The past of predicting the future: A review of the multidisciplinary history of affective forecasting

The beginnings of modern affective forecasting research

The majority of the current affective forecasting work has come out of the laboratories of Daniel Gilbert and Timothy Wilson, as well as the work of their colleagues. Gilbert describes his original interest in affective forecasting errors as stemming from a realization that he felt better than he had imagined he would after going through some negative events in his life. In fact, he recalls that he started thinking about the topic while going through a divorce in 1992 (Gertner, 2003). While affective forecasting research stems from decision-making research within economics, Gilbert’s personal experiences led him to think about how money was not the only factor that played into decisions – the other, perhaps more important, factor was happiness (ibid.). A New York Times article from 2003 quotes Gilbert as summing up his new perspective on the research by saying, ‘It all seems so small. It isn’t really about money; it’s about happiness. Isn’t that what everybody wants to know when making a decision?’ (ibid.).

The 1990s offered fertile ground for studying happiness. Referred to colloquially as ‘the best decade’ (see, for example, Wilhelm, 2018) and marked by the collapse of the Soviet Union, along with a booming economy in the United States and relative peace, the 1990s were a prosperous time in America (ibid.). However, a brief overview of the General Social Survey, a survey conducted consistently from the mid 20th century to the present, makes evident that the number of individuals who reported being ‘very happy’ did not vary enormously between the 1980s and the 1990s (‘General Happiness’, n.d.). It is within this context that researchers began to wonder: Why was happiness, particularly at the higher end of the scale, so stable despite differences in social context? Could individuals accurately determine what would affect their happiness?

Now the author of Stumbling on Happiness (Gilbert, 2006), a popular psychology book that aims to explain affective forecasting errors to the general public, Gilbert noticed a lack of research on the topic and conducted the first studies of affective forecasting errors in the late 1990s (Jaffe, 2007). Gilbert, Wilson, and their colleagues have conducted the majority of the current work on affective forecasting. The first study of affective forecasting errors, conducted in 1998, compared assistant professors who forecasted their reactions to receiving tenure to professors who did receive tenure. While the forecasters’ long-term estimates of their happiness were relatively accurate, those who had received tenure were less elated in the short term than the forecasters believed they would be (Gilbert et al., 1998). Many more studies, demonstrating that individuals made mistakes in predicting the duration and intensity of their future affect, soon followed.

At the dawn of the 21st century, attention has turned to determining why affective forecasting errors might occur. Most current theories have used a dual-processing perspective based in the work of Daniel Kahneman and Amos Tversky. Dual-processing theorists claim that decision-making is ruled by a fast, automatic, and implicit cognitive system as well as a slower explicit system. In explaining why forecasting errors might occur, Wilson, Meyers, and Gilbert (2003) theorize that the valence of affective reactions is encoded implicitly but that explicit memories are often used as guides in decision-making. They hypothesize that cognitive biases (see, for instance, Kahneman et al., 1993) infect explicit memories of emotions and affect recall of past events and predictions of future affect (Wilson, Meyers, and Gilbert, 2003). Another interesting perspective on the roles of dual systems in affective forecasting is posited by Gilbert, Gill, and Wilson (2002), whose paper notes that in predicting their reactions to future events, individuals use ‘proxies’, imagining the event and their affective reaction to it. Such mental images may be tagged with negative or positive affect, which may act as a cue when making decisions (Finucane et al., 2000). However, yet again, Gilbert, Gill, and Wilson (2002) note that proxies can be ‘contaminated’ by current affect. For example, an individual who feels upset may find it difficult to accurately imagine their future level of happiness after a positive event.

While the theories of affective forecasting mechanisms are often presented as stand-alone ideas, they are based on earlier work in economics and philosophy. Articles on affective forecasting errors and their mechanisms are possible to understand on the surface without knowing the history of the field. Nevertheless, as noted above, affective forecasting theories provide an example of the importance of multidisciplinary work in developing innovative psychological theories. The next sections will review how research on utility, starting in the 18th century, provided a basis for later affective forecasting research.

Bernoulli and the early beginnings of affective forecasting

The history of behavioural economics, and thus the history of affective forecasting, begins in the 18th century with Daniel Bernoulli. Bernoulli’s father had wanted him to study business. Reluctant to do so, Bernoulli studied medicine, eventually finding his way to a career in mathematics through his entry into a competition for the Paris Prize (Tent, 2009). In some ways, Bernoulli’s theories fitted perfectly into the context in which he grew up, mirroring ideas that were prominent at that point in history. His 1738 paper ‘Exposition of a New Theory on the Measurement of Risk’ (Bernoulli, 1954[1738]) was published during the High Enlightenment years, when rationality ruled the roost (Duignan, 2019; White, 2018). In an era where everything was subject to empirical analysis, Bernoulli’s paper included mathematical equations to summarize the previously mysterious process of decision-making. However, Bernoulli’s paper, along with including logical mathematical equations, also highlighted the importance of subjective concepts, such as utility, in the process of decision-making. The consideration of psychological factors in the process of decision-making (Kahneman, 2011) was unusual in a paper published many years prior to the birth of Romanticism, when emotions and sensations would be considered in earnest (Duignan, 2019; Kehoe, n.d.).

Bernoulli’s writings focused on the role of utility in monetary decision-making. Moreover, this article will explain how economists’ research on utility applied to affect and emotion. Therefore, prior to discussing the role of economics in the development of affective forecasting, it is essential to define the terms utility, affect, and emotion. While Bernoulli was not the first economist to make use of the term, his use of the term within the context of the St Petersburg Paradox, described below, was important to the later development of affective forecasting theory. Furthermore, the definition of utility would change as the years went on, but the concept itself remained fundamental to why a field like economics played an important role in affective forecasting. Bernoulli defined utility as ‘moral expectation’, though today it is more commonly defined as pleasure, satisfaction, or happiness (Galiani, cited in Gilbert, 2006), notably all types of affect. Utility was framed as ‘moral expectation’ in order to contrast its subjective nature with objective mathematical expectation (Zabell, 1990). Furthermore, affect and emotion are both oft-used terms within the psychological literature and are often used interchangeably. For example, the definition of affect in Wetherell (2012) is ‘the emotion associated with an idea or set of ideas’ (Collins English Dictionary definition, quoted in ibid.: 1). Meanwhile, the American Psychological Association refers to emotion as ‘a complex reaction’ (‘Emotion’, n.d.). In this article, affect will be the primary term used. The term emotion will be used colloquially when referring to specific types of affect (such as happiness).

The most important contribution of Bernoulli’s paper to affective forecasting was in its final pages. Bernoulli described a gambling vignette dubbed the ‘St Petersburg Paradox’. Though this was complex, only the essence of the game and its paradox is relevant within the context of this article. Bernoulli noted that there was a difference between what an individual really would pay to participate in this game and what a rational individual should pay, given the laws of mathematics. He noted that most reasonable people would pay only 20 ducats to participate in the gamble, though they should be willing to pay an infinite amount, had they made their decision based on mathematical expectations. Nevertheless, Bernoulli stated that it was not just the expected monetary value of the gamble (i.e. how much they could expect to earn by participating in it) but also its utility that affected how much money an individual would pay to participate (Clark, 2002; Kahneman, 2011). Therefore, Bernoulli demonstrated that individuals made choices based on subjective factors, such as utility, and on monetary factors (Kahneman, 2011). Bernoulli’s definition of utility as ‘moral expectation’ is especially important here, as ‘expectation’ implies the anticipation of something occurring in the future (‘Expectation’, 2019). However, in the years between the publication of Bernoulli’s paper and later work in behavioural economics, several contributions and detours would occur.

Bentham, James, and detours to economics and psychology

While the roots of affective forecasting lie in economics and psychology, the field’s origins took a brief detour into philosophy through the work of Jeremy Bentham in the late 1700s. A British philosopher, Bentham had studied law, though he had quickly turned from ‘the practice of law as it was to the study of law as it ought to be’ (Steintrager, 2004: xi). Little information is available about how Bentham’s historical context influenced the development of his theory. However, Bentham published his works towards the end of the Enlightenment era and was particularly influenced by the concept of empiricism. As discussed by Crimmins (2019), Bentham’s belief in empiricism led him to theorize that sensations such as pain and pleasure could be quantified, thereby influencing his definition of utility. His main theory, utilitarianism, was that individuals were ruled by pain and pleasure and that the proportion of pain to pleasure played a central role in decision-making. If a decision maximized pleasure and minimized pain, then it conformed to the principles of utilitarianism (Bentham, 2000[1781]). ¹ Moreover, Bentham noted that pain and pleasure varied in their intensity and duration, among several other aspects (ibid.). Therefore, early philosophers subscribing to utilitarianism were conceptualizing affect and its measurement in the same terms as modern researchers who measure errors in forecasting the intensity and duration of affect (Wilson and Gilbert, 2005; Wilson, Meyers, and Gilbert, 2003).

While Bentham brought the role of the emotions of pain and pleasure into decision-making prior to the majority of economists, his thoughts on the matter differed in several important ways from how we view affective forecasting and decision-making today. The most important difference is that Bentham viewed people who did not subscribe to utilitarianism as dealing in ‘caprice instead of reason’ (Bentham, 2000[1781]: 14). Therefore, Bentham never took into account the idea that decisions may be capricious, even while simultaneously subscribing to utilitarianism, simply because individuals are often incorrect in predicting what will bring them pleasure and what will bring them pain. For example, a professor might work to achieve a tenured position, predicting that this will maximize their pleasure and minimize their pain, and thus following the principles of utilitarianism. However, as professors in Gilbert et al.’s 1998 study found out, the pleasure that they derive from tenure might last for a significantly shorter time than they expected. It is conceivable that had the professors been able to more accurately predict the pleasure that they would feel from tenure, they might have made a different decision. Nevertheless, their original decision was ‘reasonable’, at least according to Bentham.

A century after Bentham had published his views on utilitarianism, William James, a pre-eminent psychologist at Harvard, released his essay ‘What Is an Emotion?’ in 1884. James was likely the first psychologist who influenced the development of affective forecasting theory, over 100 years before the theory was formalized. While some of James’ views contrasted with our current understanding of cognition and affect within affective forecasting, his understanding of human nature made it possible for affective forecasting theory to eventually develop. James, unlike his predecessors, did not consider humans to be fully rational. As one of the founders of the functionalist movement, he was more interested in discovering how people truly behaved and adapted to the world (Schultz and Schultz, 2007).

This contrast between an idealized and rational view of behaviour and an understanding of actual human behaviour is familiar, a preview of the later debates between Richard Thaler and other behavioural economists, who viewed individuals as sometimes irrational, and traditional economists, who viewed them as rational ‘Econs’ (Thaler, 2016). James noted that emotions were not necessarily logical, rational feelings. In fact, he stated that they might oppose ‘the verdict of our deliberate reason’ (James, 1884: 190). Such claims are fitting to James’ personal history. Originally planning to become an artist, James studied art in Newport, Connecticut. Later, deciding that he lacked the necessary talent for art, James studied chemistry at Harvard (Hunt, 1993). It is possible that his foundational education in the arts allowed James to discuss emotions as being highly subjective. Moreover, the Romantic era dawned in Europe in the 1800s, allowing for a rejection of Enlightenment rationality and the consideration of emotion (Kehoe, n.d.). Notably, Lacasse (2017) points out the connection between James’ view of emotions as irrational and later theories, such as Kahneman’s (2011) dual processing and Slovic’s affect heuristic (Slovic et al., 2007). Had James tried to push the idea of emotions as being subject to rational rules, the area of research that we now refer to as affective forecasting, which posits that individuals make irrational errors when predicting their own affect, might have had trouble gaining a foothold.

Early 20th-century economists’ contributions to affective forecasting

William James was the only psychologist prior to the end of the 20th century whose work went on to obviously influence the later development of affective forecasting theory. Instead, economic theories continued to develop throughout the early and mid 1900s, and several of these developments provided the basis for our understanding of affective forecasting and its importance. Twentieth-century economists focused on discerning how individuals made decisions regarding purchases. The idea discussed by economists at the time was that individuals chose the item that would provide them with the greatest utility, a word that had peaked in its usage in approximately 1796, several years after the publication of Bentham’s book. ² Theoretically, in order to choose the item that would provide them with the greatest utility, individuals had to be able to predict how their future happiness would be impacted by the purchase of that item (see Brockman, 2013). As Daniel Gilbert notes, ‘Wise choices are those that maximize our pleasure, not our dollars, and if we are to have any hope of choosing wisely, then we must correctly anticipate how much pleasure those dollars will buy us’ (Gilbert, 2006: 257). Therefore, while economists in the 20th century never overtly discussed affect, their assumptions were based on the idea that individuals had to be able to predict their future affect in order to make financial decisions.

In the early 20th century, John von Neumann was born in Budapest, Hungary. Although he showed talent in mathematics, von Neumann went on to study chemistry in Zurich; eventually, however, he earned his doctorate in mathematics in Budapest. While von Neumann made significant contributions to technology and mathematics (Poundstone, 2011), his main work of interest to this article is the development of expected utility theory, which he published in conjunction with Oskar Morgenstern in 1944. World War II made an important contribution to the historical context in which von Neumann and Morgenstern developed their theory. As described by Israel and Gasca (2009), interdisciplinary research groups of both natural and social scientists were formed to contribute to the war effort. Economists were especially influenced by the methodologies of the natural sciences. The formulas and functions that von Neumann and Morgenstern used to describe human behaviour were certainly influenced by this objective.

Expected utility theory, as described by von Neumann and Morgenstern in The Theory of Games and Economic Behavior (1944), is complex. However, two aspects of the von Neumann–Morgenstern theory are relevant to this article. First, von Neumann and Morgenstern’s expected utility theory included the concept of risk and was based on Bernoulli’s theories (Levin, 2006; Moscati, 2018). Second, von Neumann and Morgenstern’s theorem and explanation were almost strictly mathematical. For example, in attempting to define rational behaviour, the authors wrote that

a study of all these questions in qualitative will not exhaust them, because they imply, as must be evident, quantitative relationships. It would, therefore, be necessary to formulate them in quantitative terms so that all the elements of the qualitative description are taken into consideration. (von Neumann and Morgenstern, 1944: 9)

Expected utility theory had a large impact both on economics and on affective forecasting research. However, it is here that the second aspect of von Neumann and Morgenstern’s work must be considered: the mathematics of it all. It was at this point that economics hit a roadblock in what it could contribute to affective forecasting theory. Economists, up to this point, had failed to discuss decision-making errors. Given that mis-predicting the future is a fundamental part of affective forecasting, decision-making errors were becoming impossible to ignore. Unfortunately, multidisciplinary work within economics is relatively rare. Becher (1989) notes that while economics shares borders with mathematics and political science, it has less common ground with disciplines such as psychology and law. Therefore, there was little overt collaboration between economists and psychologists to write the earliest works on affective forecasting. However, in the 1970s, psychologists began borrowing the language and some of the concepts of expected utility theory to lay the foundations for what would soon become affective forecasting theory.

The entrance of Kahneman and Tversky, and a paradigm shift

Our current understanding of affective forecasting would have been impossible had it not been for the foundations developed by economists, philosophers, and psychologists in the 18th to 20th centuries. However, economics underwent a paradigm shift in the late 1960s that made it possible for Gilbert and Wilson to later come to their conclusions regarding affective forecasts. In the 1960s, Amos Tversky and Daniel Kahneman began working together in a collaboration that would last for almost 20 years, until Tversky’s death in 1996. Cass Sunstein and Richard Thaler, writing in the New Yorker, describe Kahneman as a pessimist and a worrier and Tversky as an eternal optimist (Sunstein and Thaler, 2016). Together, they wrote the papers that contributed to the development of affective forecasting.

To truly understand Daniel Kahneman’s work, it is essential to understand the context in which he grew up. During World War II, Kahneman was a Jewish adolescent, relocating constantly along with his parents and sister to escape capture by the Nazis. According to Michael Lewis, whose 2016 book The Undoing Project adeptly places Kahneman and Tversky in their historical context, Kahneman’s experiences during the war taught him that humans were ‘endlessly complicated and interesting’ (Lewis, 2016: 53). An example of humans’ complicated reasoning patterns occurred in 1941, when Kahneman’s father was jailed in Drancy. Though most Jews jailed in Drancy were sent to concentration camps, Kahneman’s father was released six weeks after his imprisonment thanks to the help of his boss, Eugène Schueller. Much later, it was discovered that Schueller had collaborated with Helmut Knochen, a commander of the SS (Sancton, 2017). However, Schueller seemed to have exempted Kahneman’s father, a successful chemist at L’Oréal, from the anti-Semitism that governed how he perceived other Jews (Lewis, 2016). From this experience, among many others, Kahneman began to learn that humans were inconsistent and often irrational. Though Kahneman would go on to win the Nobel Prize in economics, his theoretical underpinnings, along with his early work and education, were firmly based in psychology, and he therefore closely integrated psychological concepts into his economics research.

Kahneman and Tversky’s seminal 1979 paper ‘Prospect Theory: An Analysis of Decision Under Risk’ stated that most individuals, if rational, followed expected utility theory when making decisions. In essence, individuals should (and did) choose the option that would result in the greatest utility when making decisions. However, the authors noted, in a theme that became common throughout most of their work, that individuals were prone to violating expected utility theory. In fact, they were subject to cognitive biases such as the certainty effect, in which they overweighted certain outcomes over probable ones. These biases resulted in what most economists would consider irrational and sometimes faulty decision-making (Kahneman and Tversky, 1979). For example, Kahneman claimed that when individuals made forecasts of happiness, they would often use their immediate reaction to a particular situation to forecast its long-term effects on their happiness, a ‘generalization of the analysis offered in prospect theory’ (Kahneman, 2000: 705). Later on, Kahneman and Snell would further explain their view of decision-making in a 1992 paper, claiming that individuals were poor forecasters of their own predicted utility (i.e. they were unable to determine what would result in the greatest amount of utility for them). In a study of this concept, Kahneman and Snell (1992) found that participants predicted that their liking of a type of yogurt would decline if they ate it every day for a week straight. In fact, their liking for the yogurt increased, demonstrating that their predicted utility was inaccurate.

In later papers, Kahneman and Tversky formed a theory and a set of terminology to explain cognitive biases and why they occurred. Importantly, they produced a dual-processing account, positing that there were two cognitive systems, often referred to as System 1 and System 2. System 1 was fast and automatic, and worked mainly below our level of awareness, while System 2 was rule-based and logical, and worked mainly at the level of awareness (Smith and DeCoster, 2000). Crucially, the speed of System 1 resulted in efficiency but not necessarily accuracy. Therefore, thinking with System 1 could lead to cognitive errors and biases, such as the certainty effect, the peak-end rule, and the focusing illusion, among many others (see Tversky and Kahneman, 1974).

In 1998, Schkade and Kahneman published one of the earliest papers that discussed the ideas of affective forecasting, albeit with different terminology. Building off one of the most famous papers within the psychological literature (Brickman, Coates, and Janoff-Bulman, 1978), Schkade and Kahneman asked undergraduate students in California and Michigan how satisfied they were with various aspects of their life and how satisfied they were overall, as well as asking a second group in both states how satisfied someone with the same values and interests in a different state (i.e. in California, if they were a Michigan student) would be. Counterintuitively, they found that while Californian students were more satisfied than Michigan students with specific aspects of their life, overall satisfaction did not differ between the two regions. However, respondents in both regions predicted that overall life satisfaction would be higher for Californians. In essence, the paper found that individuals were inaccurate forecasters of other people’s happiness (Schkade and Kahneman, 1998).

Schkade and Kahneman chalked up the results of their study to the focusing illusion, stating that individuals made errors in assigning proper weights to considerations that were the focus of their attention. For example, when individuals made forecasts about how good winners of the lottery might feel, they might focus on the change in their state from a regular person to an extremely wealthy one; however, they failed to focus on regular aspects of that lottery winner’s life that might also impact their future happiness. The term focalism, the tendency to focus excessively on the event in question and ignore other events that may also influence your affect, would later be implemented in much of the affective forecasting literature as a potential explanation for forecasting errors (Wilson et al., 2000). For example, Gilbert et al. (1998) explained that a professor who won tenure might be elated, but eventually his attention would turn to the various other tasks that needed to be completed, such as writing a syllabus or completing a book chapter. However, an assistant professor who made a forecast of how they would feel when achieving tenure would focus on the achievement, and not on the extra work that the achievement might bring.

While Kahneman and Tversky worked on their research first in Israel, then in America, research on decision-making errors has also been developed in Europe. In Germany, Ralph Hertwig researches how individuals make decisions under conditions of uncertainty, while Gerd Gigerenzer’s work takes a particularly optimistic view of heuristics. While European psychologists rarely discuss affective forecasting per se, they have developed a rich tradition of research on how individuals employ heuristics to make optimal decisions. For example, in Taming Uncertainty (2019), Hertwig and his co-authors argue that expected utility theory and subjective utility theory are insufficient to explain how individuals make decisions. They posit that while past work has assumed that individuals under uncertainty calculate statistical probabilities of events that will occur in the future, the story is in fact more complicated. Human cognitive systems are sophisticated, and we can employ a wide range of strategies to make decisions. Hertwig’s research takes an optimistic view, discussing the myriad strategies that individuals use to make optimal decisions. Similarly, Gigerenzer’s research focuses on smart heuristics. In contrast to Kahneman and Tversky, Gigerenzer conceptualizes heuristics as tools that allow individuals to make decisions efficiently (Gigerenzer and Selten, 2001). For example, when asked to guess which of two cities has the largest population, individuals with less information tend to guess more accurately than individuals with more information, relying on the recognition heuristic. Employing this heuristic is efficient and effective – there is a correlation between the name recognition of a city and its population size (Goldstein and Gigerenzer, 2002). Though Gilbert and Wilson’s work is grounded in psychology, Hertwig and Gigerenzer have both worked with a multidisciplinary team, providing further evidence of the nature of decision-making research.

While many other researchers have contributed to the development of affective forecasting, it is Kahneman and Tversky’s work that was foundation of the emerging field. The authors’ main conclusions were that (a) humans were irrational and sometimes made faulty decisions; and (b) the majority of faulty decision-making was a result of cognitive biases and heuristics. Gilbert and Wilson applied many of Kahneman’s ideas to their explanations of the mechanisms of such errors, noting, for example, that cognitive biases might influence explicit memories of past events, thus influencing predictions of future affect. While few of Kahneman and Tversky’s papers discussed affective forecasting per se, Gilbert and Wilson went on to apply its general elements in their work. Kahneman and Tversky, by discovering cognitive biases, were among the first to challenge economists’ claims that individuals made rational decisions. Gilbert and Wilson applied their knowledge of cognitive biases to their own work, noting that such biases came into play when individuals predicted their future affect (Wilson, Meyers, and Gilbert, 2003). Meanwhile, Schkade and Kahneman’s 1998 paper was a precursor of affective forecasting work. Its findings, that people were inaccurate in predicting how happy they would be in a different situation, suggested that individuals were inaccurate in forecasting future affect. Moreover, Gilbert and Wilson went on to attribute errors in affective forecasting to biases similar to the focusing illusion. Just as Schkade and Kahneman claimed that what you focused on affected your judgements, Gilbert and Wilson have since noted that when making predictions of the future, individuals tend to focus on their current affect and use it to make forecasts (Gilbert, Gill, and Wilson, 2002). As well as this, Kahneman and colleagues’ work often focused explicitly on individuals’ liking of various things, be it California, Michigan, or yogurt. Therefore, while their work fitted within behavioural economics, it also incorporated the study of individuals’ predictions of whether various choices would make them feel positively or negatively.

Most importantly, Kahneman and Tversky, along with several other pre-eminent economists, founded behavioural economics, which blended classical economics with concepts and theories gleaned from psychology. The development of behavioural economics as an overtly interdisciplinary field of research set the stage for psychological theories to begin incorporating concepts that were originally based in economics research. The 1998 publication of Gilbert and colleagues’ paper on professors’ forecasts of their affect regarding receiving tenure signalled the official beginning of affective forecasting research. Currently, articles discussing affective forecasting focus on its mechanisms and rarely do so without making use of the terminology and theories of behavioural economics.

The future of predicting the future

While affective forecasting has a long history of precursors and building blocks that led to its development, the field also has a long way to go and continues to change. Early affective forecasting work focused on determining the reasons behind forecasting errors; more recent work has turned to the clinical implications of making forecasts and errors in forecasts. This development is to be expected. Psychology as a whole has steadily moved towards a heavier focus on applicable research and away from basic cognitive science. The focus on applicability in fact has its origins in the early 1900s, when funding in psychology began to be determined by the ability of one’s research to ‘cure society’s ills’ (Schultz and Schultz, 2007: 225). This trend continued in the decade following World War II, when the foundation that is currently called the Department of Veterans Affairs in the United States led psychologists to discover that clinical work could earn them a living (Seligman and Csikszentmihalyi, 2000). Currently, the movement towards practical work continues across all the sciences. Researchers in Canada complain of low fundamental science funding, with grants funding this sort of research down from 50.1% of the National Sciences and Engineering Council's budget to 38.4% at the start of the 2010s (Birchard and Lewington, 2014).

Given this change in focus, affective forecasting research in 2015 and beyond has increasingly focused on applications in the understanding and treatment of depression and anxiety, as well as symptoms of other psychological disorders. In a 2019 study, Zetsche and colleagues find that individuals with major depressive disorder are less accurate than healthy individuals in forecasting their future mood. Most importantly, the authors urge that treatment of major depressive disorder take biases in forecasting into account. Clinicians should work with clients to correct distorted forecasts, perhaps even by making use of technology (such as smartphone applications that can assess daily affect; Zetsche, Bürkner, and Renneberg, 2019). In clinical work on trauma, Rizeq and McCann (2019) have found that trauma experience leads individuals to forecast greater negative affect related to life events, and that this relationship is mediated by emotional dysregulation, another concept that is particularly clinically relevant. Other authors have found that a lack of forecasting errors may be one of the keys to certain disorders. For example, Moore et al. (2019) have determined that individuals with social anhedonia, a symptom of schizophrenia and other disorders, are more accurate in forecasting negative affect. Such individuals expect to receive little pleasure from social interactions and do not in fact experience much pleasure during such interactions, potentially leading to increased desire for solitude.

While a clinical focus has pushed affective forecasting in a new direction, researchers within clinical psychology have built upon existing research on cognitive biases within psychological science, particularly the research of Gerd Gigerenzer. Research into smart heuristics has looked at cognitive biases from a positive angle – heuristics can lead to mistakes, but they are also essential for efficient decision-making. Similarly, cognitive biases can lead to errors, but errors and biases are also adaptive. In a recent article, Zetsche and colleagues find that individuals with obsessive-compulsive disorder (OCD) do not have an ‘unrealistic optimism bias’ when predicting the likelihood of various events (Zetsche, Rief, and Exner, 2015: 510). That is, such individuals lack the (sometimes inaccurate) tendency that healthy individuals have to overestimate the probability of good things happening to them. The lack of this cognitive bias plays a role in the symptomatology of OCD. Accordingly, Wenze, Gunthert, and German (2012) have found that individuals with anxiety make particularly inaccurate forecasts of future negative mood, in comparison to the average participant. While this branch of research is continuing to develop, it is evident that clinical psychologists have added a useful new perspective to past cognitive research findings.

Conclusion

Affective forecasting is a relatively modern field of research within psychology. However, its historical roots go as far back as the 18th century, to the birth of utilitarianism as a philosophy and Bernoulli’s St Petersburg paradox. Later, economics research in the 20th century led to our current understanding of how the considerations of future affect play a role in day-to-day decision-making. Most importantly, the classical economic theories of the 20th century and earlier allowed Kahneman and Tversky, among others, to react to what was missing from economics, eventually leading to the birth of behavioural economics as a research area (see Thaler, 2016). With the founding of behavioural economics, the role of affective errors in decision-making was better understood, and the way was paved for Gilbert and Wilson’s research. Current work on affective forecasting aims to apply the tenets of affective forecasting to clinical work. In order to develop this theory, it is imperative that we appreciate the roots of affective forecasting errors and the multidisciplinary nature of affective forecasting research.