Newell & Shanks (N&S) raise important questions about whether people possess self-insight into the factors that shape their decisions. Our critique is motivated not by the evidence that was reviewed in the article, but by a large and essential body of literature that was not – literature suggesting that people are unaware of how contextual factors influence their use of information and the eventual decisions they reach.

Suppose that for any decision two broad types of variables come into play: input variables and context variables. Input variables are informational variables that people use in determining what to think or how to act. For example, someone decides to vote for a candidate based on his or her perceptions of that candidate's political leanings, competence, and likability. Context variables, by contrast, are “background” variables that affect the way people think about a particular decision or weight relevant pieces of information. In one context, a voter might weight the candidate's competence over likeability; in another context, the same voter might weight the candidate's likeability over competence.

On our reading, N&S focus their critique of awareness on input variables, making the claim that people have insight into the various inputs that shaped their responses (e.g., Harries et al. Reference Harries, Evans and Dennis2000). Even if we stipulate this as true, research not reviewed by the authors has shown that people remain unaware of how their use of those inputs is influenced by the context in which the decision was made. People may know which input variables shaped their decision, but they still may lack insight into how context nudged them to rely on those variables.

Imagine someone deciding on Kahneman and Tverksy's classic Asian Disease problem. They have a chance to save the lives of 600 people by choosing which of two drugs to recommend. In one version of this problem, they learn that Drug A will save 200 people with certainty and Drug B has a 1-in-3 chance of saving everyone (otherwise, no one is saved). Most people in this situation recommend Drug A. Let's assume they know why: They value the certainty of saving 200 over the uncertainty of Drug B. However, in another version of this problem, the same question is framed in terms of losses, such that Drug A will result in the death of 400 people with certainty and Drug B carries the 2-in-3 chance that everyone will die (with a 1-in-3 chance that no one dies). Asked with this framing, people now prefer Drug B, and again can articulate why – it's better to take a chance on saving everyone than kill 400 for sure. Thus, depending on the question asked, people arrive at two very different answers, and they can describe accurately why they decided as they did. To our minds, though, the interesting self-insight question still remains: Do people know how the context of the decision affected them? Do they know that their use of the information and eventual decision was shaped in large part by the question framing?

Emerging evidence reveals that people tend to have inadequate insight into the impact of significant context variables on their decisions. In this work, people are asked how their preferences or decisions would change if the circumstances surrounding their decisions were different. For example, they are asked how likely they would be to help another person if placed in a group that passively did nothing versus in a room alone with the other person—the classic bystander intervention effect. People claim that being in a group would do nothing to their tendency to help, but in reality people are much less likely to help in those circumstances (Balcetis & Dunning Reference Balcetis and Dunning2013; see Epley & Dunning Reference Epley and Dunning2000, for conceptually similar results).

Similarly, people fail to understand how emotion changes their preferences (Van Boven et al. Reference Van Boven, Loewenstein, Dunning, Nordgren, Olson and Zanna2013). Many people will say, when asked hypothetically, that they would be willing to dance in front of a classroom full of peers for $5; a substantially smaller number actually accept this opportunity when faced with the real choice. What may explain this difference between hypothetical and real decisions is that people fail to understand how the context of the decision – real versus hypothetical – changes their decision-making process. In the former case, people weight the $5 compensation over their fear of social evaluation. In the latter, that weighting flips such that the fear of social evaluation trumps the lure of compensation (Van Boven et al. Reference Van Boven, Loewenstein, Welch and Dunning2012).

A range of contextual variables known by social psychologists to affect decision making appear to go largely undetected by decision makers themselves. People do not recognize how sensitive their decisions are to social norms (Nolan et al. Reference Nolan, Schultz, Cialdini, Goldstein and Griskevicius2008). They believe that they are more immune to conformity pressures than they really are (Pronin et al. Reference Pronin, Berger and Molouki2007). And they similarly overestimate their resistance to media influence (Davidson Reference Davidson1983; Douglas & Sutton Reference Douglas and Sutton2008; Reference Douglas and Sutton2010; Perloff Reference Perloff1993; Reference Perloff1999).

Or, consider the following: We (Helzer & Dunning Reference Helzer and Dunning2013) have asked people to make a decision between X and Y, with knowledge that most of their peers think X. The opinions of their peers have an effect on the participants' ultimate decision, but do the participants know this? After they reach a decision, we ask them what their decision would have been if, instead, their peers had favored Y. As in the studies cited above, people consistently underappreciate the impact this shift will have on their decisions, thinking that their decisions are insensitive to this changing context, when, in fact, they are quite sensitive.

Thus, even if people are experts at understanding how various input cues landed them at a particular decision (something N&S refer to as cue utilization), we assert on the basis of relevant empirical findings that they fail to appreciate how context influences the weight given to various decisional inputs. This leaves them with some recognition of how they arrived at their ultimate decision, but not necessarily why the decision unfolded as it did.