Introduction

The idea that individuals make decisions by “satisficing” rather than maximizing began with Herbert Simon. In his autobiography, Simon (1991) describes its origins in an undergraduate term paper he wrote on the budget process for public recreation in Milwaukee. Finding that decision makers did not view or reason about the division of funds in the ways neoclassical economic theory predicted they should, that there were limits seemingly to their motivation and capacity to reason rationally, he began to ask how individuals actually process information and make decisions under conditions of “bounded rationality.” Simon’s answers to this question developed in several papers published in the 1950s, in which he reiterated that the notion of maximizing (or optimizing), which was dominant in economics and would soon exert similar influence on political science, is an unrealistic portrayal of the way in which decisions are generally reached (Simon, 1955; see also March & Simon, 1958; Simon, 1956, 1957). Rather, individuals are “heuristic problem solvers” (Simon, 1991, p. 163). Moreover, it is not even the case that people make decisions “as if” they are utility maximizers: the maximizing idea also does not have predictive validity. Instead, Simon (1956) contended that

Since the organism, like those of the real world, has neither the senses nor the wits to discover an “optimal” path—even assuming the concept of optimal to be clearly defined—we are concerned only with finding a choice mechanism that will lead it to pursue a “satisficing” path, a path that will permit satisfaction at some specified level of all of its needs. (p. 136)

Simon argued that individuals begin with aspirations—minimum thresholds that must be reached to be satisfied with a decision. Rather than conducting a full search of the universe of alternatives open to them and choosing the one that provides the greatest benefits relative to costs, individuals consider alternatives sequentially until they reach one that meets their aspiration level. At that point the consideration of options ends and the decision is made: “The search is incomplete, often inadequate, based on uncertain information and partial ignorance, and usually terminated with the discovery of satisfactory, not optimal, courses of action” (Simon, 1985, p. 295). The individual has satisficed rather than maximized: the decision may coincide with the decision he or she would have made if maximizing, but much (most) of the time it will not.¹ Finally, rather than reflecting irrationality, satisficing is boundedly rational in that it is still based on expected utility but that utility stems from limited information search.²

As this article outlines, in addition to computer science, economics, evolutionary biology, marketing and advertising, philosophy, and psychology, among other fields, this deceptively simple idea has influenced thinking in a range of subfields in political science from elite to mass decision making, as well as one of the core methods of the social sciences: survey research.³ In addition to providing explanatory power for how a variety of political decisions are made by different kinds of decision makers, research on satisficing has turned to questions of whether satisficing and maximizing are personality traits, implying less variation within individuals from the interaction of cognitive limitations, motivation, and task difficulty than Simon’s initial formulation of the idea perhaps suggested. This has in turn led to consideration of whether satisficing or maximizing results in systematically better decisions, with which individuals are more confident and comfortable. This article surveys the current state of knowledge about satisficing in the political world, across a range of groups (from elites to voters) and a range of decisions (from conflicts to vote decisions to survey responses). It then identifies and discusses some of the problems with research on satisficing before concluding by considering opportunities for future research.

Satisficing and Heuristics

The notion of satisficing is based on two central insights: acknowledgment that the cognitive limitations of individuals constrain their capacity to search for information and, concomitantly, that individuals vary in their motivations to consider a greater or lesser range and depth of information before coming to a decision, which will also be influenced by task difficulty. As a result, given a decision whose demands engage cognitive limitations (clearly, if the decision at hand is such as to render cognitive limitations irrelevant, then the possibility of satisficing does not arise), individuals will often draw on heuristics or shortcuts, rules of thumb that allow them to reach decisions that are “good enough” rather than optimal.

Why, when, and how do people satisfice? The cognitive and motivational limits of individuals are well established; the question then becomes how people form judgments and make decisions within those constraints. The idea that individuals satisfice is consistent with psychological models of judgment, in which, for example, attitudes are composed of

[a] latitude of acceptance [that] includes the person’s most preferred position but also includes the range of other opinions on an issue that the person finds acceptable . . . [and a] latitude of rejection [that] comprises the range of opinions that the person finds objectionable. . . . If the position advocated by a communication falls within a person’s latitude of acceptance, assimilation occurs, but a position that falls within the latitude of rejection is contrasted.

(Petty & Cacioppo, 1981, pp. 101–102)

Considering a problem and making a decision involves framing the problem, memory search, and evaluation (Newell & Simon, 1972). The pertinent question is what is involved in that framing and in deciding on an aspiration level deemed “good enough.”

Prospect theory provides one influential set of answers. It generally offers an explicit aspiration level, for example, a return to the status quo when individuals are in the domain of losses, as well as examples of heuristics that influence the framing of the decision. Those heuristics include availability (how easily examples come to mind), representativeness (whether there are examples that appear similar on salient criteria and thus place them in a category from which the individual can infer desirable actions based on schemas), and simulation (how easy it is to imagine one outcome rather than another). In the political realm, such as in evaluating candidates, these heuristics translate to considerations like party, ideology, familiarity, and person stereotypes. A voter may not know enough about the candidates, or be sufficiently motivated to make a detailed evaluation of each one, so perhaps it is enough to know that there is a Republican or Democrat on the ballot, find that candidate, and cast a vote accordingly.

Satisficing and Elite Decision Making

We may think of elite decision makers in government or legislatures as possessing more motivation and cognitive capacity to consider the full range of possible options when making decisions, but what applies to ordinary individuals—limited cognitive abilities, information overload, and task complexity—is both still relevant to elites and exacerbated by decisions that often have to be made under time pressure and that can have life and death implications. Certainly, elite decision makers are not rational actors who “gather as much information about every conceivable alternative course of action as they can before making a decision” (Lau & Redlawsk, 2006, p. 225). Indeed, much of the intersection of political psychology and international relations concerns how and why elite decision makers often fail to consider the full range of alternative pathways and the biases that lead to flawed decision making.

The “cybernetic” theory of elite decision making explicitly draws on Simon’s work to argue that elite choices are the result of an initially limited consideration of alternatives and then of choosing the first one that is good enough. This work also suggests that elite evaluations of situations, leading to decisions, will be quite stable because decision makers will stick with established paths and only deviate from them when new information or a radically new context demands it. Cybernetic theory has been used to explain the way in which elite decision makers, particularly in the United States, from presidents (Ostrom & Job, 1986) to the Supreme Court (Hagle, 1991), process information and reach decisions on issues such as security (Steinbruner, 1974). More recently, attention has turned to other noncompensatory decision making processes—in which alternatives are not even considered if they do not satisfy salient criteria regardless of what other benefits they might offer—that appear to characterize a great deal of elite decision making, suggesting that elites often also make choices that are “good enough” within a set of bounded criteria. Much of this research concentrates on the influences on alternatives that are considered acceptable and alternatives that are simply not considered.

Prospect theory (Kahneman & Tversky, 1979; Tversky & Kahneman, 1992) also extends Simon’s ideas of bounded rationality and satisficing, suggesting that the satisficing actor will make different decisions when in the domain of losses versus the domain of gains. A satisficing actor in the domain of losses will be more inclined to be risk acceptant and choose a risky course, whereas a satisficer in the domain of gains will be risk averse and therefore only likely to consider alternatives that limit his or her exposure to risk. In either case, how the problem is framed for the actor also means that some options are simply not considered. In addition, of those options that remain, the maximizing rather than the minimally acceptable option will not necessarily be chosen. Farnham (1992) argues that prospect theory accounts for Franklin Roosevelt’s change of policy to a more interventionist approach during the Munich crisis, and Haas (2001) states that both Kennedy and Khrushchev’s behavior conformed with the expectations of prospect theory during the Cuban Missile Crisis, during which both leaders made risky and suboptimal decisions when in the domain of losses.

In addition, Kahneman and Tversky’s representative heuristic lies behind the analogical reasoning that characterized the limited consideration of alternatives in Lyndon Johnson’s decisions surrounding “Rolling Thunder” in Vietnam in 1965. Johnson and his principal advisers were so persuaded by the similarities of Vietnam to the Korean War of the previous decade that alternative courses of action that were inconsistent with that analogy, such as withdrawal or sticking with the status quo, were neither considered nor understood (Khong, 1992). Having eliminated those options, Johnson and his advisers made a choice from those available that satisfied their aspirations at that time. Similarly, for Jimmy Carter in 1980, the analogy of Israel’s rescue mission at Entebbe assumed such resonance that diplomatic routes to ending the crisis were no longer considered (Houghton, 1996), while Secretary of State Madeline Albright’s experience as a child in 1930s Czechoslovakia is argued to have made the analogy of Munich particularly salient to her during the Kosovo crisis in the 1990s, precluding softer diplomatic approaches to Milosevic (Hehir, 2006).

Whether decisions such as Rolling Thunder and the Iran hostage rescue mission were the product of satisficing is, however, an open question: while they are consistent with satisficing in the failure of the decision makers to consider the full range of alternatives, to be satisficing the order in which alternatives were evaluated becomes important; there needs to be evidence that the consideration of alternatives stopped with the first set of options that surpassed the aspiration level. Accounts of such decisions are often also consistent with other decision making strategies such as attribute-based “elimination by aspects” or with poliheuristic reasoning in which decision makers first eliminate options that do not satisfy fundamental criteria—usually based on political considerations such as enhancing the chances of reelection—and then make the optimal choice on those criteria from those remaining. That is, their search does not end with the first choice that is good enough, nor need it have the highest expected utility based on all criteria, just on the criterion or criteria that are of critical importance to the decision maker (Mintz, 1993, 2004; Mintz & Geva, 1997). Indeed, the poliheuristic framework has been applied to some of the same examples of foreign policy decision making as prospect theory, such as the Iran hostage rescue mission (Brulé, 2005) and Kosovo (Redd, 2005). The broader point remains, however, that elite decision makers in government are also likely to consider a circumscribed range of alternatives and to satisfice, at least sometimes, due to cognitive and informational limitations and the pressures of time.

In these psychological accounts of elite decision making, much of the emphasis is on how the (mis)use of heuristics leads its protagonists astray. This is likely to become more pronounced during conflicts as time, moral dilemmas, and the limited duration of rally-round-the-flag effects add to the pressure on leaders (Stanley, 2018). In other areas of political science, however, the emphasis is less on decision making flaws than on how satisficing helps to explain empirical regularities in the behavior of elites. For example, the “political business cycle” implies that maximizing incumbent governments will try to maneuver key economic indicators such as growth in gross domestic product, inflation, and unemployment so that they are moving in positive directions for the incumbents as an election approaches. However, the empirical evidence indicates that incumbent governments do not always engage in this kind of behavior. A satisficing account says that this is because the government’s aspiration of reelection means that if it is already at a level of popularity that does not demand efforts to boost the economy, it will not do so:

If its popularity is high, the government pursues its ideological goals; if its popularity is low, it concentrates on securing reelections. That is, politicians attempt to affect macroeconomic variables before election time only if previous macroeconomic outcomes threaten their political popularity and their reelection chances.

(Davidson et al., 1992, p. 22, italics in original)

In U.S. elections, changes in inflation and unemployment are indeed consistent with satisficing behavior rather than with an unwavering political business cycle. Similarly, the well-known median voter theory predicts that parties will converge on the median voter, which is usually also taken to imply policy moderation. Yet in reality, parties frequently do not seem to try to locate the median voter and sometimes diverge rather than converge: If incumbents and challengers recognize that voters are satisficers rather than maximizers, such behavior makes sense (Bernhardt & Ingeman, 1985).

Satisficing may also enter into the decision making of legislators (Weaver, 2018). If good public policy is one of the primary goals of legislators, while some legislators may maximize in pursuing their policy objectives, many are likely to use heuristics that allow them to make policy choices that are good enough rather than optimal. Similarly, seeking reelection means paying attention to some voter signals and not others due to time pressures (Henderson & Brooks, 2016). It would be fair to say, however, that satisficing among legislators has received far less attention than satisficing among other elites or in the mass public.

Satisficing and Mass Decision Making

Simon’s puzzle has also informed the considerable body of research that has followed Converse’s (1964) seminal work on what lies behind mass political attitudes. The tenor of much of that research is not to argue that Converse was wrong—the mass public displays high levels of ignorance about politics and mostly does not reason from abstract principles such as liberalism or conservatism, nor does it possess “belief systems”—but to offer a reinterpretation of what such mass ignorance implies. For the public, aspiration levels for political decisions do not generally warrant full consideration of alternatives and extensive cognitive effort. Instead, the public makes decisions that are “good enough” using an array of cues and heuristics. But what is the evidence, and what are the consequences of mass satisficing?

One of the principal foci of those interested in mass decision making is elections, in which there are two main topics of scrutiny: vote preferences and turnout. Satisficing offers insights into both. For vote preferences it has implications for voter knowledge and for how voters make decisions. It accounts for the lack of political knowledge observed by Converse (1964), suggests that individuals will rely on inferences and heuristics from incomplete information in order to make evaluations, and, given variation in motivation, that this will be increasingly true as races and candidates become more obscure. It also indicates change over time if information becomes easier to acquire and motivations to engage with it change.

In terms of decision making, simple differences such as ballot order will have consequences if individuals consider parties or candidates sequentially until they find a choice that meets or surpasses their minimal desires (Miller & Krosnick, 1998). Having met their aspiration level, the further cognitive costs of considering additional potentially acceptable candidates down the list, and then endeavoring to distinguish the best candidate among those deemed acceptable, outweigh the benefits. Such behavior may be considered an undesirable outcome of satisficing but (a) the effects are small and (b) research on information processing and vote preferences suggests that a satisficing decision rule leads more voters to “vote correctly.”⁴ Indeed, the result is choices that seem close to, or even the same as, those that would be made if the public maximized (Lupia, 2016). This is discussed in more detail in the section on the normative implications of satisficing.

While the notion that turnout in elections increases when the costs of voting are reduced and decreases when the costs of voting are increased is widely accepted, a strict rational choice account of turnout famously implies that in most circumstances it is irrational to vote (Green & Shapiro, 1994). Although there have been many revisions to that strict model that address such dismal implications, a satisficing account based not on a voter’s estimate of the likelihood of casting the pivotal vote but on the idea of aspiration levels, trial and error, and sticking with habits whose outcomes are satisfactory while avoiding habits and behavior that are not proves to be highly predictive of turnout over time in several democracies (Bendor, Diermeier, & Ting, 2003; Collins, Kumar, & Bendor, 2009). This simply requires knowing whether an individual voted in the last election (whether he or she was a “voter” or a “shirker”) and whether or not the preferred party won (whether he or she was on the “winning” or “losing” side). The implication is that certain “satisfied voters” will sit out the next election, while other “dissatisfied shirkers” will be motivated to participate.

Is Satisficing a Stable Personality Trait?

In Simon’s description of satisficing it is a decision strategy influenced by cognitive constraints: while there is individual-level variation in cognitive capacity and the motivation to consider and weigh alternatives, satisficing is a “universal behavioral tendency” (Misuraca, Faraci, Gangemi, Carmeci, & Miceli, 2015, p. 111) affected by aspects of the decision such as its subject matter and difficulty. Thus Simon made no suggestion that satisficing is a stable personality trait. Nevertheless, psychologists became interested in the question of whether some individuals employ maximizing strategies all or most of the time as a function of their personality, while others adopt satisficing strategies all or most of the time when they make decisions. This led Schwartz et al. (2002) to develop the “maximization scale” based on agreement or disagreement with 13 statements such as, “When I watch TV, I channel surf, often scanning through the available options even while attempting to watch one program,” “I treat relationships like clothing: I expect to try a lot on before I get the perfect fit,” and “No matter what I do, I have the highest standards for myself.” This results in a unidimensional continuum with end points of (extreme) maximizer and satisficer.

Much of the debate since has centered on the construct and content validity of the maximization scale, prompting numerous refinements to the items as well as theories that there are in fact two or three dimensions to maximization (for a summary of measures, see Misuraca & Fasolo, 2018, Table 1). Maximization scales also exhibit relationships with some of the “Big Five” personality traits: conscientiousness, extraversion, and openness to experience are all associated with maximizing while neuroticism and agreeableness appear unrelated. For the purposes of this article, however, regardless of debates about scales, what is important is that the preponderance of this line of research indicates that satisficing is indeed a stable individual trait (see Harman, Weinhadt, & Gonzalez, 2018, for a contrary viewpoint): some individuals have a predisposition to satisfice and others to maximize when making decisions. The two main characteristics of a maximizer are seen to be the goal of making the best possible choice and the strategy of alternative search (Cheek & Schwartz, 2016), although there remains some disagreement about whether alternative search should be considered an intrinsic feature of maximizing. This personality trait manifests itself in empirical regularities such as how individuals make decisions affecting themselves versus the advice they give to others: maximizers maximize in decisions affecting self and offer the same advice to third parties, whereas satisficers advise third parties to maximize (Luan et al., 2018).

Is Satisficing “Better” Than Maximizing?

What does it mean to say that one decision rule is better than another? In many realms of politics and decision making such as voting, it implies that individuals decide or vote “correctly,” as they would if they had full information, more often. It is also preferable if individuals are more certain about, content with, and have a sense of efficacy about their decisions, all of which, at the mass level, are associated with engagement and participation in politics and thus with accountability. Before considering this question further, it is worth reiterating that satisficing need not imply laziness. Indeed, if an individual’s aspiration levels are sufficiently high, satisficing will necessitate careful consideration of alternatives until an alternative meets or exceeds the desired levels. There has been some debate over whether under certain conditions maximizers might accept lesser outcomes than satisficers. Some argue that satisficers can have higher standards than maximizers: using an example of house prices, Cheek and Schwartz (2016) contend that if a maximizer decides to accept the highest offer while a satisficer has an aspiration level that exceeds the best offer, the maximizer will accept the offer while the satisficer will take the house off the market. This is controversial, however, because it seems to violate Simon’s contrast between maximizing and satisficing and has resulted in a reassertion of seeking the highest standards as a necessary (and some argue sufficient; see later discussion) characteristic of maximizers.

Returning to the question of whether satisficing is in some senses a “better” decision rule than maximizing, we might think that in an ideal world people would be maximizers. In an election, for example, the perfect voter gathers information about the candidates or parties—their policies and their records—weighs up that information in the light of his or her own interests, and casts a vote for the candidate or party he or she feels is closest and most likely to be able to deliver what he or she wants. While it is well known that there are very few such maximizing voters in any electorate, as well as being an unrealistic expectation, should we be content or even want people to satisfice in elections? The evidence suggests that perhaps we should. For example, older citizens, who are usually seen as closer to the ideal citizen in that they engage with election issues and are most likely to vote, are also more likely to satisfice in forming their voter preferences (Riggle & Johnson, 1996). Lau and Redlawsk’s (2006) work also suggests that we should: If we follow the rule of thumb that we want individuals to vote their interests, satisficers appear to outperform maximizers as well as other types of decision makers. Indeed, Lau and Redlawsk’s findings indicate that “excess information” may actually impair decision making such that maximizers are less likely to make optimal choices.

On the other hand, how do we square this with the evidence that something as simple as ballot order can affect who wins an election? Ballot order effects imply that lower quality candidates could win office as an artifact of being the first, or high up, on the ballot, while higher quality candidates further down the ballot miss out. Clearly this has negative implications for the desirability of satisficing in elections. However, the evidence is that such order effects, while present, are small and most likely to affect elections where levels of information are particularly low and cues such as party and incumbency are lacking; indeed, Miller and Krosnick (1998) conclude that “they have probably done little to undermine the democratic process [in contemporary America]” (pp. 291–292).⁵

The notion that satisficing is a stable individual trait provides additional perspective on the question of its desirability. For if satisficing is a stable individual trait, it implies that an examination of such individuals to see whether they tend to be more confident about their decisions, happier with the results, and so on is possible. Regardless of how satisficing has been measured as a trait, most studies have associated it with positive outcomes of well-being and mental health such as greater life satisfaction, happiness, optimism, self-esteem, and less depression than maximizers (Chang et al., 2011; Iyengar, Wells, & Schwartz, 2006; Purvis, Howell, & Iyer, 2011; but see Diab, Gillespie, & Highhouse, 2008, for contradictory claims on some of these outcomes), although the evidence surrounding behavioral outcomes (such as securing a higher paying job) is more mixed (Iyengar et al., 2006).⁶ Thus not only is satisficing associated with voting correctly, but it also appears related to individual-level downstream effects of decision making that are desirable in the social and political world.

Satisficing in Surveys

Future Directions

There are two primary areas that are ripe for future research into satisficing—research in which satisficing is currently largely or wholly absent and topics on which satisficing is already prominent but in which research could be improved or extended—as well as the need for further research into the normative implications of satisficing. Beginning with topics for which satisficing is not currently prominent, while there has been a great deal of attention paid to satisficing in surveys, much less focus has been given to satisficing in experiments and seemingly none whatsoever to satisficing in qualitative research. Yet acknowledgment of the problem, how to identify it, its frequency, and possible remedies appear just as important in other research methods as in surveys. Qualitative methods such as in-depth interviews or focus groups can demand a great deal from respondents. While the objectives are different from surveys, interpretation of what is said is still very much dependent on the researchers’ recognition of issues in the production of the data. Satisficing is very likely in responses to questions, prompts, or group tasks that are often cognitively demanding during lengthy interviews or focus groups, but there is no equivalent literature on how to recognize or address satisficing in qualitative research—how one might, for example, distinguish satisficing from meaningful silence or acquiescence in a focus group—or its implications for the quality of the data that result.

With experiments, avoiding “observational studies” in service of greater internal validity frequently appears to be regarded as inherently sufficient for the quality of causal inferences. While problems of external validity are given lip service, the possibility of satisficing and its implications tends not to be (for an exception using conjoint analysis, see Bansak, Hainmueller, Hopkins, & Yamamoto, 2018). Yet experiments that involve, for example, fictitious candidates or scenarios from imaginary countries, let alone panels of Internet respondents who take numerous surveys for minimal rewards, are unlikely to engage respondents in maximal decision strategies. More research on how such features of experiments affect responses seems overdue (see Enns & Richman, 2013, p. 816, for a similar discussion).

Maximization scales in social psychology allow the development of a continuum of satisficing that differentiates between individuals who tend to engage in extensive satisficing in decision making and individuals who maximize. Individual-level measures of participants are therefore possible in all modes of research, including qualitative studies and experiments, and would allow insight into the effects of satisficing on response in research across the spectrum.

In research in general, also regardless of design, there is a need to distinguish between satisficing as “nuisance” and satisficing as explanation. The focus has been on satisficing compromising data quality, but there would seem to be a parallel here to framing effects. Just as framing effects from question wording in surveys can both impair data quality and be used to the researcher’s advantage by providing insights into the influences on political attitudes and behaviors, manipulating motivation, task difficulty, and so on could be used not only to enhance data quality in research but to provide greater insights into triggers of variation in satisficing decision making within and across individuals and its consequences. Even if satisficing is a reasonably stable personality trait, it implies that there is variation (i.e., there are times when satisficers will maximize). While political science has engaged in a great deal of research to find out when this will happen in surveys, it knows far less about this question in the real world of political decision making.

Finally, while Simon’s initial notion of an isolated, individual decision maker has been influential, it needs further development to account for the group contexts that are salient to so many decision tasks in politics. In elite decision making, for example, the relevance of the ways in which leadership groups are structured, questions such as whether or not reasoning and decision styles are mirrored by conversational partners, and how those affect the quality of decisions, are the kinds of questions political scientists have asked with respect to groupthink (Janis, 1982) and analogical reasoning (Khong, 1992) but not satisficing.

In the category of areas of political science in which satisficing is already prominent but where research could be improved or extended, clearly satisficing is a concept that is well known and used in several areas of elite and mass decision making. But satisficing began from recognition of the cognitive limitations of individuals; perhaps as a result, the role of affect as an influence on satisficing has been underexplored (it is noteworthy that Schwartz et al.’s, 2002, maximization scale has included statements such as, “I am uncomfortable making decisions before I know all my options” that suggest some affective component to satisficing). There are two clear stages of decision-making in which affect may influence satisficing: in signaling whether or not an impending decision is routine or worthy of more scrutiny and as information in deciding whether an alternative reaches or surpasses the desired aspiration level. With regard to the first stage, in later work Simon (1967) himself wrote about emotions as an “interruption mechanism” that alerts individuals to respond to urgent needs (i.e., jolts individuals out of habitual responses to stimuli, redirecting their attention). This is similar to the notion of affective intelligence and the role of emotions such as anxiety in signaling to the individual whether or not more elaborate scrutiny of the environment is necessary (Marcus, Neuman, & MacKuen, 2000). Affect may also play a part in the second stage of decision making: in evaluating alternatives individuals may reason backward from their emotions about an issue to where they stand on that issue; that is, affect will inform or interact with cognition in signaling whether or not an alternative satisfies aspiration levels. While these are not new insights into affect, its role in satisficing specifically is understudied but is important to our understanding of what triggers satisficing as a decision strategy and the alternatives that are then deemed good enough.

In areas where satisficing is already prominent, future research should do more to demonstrate that what we are observing is indeed satisficing. There are two related questions here: first, how strong is the evidence for satisficing, and, second, what are the competing accounts for what is taking place? Whether it be surveys or elite decision making, the researcher may observe behavior that is consistent with satisficing in that it appears to involve a less than full information search, the use of heuristics, and selection of an alternative that is good enough rather than optimal. In neither case, however, is it common practice to look at whether processes and outcomes that are consistent with satisficing actually seem to be satisficing upon deeper examination. Researchers often appear comfortable with the idea that if it “looks” like satisficing and “walks” like satisficing, it must be satisficing. This means that there could be slippage between when shortcuts are satisficing and when shortcuts are something else.

Satisficing involves the sequential consideration of alternatives, and that is not always apparent in decisions that are attributed to satisficing (i.e., satisficing will take place for a subset of those decisions that involve shortcuts). Indeed, a rare study that goes deeper by having respondents talk through answers to survey questions that are seen to signify satisficing, such as opting for the midpoint of a scale or acquiescence, suggests both that these patterns in answers may not derive from satisficing at all and that measures such as response latency as an indicator of satisficing could be misleading (Campanelli, Gray, Blake, & Hope, 2016). While some may regard self-reports of the considerations that informed a survey answer as unreliable, there are alternatives such as eye tracking that would allow additional perspective on a respondent’s focus prior to answering questions. Eye tracking has been used in marketing, for example, to look at how consumers consider brands and products (Stüttgen, Boatright, & Monroe, 2012) and would provide further clarity on questions like ballot order effects.

Future research should also examine more rigorously competing explanations to satisficing. This echoes Bendor’s (2003) point that “if we inspect the literature in political science for work that either tests his [Simon’s] ideas empirically or develops them theoretically, the pickings are lean” (p. 445). Too much research on satisficing ignores common practice in the social sciences by failing either to articulate what would falsify the satisficing hypothesis or to pit satisficing against competing possibilities such as social desirability bias (Prior, 2009) or “fast and frugal” decision making (Lau and Redlawsk, 2006). While little research sees the need to go further, when it does, it suggests that reality is often more complex. For example, Riggle and Johnson (1996) clearly pit the noncompensatory decision strategies of elimination by aspect and satisficing against each other by postulating that evidence for the former is a search of all candidates on a particular issue, followed by a search of the remaining candidates on a second issue, and so on, while satisficing involves searching for information about a candidate rather than an issue first, with that candidate either proving satisfactory (i.e., reaching the aspiration level) or prompting a search for information about a second candidate until one candidate proves satisfactory. Yan and Keusch (2015) find patterns of survey responses that could be attributed to satisficing, but in examining the possibility that response patterns could also be due to anchoring and adjustment they find more evidence for the latter than for satisficing. Regardless of whether or not one agrees with these authors’ distinctions, they make it clear what is and is not evidence for satisficing. To be able to test competing explanations, however, there also needs to be more agreement on the empirical manifestations of satisficing behavior. Currently, while there is some consensus that maximizers will have higher standards or aspiration levels, there is disagreement over whether this distinction alone is sufficient or whether the tendencies to search more or less alternatives and to find decision making difficult are also integral to maximizing as opposed merely to being the outcomes of having higher standards.

Some of the testable implications of satisficing in surveys also suggest more complexity in general than is often allowed. If satisficing in surveys, for which the indicators of satisficing are established and clear, is far more complex than acknowledged, it would seem even more likely in areas where the evidence for satisficing is opaque and indirect. For example, questions that are challenging and difficult to answer seem to result in less satisficing than questions that are familiar and have scales that are easier to understand, because their difficulty arouses the attention of otherwise quiescent respondents (Malholtra, 2009). “Heaping” of answers in surveys appears to reflect satisficing for some kinds of questions but thoughtful engagement for others (Holbrook et al., 2014). Moving away from surveys and returning to ballot order effects, some evidence suggests that candidates at the top and at the bottom of the list do better than they should based on ballot position, while candidates in the middle do worse. Thus straightforward satisficing on the part of voters, in which parties or candidates are considered in order until an alternative reaches the voter’s aspiration level, appears not to provide a full explanation; it may still be that voters are satisficing, but if so, it is in a way that is more complex than a sequential consideration of candidates. This needs additional theorizing and testing (Meredith & Salant, 2013).

The normative dimension to satisficing also needs more considered research. To what extent is satisficing a desirable decision making rule, in which elites and masses are more likely to make correct decisions with certainty and with which they are content? This is a critical question, both for elite and mass decision making. With respect to political knowledge, for example, a wealth of research beginning with Converse (1964) has demonstrated that ordinary individuals appear to have very shallow knowledge of public officials and the political system, as well as harboring inconsistent and unstable policy preferences. If, however, satisficing is the most effective decision rule by which individuals can “vote correctly,” perhaps we are asking too much unnecessarily and should be looking at political knowledge differently (Lupia, 2016).

This seems a particularly urgent question in the Internet age. Just as changes in survey designs can inhibit satisficing, the Internet has rendered it far easier to make an informed comparison of alternatives in a variety of domains from relationships (e.g., Tinder) to products (e.g., Consumer Reports). Thus maximizing may be an easier decision making rule in the Internet age and more prevalent (Simonson & Rosen, 2014). To take just one example from politics, Vote Advice Applications, designed to help prospective voters choose the candidate or party that most closely reflects their policy preferences, have become commonplace around the world, with the numbers of visitors to these sites during some countries’ elections numbering in the millions (see special issue of the Journal of Elections, Public Opinion and Parties, 2017, volume 1). This may seem like a boon because it provides individuals with a fuller picture of the electoral map and offers access to additional information. But perhaps it leads individuals toward maximizing rather than satisficing and a less desirable decision rule (Lau & Redlawsk, 2006). Drawing on another of Simon’s interests—artificial intelligence (e.g., Newell & Simon, 1956, 1972)—simulations using agent-based models (e.g., Fowler & Laver, 2008) are a further promising route to gaining insights into questions of the desirability of different decision rules, how agents may respond to failed decision rules, and some of the issues about group dynamics referred to earlier.

Finally, a more ambitious agenda for future research goes beyond satisficing to Simon’s broader insights about the implications of “bounded rationality” as an alternative grand theory to rational choice. In 2003 Bendor argued that

Too much of the potential in Simon’s work has gone untapped. We have often waved a magic wand, labeled either “bounded rationality” or “satisficing,” but such wands by themselves do little scientific work. It is, however, possible to build an alternative research program. (p. 434)

Some of the more recent work on poliheuristics, political knowledge, and voting behavior has the potential to build this alternative, in combination with theories such as affective intelligence (Marcus et al., 2000) that suggest a central role for emotions in determining when individuals will rely on shortcuts such as satisficing and when they will conduct fuller information searches, and with what consequences.