There is a particular kind of mild domestic spectacle that occurs in the living rooms of people who watch Jeopardy. The contestants on the screen are facing the question. They are standing behind their podiums, lights on them, audience in the room, host reading the clue at a particular paced cadence that has been calibrated over decades of broadcast television to produce a specific kind of mild competitive tension. The contestants have a few seconds. Most of the time, in any given clue, one of them buzzes in and produces the correct response. Sometimes, however, all three contestants pause. The pause is the moment the spectacle becomes interesting. In the pause, someone on the couch, watching the broadcast, will, with a particular kind of slightly indignant confidence, shout the answer at the screen.
The shouter, in this moment, is in a particular psychological configuration. The configuration is the implicit conviction that, if the shouter had been on the stage, they would also have produced the answer, and produced it more quickly than the contestants who are visibly struggling. The conviction is, on close examination, often inaccurate. The shouter is not, in most cases, smarter than the contestants. The shouter is, more accurately, working with a less filtered version of the question than the contestants are working with, and the gap between these two configurations is what the article is about.
What the contestant is actually processing
It is worth being precise about what is happening, cognitively, in the moment a contestant on stage is facing a Jeopardy clue. The cognitive load is, on close examination, considerably larger than what the shouter on the couch is processing.
The contestant is, first, parsing the clue itself. Jeopardy clues are, by design, multi-layered. The clue is rarely a direct question. The clue is, more accurately, a small puzzle that one has to disassemble in order to identify the actual question being asked. The disassembly takes cognitive bandwidth. The bandwidth is being deployed in real time, under the time pressure of the show’s format.
The contestant is, second, deciding whether to buzz in. The buzzing-in decision is, in the actual mechanics of the show, the most important cognitive decision in any given clue. The contestant cannot, by the rules of the show, buzz in before the host has finished reading the clue. The buzzing-in is, accordingly, calibrated to a specific millisecond window. The contestant is, in some real way, expending considerable cognitive resources on the buzzing-in decision itself, in parallel with the parsing of the clue and the retrieval of the answer.
The contestant is, third, retrieving the answer under conditions that the wider cognitive literature has identified as among the most difficult conditions for retrieval. Research on choking under pressure has established that high-stakes performance conditions impose a particular kind of cognitive load that reduces the available working memory for the actual task. The paradoxical finding from this literature is that the effect is, in fact, more pronounced in individuals with high working memory capacity, because their characteristic style of solving problems relies more heavily on working memory in the first place. The high-capacity individual, accordingly, has more to lose when the pressure starts consuming the working memory they normally have access to. The contestant on stage, who has, by the structural fact of having qualified for the show, demonstrated a high working memory capacity, is operating in conditions that specifically reduce the very resource they would otherwise be deploying.
The contestant is, fourth, managing the social presence of the other contestants and the audience. The social presence is, on the available evidence, an additional cognitive load. Research on the neuropsychological mechanisms of choking has identified social monitoring as a particular contributor to the performance decrement, because the part of the cognitive apparatus that is being deployed to track the social situation is, in some real way, drawing on the same resource that the actual task requires.
The cumulative cognitive load on the contestant is, accordingly, considerable. The contestant is parsing, deciding, retrieving, and socially monitoring, all in parallel, in a window of a few seconds, under conditions that have been specifically engineered to maximize the pressure on the cognitive apparatus.
What the shouter is actually processing
The shouter on the couch is, on close examination, working with a considerably less filtered version of the same situation.
The shouter is, first, not under the time pressure of the show’s format. The shouter has, in practice, as long as they need to produce the answer. The “shout” is, in most cases, occurring somewhere between two and five seconds after the host has finished reading the clue, which is, by the standards of the show, an enormous window. The contestants, by the time the shouter shouts, have already had to decide whether to buzz in. The shouter’s window is, accordingly, several times larger than the window the contestants are operating in.
The shouter is, second, not managing the buzz-in decision. The shouter does not have to time the buzz-in to the millisecond window the show requires. The shouter does not have to commit to producing an answer before they know whether they have one. The shouter is, in some real way, free to wait until the answer has fully formed before producing it. The contestants do not have this option.
The shouter is, third, not under the social pressure of standing on a stage in front of an audience and three cameras. The shouter is in their living room, possibly in their underwear, with no consequences attached to either being right or being wrong. The cognitive load associated with social monitoring is, in the shouter’s case, essentially zero. The retrieval bandwidth is, accordingly, fully available for the actual task of retrieving the answer.
The shouter is, fourth and most importantly, not facing any consequence for being wrong. The shouter who shouts a wrong answer at the screen pays no cost. The shouter who shouts a right answer receives, internally, a small piece of confirmation that they are clever. The asymmetry between these two outcomes is what produces, over time, the shouter’s accumulated sense of being smarter than the contestants. The shouter remembers the times they were right. The shouter does not, in most cases, remember the times they were wrong. The selective remembering produces, in the shouter’s internal accounting, a track record that is, on examination, considerably better than the actual one.
What this tells us about the gap
The gap between knowing something and retrieving it under pressure is, on the available evidence, one of the most consistently underestimated cognitive variables in adult life. The wider culture tends to treat knowledge as a roughly binary thing. Either one knows the answer or one does not. The treatment is, on close examination, not quite accurate. The actual structure is more like the following.
One can know something in the sense that the information is, in some form, present in one’s long-term memory. The presence is verifiable by various offline tests. Show the information to the person in the right cue, and the person will, in most cases, recognize it.
One can know something in the sense that one can retrieve it under low-pressure conditions. The retrieval requires the cognitive apparatus to bring the information from long-term memory into working memory, under conditions where the working memory has the available capacity to do the bringing. Most adults can do this for most of the information they know, given enough time and the absence of pressure.
One can know something in the sense that one can retrieve it under high-pressure conditions, in real time, with social monitoring, with consequences for failure, and with the additional cognitive load of various parallel demands on the apparatus. This is, on the available evidence, a structurally different capacity from the previous two. Research on classifier performance under pressure has shown that the same person, with the same underlying knowledge, can perform dramatically differently across these conditions, and that the differences are not noise. They are systematic features of how the cognitive apparatus operates under different loads.
The Jeopardy contestants are operating at the third level. The shouter on the couch is operating at the first or second level. The shouter is not, on close examination, smarter than the contestants. The shouter is, more accurately, operating with a much less constrained version of the cognitive task. The constraint is what separates the couch from the stage. The constraint is what makes the stage difficult in a way that the couch is not.
What this implies, more generally
The implication of this, for the wider culture’s reading of the various performances it observes, is uncomfortable. The wider culture tends to evaluate cognitive performance under high-pressure conditions and to attribute the results to the underlying cognitive capacity of the performer. The attribution is, on close examination, partially wrong. The performance under high pressure is the joint product of the underlying capacity and the various pressure-related variables that affect how much of the capacity is, in real time, available for deployment. Two people with identical underlying capacity can perform very differently under pressure, depending on how their particular nervous systems respond to the pressure conditions. The wider culture, in attributing the performance entirely to the underlying capacity, is missing the structural variable that is, in many cases, doing more of the explanatory work.
This applies, on the available evidence, to a great many cognitive performances beyond Jeopardy. It applies to interviews, where the candidate’s interview performance is being evaluated as a proxy for their underlying capacity. It applies to academic examinations, where the student’s exam performance is being evaluated as a proxy for their understanding. It applies to the various small daily performances of adult life, in which one’s ability to deploy one’s underlying capacity under pressure is, in most cases, doing more of the visible work than the underlying capacity itself.
The shouter on the couch is, in some real way, the perfect illustration of what the underlying capacity looks like when it is unburdened by the pressure conditions. The shouter has the capacity. The shouter does not have, in any single instance, the conditions under which the capacity is being tested in the form it is tested on stage. The couch is not the stage. The shouter, accordingly, is not the contestant. The shouter’s success in their own living room is, in some real way, the structural artifact of the comparison being made without the pressure that is, on the available evidence, what separates ordinary cognitive performance from the kind that the wider culture, in evaluating it, treats as the genuine measure of capacity.
The contestants on stage are, by every honest accounting, doing something considerably harder than the shouter is doing. The doing is what the shouter, in most cases, has never had to do, and what the shouter, if placed on the stage, would in most cases discover is considerably harder than it looked from the couch. The discovery would be, in some real way, the most useful piece of cognitive humility available to anyone who has spent decades shouting answers at television screens. The discovery is not, on the available evidence, one that most shouters ever get to make. The couch protects them from it. The protection is, in some real way, what allows the shouting to continue.
