed, has performed the most directly available IOED correction. The explanatory depth test does not require a questioner; it requires the discipline of the causal “why?” applied to the founder’s own confident assumptions until the explanation either grounds out in genuine mechanism or reveals the point where familiarity ends and genuine understanding has not begun.

If the experiences described in this article are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Rozenblit, L. and Keil, F. (2002). The misunderstood limits of folk science: An illusion of explanatory depth. Cognitive Science, 26(5), 521-562. Keil, F.C. (2003). Folkscience: Coarse interpretations of a complex reality. Trends in Cognitive Sciences, 7(8), 368-373. Kruger, J. and Dunning, D. (1999). Unskilled and unaware of it. Journal of Personality and Social Psychology, 77(6), 1121-1134. Tetlock, P.E. (2005). Expert Political Judgment. Princeton University Press. Johnson-Laird, P.N. (1983). Mental Models. Harvard University Press. Sloman, S. and Fernbach, P. (2017). The Knowledge Illusion. Riverhead. Meadows, D.H. (2008). Thinking in Systems. Chelsea Green.