Advertisement

Sunk Cost Effect in Individuals with Autism Spectrum Disorder

  • Junya Fujino
  • Shisei Tei
  • Takashi Itahashi
  • Yuta Aoki
  • Haruhisa Ohta
  • Chieko Kanai
  • Manabu Kubota
  • Ryu-ichiro Hashimoto
  • Motoaki Nakamura
  • Nobumasa Kato
  • Hidehiko Takahashi
Original Paper

Abstract

The sunk cost effect, an interesting and well-known decision bias, is pervasive in real life and has been studied in various disciplines. In this study, we modified a task exemplifying the sunk cost effect and used it to evaluate this behavior in individuals with autism spectrum disorder (ASD). The control group exhibited a typical sunk cost effect in our task. We found that the sunk cost effect was lower in the ASD group than in the control group. The results agree with previous evidence of reduced sensitivity to context stimuli in individuals with ASD and extend this finding to the context of the sunk cost effect. Our findings are useful in addressing the practical implications on their socioeconomic behavior.

Keywords

Autism spectrum disorder Decision making Sunk cost effect Behavioral economics 

Introduction

Decision making, which is the process resulting in the selection of a belief or a course of action among several alternative possibilities, is central to daily functioning (Fujino et al. 2016a; Mosner et al. 2017; Sharp et al. 2012; Thiébaut et al. 2016). Individuals with autism spectrum disorder (ASD), which is characterized by altered social interaction and atypical interests, frequently show altered decision making in various situations (Boulter et al. 2014; Brosnan et al. 2014; Luke et al. 2012; Minassian et al. 2007). This, in turn, can influence their economic and social functioning (Farmer et al. 2017; Mussey et al. 2015).

Studies on decision making and behavioral economics have rapidly expanded in recent years and have demonstrated that humans are susceptible to various types of biases in decision making (Camerer 2011; Mobbs and Kim 2015; Takahashi 2013; Thaler 1994; Tversky and Kahneman 1992). In line with their advancements, these disciplines have been used to assess altered decision-making behavior in individuals with ASD. For example, previous studies have shown that loss/gain framing effects were reduced when individuals with ASD made choices between gambles (De Martino et al. 2008; Shah et al. 2016). Farmer et al. (2017) have recently demonstrated that individuals with ASD were less influenced by decoy options in an attraction effect task. Our work has also shown that sensitivity to context changes under ambiguity was reduced in individuals with ASD (Fujino et al. 2017a). These previous findings are in agreement regarding the fact that individuals with ASD show reduced sensitivity to context stimuli and make more rational and consistent choices in experimental situations. Along this line, Brosnan et al. (2016, 2017) have proposed a Dual Process Theory account of ASD, that is, reasoning by individuals with ASD can be characterized by reduced intuitive (Type 1) and greater reflective (Type 2) processing. Briefly, Dual Process Theory, a major theory of reasoning and decision making within psychology, posits two different modes of processing: Type 1 is intuitive, unconscious, and experiential and Type 2 is reflective, conscious, and rational (Evans 2008; Lieberman 2007). This issue deserves further exploration because choice consistency is regarded normative in conventional economic theory, and reduced context sensitivity in ASD would demonstrate that ASD is not a disability in all respects (Baron-Cohen 2000; Farmer et al. 2017).

In this study, we investigated whether the reduced context sensitivity that characterizes ASD extends to decision making in the context of the sunk cost effect. The sunk cost effect is the tendency to continue an investment or take an action even though it has higher future costs than benefits, if costs of time, money, or effort were previously incurred (Arkes and Blumer 1985). A clear example of the sunk cost effect was proposed by Arkes and Blumer (1985): Suppose that “you have spent $100 on a ticket for a trip to Michigan. Several weeks later you buy a $50 ticket for a trip to Wisconsin. You think you will enjoy the Wisconsin trip more than the Michigan trip. When you are putting your just-purchased Wisconsin trip ticket in your wallet, you notice that the Michigan trip and the Wisconsin trip are for the same weekend! It is too late to refund either ticket. You must use one ticket and not the other. Which trip will you go on?”

According to an assumption of traditional economics theory, decisions should be made on the basis of the costs and benefits expected to arise in the future from the choice of each option (Thaler 1980). According to this assumption, everyone would be expected to choose the trip considered more enjoyable (i.e., the trip to Wisconsin). However, approximately half of the participants in the experiment chose the Michigan trip instead, which had incurred the larger sunk cost (Arkes and Blumer 1985). Considering that an invested sunk cost cannot be recovered, a rational forward-looking decision maker should ignore sunk costs (Arkes and Blumer 1985; Augenblick 2016; Friedman et al. 2007; Thaler 1980). However, nonhuman species (Magalhães et al. 2012), humans (Strough et al. 2008), companies (Tan and Yates 1995), and even governments (Sleesman et al. 2012) often do not ignore sunk costs during decision making. For example, one is likely to wear a piece of ugly clothing if it has cost him more money (Zeng et al. 2013). People who pay less at an all-you-can-eat pizza restaurant because of a surprise reduction in the price tend to eat a smaller amount of pizza (Just and Wansink 2011). The sunk cost effect is pervasive in real life and influences many types of socioeconomic behavior, such as entertainment (Roth et al. 2015), investments (Bogdanov et al. 2017), management (Conlon and Garland 1993), and interpersonal relations (Arkes and Blumer 1985). Therefore, understanding how individuals with ASD behave under sunk costs can potentially reveal new insights into the practical implications of their socioeconomic behavior. However, to the best of our knowledge, no study has investigated the sunk cost effect in individuals with ASD.

If the tendency of individuals with ASD to prioritize local information and to be relatively insensitive to context stimuli extends to decision making in the context of the sunk cost effect, they would be less influenced by sunk costs and would make more rational choices under such situations. Correspondingly, we hypothesized that the sunk cost effect would be reduced in individuals with ASD compared with those without ASD. In this study, we modified a task that clearly shows the sunk cost effect (Arkes and Blumer 1985; Fujino et al. 2016b) and used it to test our hypothesis.

Methods

Participants

Twenty-nine adults with ASD and 29 healthy control adults participated in this study. The sample size was determined on the basis of previous studies on the decision making of individuals with ASD (e.g., Luke et al. 2012; Robic et al. 2015; Shah et al. 2016). Participants with ASD were recruited from a database of volunteers who had received a clinical diagnosis of ASD in the outpatient units of Showa University Karasuyama Hospital. The diagnostic procedure to identify individuals with ASD was the same as in our previous studies (Fujino et al. 2017a; Itahashi et al. 2015). Briefly, at least three experienced psychiatrists and a clinical psychologist assessed all the patients using the criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition text revision (DSM-IV-TR). The assessment comprised participant interviews about developmental history, present illness, life history, and family history. Patients were also asked to bring suitable informants who had known them in early childhood. This process required approximately 3 h. A diagnosis of ASD was made only when there was a consensus between the psychiatrists and clinical psychologist. At the time of testing, an experienced psychiatrist evaluated psychiatric comorbidity by using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). No participants with ASD satisfied the diagnostic criteria for substance use disorder, bipolar disorder, or schizophrenia. The healthy controls were recruited from the general population in the Tokyo area via advertisements and acquaintances. They did not meet the criteria for any psychiatric disorders according to the evaluation of an experienced psychiatrist using SCID. The control and ASD groups were matched for age, gender, current smoking status, education, and estimated full-scale intelligence quotient (IQ) level; previous studies reported that IQ level could have an impact on decision making under sunk costs (Stanovich and Stanovich 2010). No participant (participant with ASD or control) had any history of head trauma, serious medical or surgical illness, or substance abuse.

The ASD symptom severity and the IQ levels of the participants with ASD had been evaluated before the study. With regard to the severity of ASD symptoms, the Autism Diagnostic Observation Schedule (ADOS) (Lord et al. 2000) was used. The ADOS is a research-standardized test that provides an index of autistic symptoms observed on a specific occasion. The IQ levels of the participants with ASD were evaluated using either the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III) or the Wechsler Adult Intelligence Scale-Revised (WAIS-R). Although there are several minor changes in WAIS-III from WAIS-R (e.g., more items), the number of core items remains largely unchanged. Therefore, we considered them to be essentially identical in the full-scale IQ measurement of individuals with ASD. Each participant with ASD was considered high functioning because his or her full-scale IQ score was above 80. The IQ scores of the controls were estimated using a Japanese version of the National Adult Reading Test (JART) on the basis of previous findings that JART successfully predicted full-scale IQ scores in healthy populations (Matsuoka and Kim 2006; Matsuoka et al. 2006).

This study was approved by the Committee on Medical Ethics of Kyoto University and the institutional review board of Showa University Karasuyama Hospital, and was conducted in accordance with The Code of Ethics of the World Medical Association. After providing a complete description of the study to all participants, written informed consent was obtained from all of them.

Sunk Cost Effect Task

We modified a clear example of the sunk cost effect described in the Introduction (Fig. 1). The task was similar to that used in our previous study (Fujino et al. 2016b). Participants were initially instructed to select their preferred option from two travel destinations (preference phase). Thereafter, participants received the following instructions: “You mistakenly purchased both of the tickets. You cannot get a refund, and the departure date is the same for both. Which one will you choose?” The task comprised 46 trials. In half of the trials (control condition), the costs of the two tickets were identical. Twenty-three different pairs of destinations in various countries (e.g., New York vs. Los Angeles, London vs. Manchester, Kamakura vs. Hakone) were presented. Each destination was presented only once (all 23 destination pairs are shown in Table S1). The costs of the tickets were realistically defined from among five values (¥30,000, ¥40,000, ¥50,000, ¥80,000, and ¥100,000). One hundred Japanese yen corresponds to approximately 1.0 US dollar. In the other half of the trials (sunk cost condition), the participants selected from the same destination pairs and ticket costs. However, tickets for the destinations not selected by participants in the preference phase, i.e., nonpreferred, were presented as 1.5 times more expensive than tickets for the preferred, i.e., chosen in the preference phase, destinations. A decision maker sufficiently influenced by the sunk costs would switch his/her choice in this condition. The trial order was randomized across participants. The time course is shown in Fig. 1.

Fig. 1

Experimental design. In each trial, subjects first express a preference for one of two cities to travel to (preference phase). They are then told that they have, by mistake, previously paid for nonrefundable tickets to both cities (which are sunk costs). In the control condition, the tickets are reasonably priced, and the costs of the two tickets are identical. In the sunk cost condition, the ticket price for the nonpreferred destination is 50% higher than the ticket price for the preferred destination. A decision maker sufficiently influenced by the sunk cost will switch his or her choice in this condition. This figure depicts the case in which New York is chosen in the preference phase

All participants were quizzed about how well they understood the task (Supplementary methods) and practiced at least once on a shorter version of the task. Only after they successfully completed the quiz were they allowed to proceed with the experiment. The experiment was conducted using E-Prime software (Psychology Software Tools, Inc., Pittsburgh, PA, USA).

Statistical Analyses

On the basis of our previous study (Fujino et al. 2016b), the strength of each participant’s sunk cost effect was estimated as follows:
$${\text{Sunk cost effect score}}=\left( \begin{gathered} \frac{{{\text{Number of trials where the option that would be enjoyed less was chosen under sunk cost condition}}}}{{{\text{Total number of trials in the sunk cost condition}}}} \hfill \\ - \frac{{{\text{Number of trials where the option that would be enjoyed less was chosen under control condition}}}}{{{\text{Total number of trials in the control condition}}}} \hfill \\ \end{gathered} \right) \times 100\,(\% )$$

A higher sunk cost effect score indicates a stronger sunk cost effect.

Furthermore, to further explore how the participants behaved when considering sunk costs, we separated the trials on the basis of ticket costs; previous studies have shown that healthy subjects are more susceptible to the sunk cost effect when larger sunk costs are incurred (Bogdanov et al. 2017; Haller and Schwabe 2014).

High-cost trials: trials with costs of the preferred options (chosen in the preference phase) being ¥80,000 or ¥100,000 (22 trials: 11 trials for sunk cost condition and 11 trials for control condition).

Low-cost trials: trials with costs of the preferred options (chosen in the preference phase) being ¥30,000, ¥40,000, or ¥50,000 (24 trials: 12 trials for sunk cost condition and 12 trials for control condition).

We then calculated the difference in the sunk cost effect score between the high-cost and low-cost trials (sunk cost effect score in the high-cost trials–sunk cost effect score in the low-cost trials).

Finally, we performed correlation analyses between the sunk cost effect score/the difference in the sunk cost effect score between the high-cost and low-cost trials and the severity of clinical symptoms evaluated using the ADOS (communication subscale score, social interaction subscale score, and communication subscale + social interaction subscale score) across the participants with ASD (after controlling for age, gender, and IQ levels).

Because some of our continuous measures were not normally distributed (Shapiro–Wilk test, p < 0.05), we chose Mann–Whitney tests for comparing group differences and Spearman’s rank correlations for the correlation analyses. For analyses of variance (ANOVA), we performed robust ANOVA using WRS2 package (Mair et al. 2016) within the statistical software R (R Core Team 2016), which is comparable to the non-parametric analyses applied in the current study on the basis of previous studies (e.g., Desir and Karatekin 2018; Krzyzanowska et al. 2017; Tei et al. 2018). Specifically, we applied the bwtrim function, which returns the test statistic value of Q that is approximately F-distributed (Krzyzanowska et al. 2017; Mair et al. 2016; Tei et al. 2018). Statistical analyses, except for robust ANOVA, were performed using SPSS 21 (IBM, Armonk, NY, USA). Results were considered statistically significant at p < 0.05 (two-tailed).

Results

Two participants with ASD were excluded from the analyses because even under the control condition, they frequently chose the option that was expected to be less enjoyable, suggesting that they did not understand the task. They chose the less enjoyable option at rates of 56.5 and 52.2% each, which were extremely high compared with the rates of the participants [> 3 standard deviation (SD) + mean]. Thus, data from 27 participants with ASD and 29 controls were analyzed (age range: 20–47 years). Demographic and clinical data are shown in Table 1. There were no significant differences between the groups in age, gender, current smoking status, education, and estimated full-scale IQ levels. In total, 12 participants with ASD were administered psychotropic drugs comprising anxiolytics (N = 6), antidepressants (N = 5), antipsychotics (N = 5), antiepileptics (N = 1), hypnotic drugs (N = 6), and other psychotropic drugs (N = 3).

Table 1

Demographic and clinical characteristics of participants

 

Control group

ASD group

Statistics

(N = 29)

(N = 27)

p

Age (years)

30.9 ± 8.4

29.2 ± 4.2

0.71a

Gender (male/female)

26/3

24/3

0.93b

Current smoker/nonsmoker

6/23

3/24

0.33b

Education (years)

14.7 ± 2.0

15.5 ± 1.9

0.09a

Estimated full-scale IQ

106.1 ± 7.8

108.0 ± 12.8

0.59a

ADOSc

 Communication

 

3.8 ± 1.4

 

 Social interaction

 

7.0 ± 2.1

 

 Communication + social interaction

 

10.8 ± 3.3

 

ADOS autism diagnostic observation schedule, IQ intelligence quotient

aMann–Whitney test

bTwo-tailed Chi square test

cData not available for one participant

Overall, the participants [N = 56 (control 29, ASD 27)] performed the task well, missing an average of only 0.11 ± 0.37 (mean ± SD) trials (total, 46 trials). Briefly, a missed trial is a trial in which a participant could not make a decision within the time the option was presented (5 s, Fig. 1). There was no significant difference between the groups in the number of the missed trials (p = 0.21). We also found no significant group differences in the reaction time in the preference phase (p = 0.90).

Figure 2a shows the reaction time in the sunk cost condition and the control condition in both groups. A robust two-way mixed ANOVA revealed the main effects of both group (Q = 7.00, p = 0.013) and condition (Q = 58.29, p < 0.01) as well as a group by condition interaction (Q = 13.33, p < 0.01). Post hoc Mann–Whitney tests revealed that this interaction was driven by a group difference in the reaction time under the sunk cost condition (control 1.60 ± 0.57, ASD 1.10 ± 0.38, p < 0.01), whereas no such difference was observed under the control condition (control 0.96 ± 0.18, ASD 0.91 ± 0.28, p = 0.42).

Fig. 2

Behavioral data in the sunk cost effect task. a Reaction time in the sunk cost and control conditions. The reaction time under the sunk cost condition was significantly shorter in the ASD group than in the control group (p < 0.01), and no such difference was observed under the control condition (p = 0.42). Error bars indicate ± standard errors. b The sunk cost effect score in the control and ASD groups. The sunk cost effect score was reduced in the ASD group compared with that in the control group (p < 0.01)

The mean sunk cost effect score in the control individuals was 58.5 ± 28.5%, thus indicating that their decisions were considerably influenced by sunk costs. This susceptibility to the sunk cost effect was significantly reduced in the ASD group (33.4 ± 39.9%) compared with that in the control group (p < 0.01) (Fig. 2b). Considering that 12 participants with ASD took psychotropic drugs, we compared the sunk cost effect score of participants with ASD who were not taking psychotropic drugs (N = 15) with that of the control participants. The analysis did not materially change the result: the sunk cost effect score among participants with ASD who were not taking psychotropic drugs was marginally significantly reduced compared with the control group (p = 0.08). In addition, to take into account the participants’ preference for a city as a possible confounding factor, we performed the additional analyses using the reaction time in the preference phase. The analyses did not materially change the results. Please see Supplementary Results for details.

To further explore how the participants behaved when considering sunk costs, we separated the trials on the basis of ticket costs, as described in the Methods. The difference in the sunk cost effect score between the high-cost trials and the low-cost trials was significantly reduced in the ASD group compared with the control group (control 15.1 ± 24.9%, ASD 0.8 ± 17.6%, p = 0.02, Fig. 3a). The significance of this observation was confirmed in an interaction analysis showing that the sunk cost effect score was less influenced by ticket costs in the ASD group than in the control group [robust ANOVA; group (control, ASD) × ticket costs (high, low); Q = 4.37, p = 0.045, Fig. 3b].

Fig. 3

Sunk cost effect score in the high- and low-cost trials. a The difference between the sunk cost effect score in the high- and low-cost trials was significantly reduced in the ASD group compared with that in the control group (p = 0.02). Error bars indicate ± standard errors. b A robust two-way mixed ANOVA revealed a significant interaction between group and ticket costs (p = 0.045). Error bars indicate ± standard errors

Thereafter, we performed correlation analyses between the sunk cost effect score/the difference in the sunk cost effect score between the high-cost and low-cost trials and the severity of clinical symptoms evaluated using ADOS (communication subscale score, social interaction subscale score, and communication subscale + social interaction subscale score) among the ASD participants. We found no significant relationship between the sunk cost effect score/the difference in the sunk cost effect score between the high-cost and low-cost trials and the severity of the clinical symptoms (all, p > 0.50).

Discussion

To the best of our knowledge, this is the first study to investigate the sunk cost effect in individuals with ASD. Our results accord with previous evidence of reduced sensitivity to context stimuli in ASD, and extend this observation to the context of the sunk cost effect.

The control group exhibited a typical sunk cost effect (Arkes and Blumer 1985; Fujino et al. 2016b; Haller and Schwabe 2014). Given that an invested sunk cost cannot be recovered, a rational forward-looking decision maker would be expected to ignore sunk costs (Arkes and Blumer 1985; Augenblick 2016; Friedman et al. 2007; Thaler 1980). However, under the experimental sunk cost condition in this study, the control participants frequently (on more than half the occasions on average) chose the option that was considered less enjoyable. Furthermore, the sunk cost effect score was higher in the high-cost trials than in the low-cost trials.

Consistent with our hypothesis, the sunk cost effect was reduced in the ASD group compared with that in the control group. Furthermore, unlike in the control participants, the sunk cost effect score among participants with ASD was not increased when larger sunk costs were incurred. Individuals with ASD have frequently been reported to show atypical performance on tasks that require processing of local information independently of its context. For example, individuals with ASD are better at finding figures embedded in complex shapes than are control participants (Jolliffe and Baron-Cohen 1997); in addition, individuals with ASD are less affected by the distractors in visual search (Plaisted et al. 1998). Many of these studies focused on perceptual tasks, such as pitch discrimination, visual search, and motion-coherence detection, with corresponding theoretical frameworks that emphasize “low-level” processes, such as enhanced perceptual discrimination (Farmer et al. 2017; Happé and Frith 2006). Recently, studies on healthy subjects have shown a substantial overlap in brain activation between perceptual processing and “higher-level” cognitive processing such as decision making and moral judgment (Decety and Lamm 2007; Tei et al. 2017). The authors proposed that social cognitive processing abilities depend at least in part on perceptual processing (Decety and Lamm 2007; Tei et al. 2017). In line with this notion, recent studies of ASD have shown that the reduced context sensitivity that characterizes ASD can be seen in “high-level” decision-making tasks. Specifically, De Martino et al. (2008) and Shah et al. (2016) examined the framing effect on monetary decisions in individuals with ASD and showed that individuals with ASD are less susceptible to the framing effect and make more consistent choices. Farmer et al. (2017) compared a series of choices regarding consumer products between participants with ASD and controls and showed that participants’ preferences between a given pair of options frequently switched when the third item in the set was changed. This tendency was reduced among participants with ASD, thereby indicating more consistent and conventionally rational choices. Taken together, our results are in line with these previous experimental findings and demonstrate that the conventionally more rational responses of individuals with ASD can also be seen in the context of the sunk cost effect.

Individuals with ASD often take longer to make decisions (Brosnan et al. 2014; Luke et al. 2012). Interestingly, however, the choice reaction time under the sunk cost condition in our task was significantly shorter in the ASD group than in the control group. The weak central coherence hypothesis is one of the most prominent theories concerning the atypical performance of individuals with ASD on tasks that involve local and global processing (Happé and Frith 2006; Plaisted et al. 2003). In individuals with typical development, a common effect is that the global level of the stimulus dominates responding, with slower and less accurate responses at the local level (Navon 1977; Plaisted et al. 2003). Although there are mixed findings, a number of recent studies have shown that individuals with ASD can respond to the global level of a hierarchical stimulus in the same way as comparison individuals (Mottron and Belleville 1993; Plaisted et al. 1999, 2003; Happé and Frith 2006). In addition, under some circumstances, individuals with ASD have been found to show faster and more accurate responses to the local level than comparison individuals (Mottron and Belleville 1993; Plaisted et al. 1999, 2003; Happé and Frith 2006). In light of these studies, the weak coherence account has recently refined toward an emphasis on superiority in local processing rather than deficit in global processing (Happé and Frith 2006; Plaisted et al. 2003). The results of the decision-making task in the study by Farmer et al. (2017) were also in line with these previous findings, and the authors highlight that ASD is characterized by a wide ranging enhancement of or preference for local information processing (Farmer et al. 2017). Although this is speculative, the current results may agree with the recent suggestions (Farmer et al. 2017; Happé and Frith 2006; Plaisted et al. 2003) and illuminated the positive aspect of the atypical decision-making styles in individuals with ASD. Regarding decision making in individuals with ASD, Brosnan et al. (2016, 2017) applied Dual Process Theory (Evans 2008; Lieberman 2007) and recently proposed a useful framework that reasoning by individuals with ASD can be characterized by reduced intuitive (Type 1) and greater reflective (Type 2) processing. Previous functional magnetic resonance imaging studies on healthy subjects have revealed that several brain regions, including the medial prefrontal cortex, lateral prefrontal cortex, anterior cingulate cortex, and insula, are involved in decision making under sunk costs (Fujino et al. 2016b; Haller and Schwabe 2014; Zeng et al. 2013). These areas play key roles in Type 1 and Type 2 processing (Fujino et al. 2017b; Hallsson et al. 2018; Lieberman 2007) and have been repeatedly reported to be altered in individuals with ASD (Philip et al. 2012). Future neuroimaging studies, together with a framework proposed by Brosnan et al. (2016, 2017), should provide deeper insights into the mechanisms of altered decision making among individuals with ASD when considering sunk costs.

The findings of the present study have practical implications on the socioeconomic functioning of individuals with ASD because the sunk cost effect influences many types of socioeconomic behavior, such as entertainment (Roth et al. 2015), investments (Bogdanov et al. 2017), management (Conlon and Garland 1993), and interpersonal relations (Arkes and Blumer 1985). The sunk cost effect is considered as an irrational behavior according to economic theory, and such a decision bias can lead to severe financial or political consequences, such as continuation of an unprofitable building project or even war (Arkes and Blumer 1985). By providing additional evidence, the present study supports the notion that ASD is not a disability in all respects (Baron-Cohen 2000; Farmer et al. 2017). Elucidating the heterogeneity of symptom expression in ASD is crucial for obtaining better understanding of the underlying neurobiological mechanisms as well as for the establishment of precise treatment strategies (Happé and Frith 2006). A behavioral economics approach can help in elucidating existing symptomatology, or inform the development of new mediating markers and the personalization of treatment (Fujino et al. 2017a; Ruderman et al. 2016; Sharp et al. 2012). In this study, we did not find any significant correlation between the measures related to the sunk cost effect and severity of clinical symptoms among participants with ASD. However, the distributions of the sunk cost effect score in our task in ASD participants were diverse. Comprehensive measurements with various behavioral economics tools in larger samples would lead to a better understanding of the heterogeneity of altered decision-making in ASD.

There were several limitations to our study. First, we used a hypothetical choice experiment to understand actual choices under the influence of sunk costs. Although decision making regarding hypothetical rewards does not necessarily reflect real-life decision-making behavior, the validity of the results of experiments with hypothetical rewards has been reported (Bray et al. 2010; Kang et al. 2011; Locey et al. 2011). Therefore, we consider our findings to be useful in understanding and predicting real choices under the influence of sunk costs. Second, nearly half of participants with ASD were administered psychotropic medication, thus indicating that we cannot exclude the possibility of a medication effect. For example, dopaminergic agents have been reported to influence value-based decision making (Doya 2008). In addition, previous studies have shown that serotoninergic agents have effects on discovering bad decision outcomes and risk-seeking behavior (Rogers 2011). However, the sunk cost effect score of participants with ASD who were not taking psychotropic drugs was also marginally significantly reduced compared with that of the control participants. Third, comorbid anxiety and depression are prevalent in individuals with ASD (Boulter et al. 2014; Luke et al. 2012). However, in this study, we did not quantify the participants’ levels of anxiety/depression or control these variables. Individuals with anxiety/depression have been reported to show altered performance in various types of decision-making tasks in clinical groups and following mood induction in nonclinical samples (Blanchette and Richards 2010; Hartley and Phelps 2012; Pulcu et al. 2014). In addition, previous studies have suggested that difficulties in decision making by individuals with ASD may be exacerbated by higher levels of anxiety and depression (Luke et al. 2012). Therefore, future studies should focus on detailed research on how symptoms of anxiety and depression influence decision making under sunk costs in individuals with ASD. Fourth, our ASD sample consisted of only high-functioning individuals with ASD. Future studies recruiting a larger number of ASD individuals with diverse IQ levels and not taking medication are needed to replicate and strengthen our findings.

Notwithstanding these limitations, the current results extend previous findings by showing that the reduced context sensitivity that characterizes ASD spills over into the context of the sunk cost effect. Our findings will contribute to a better understanding of altered decision making in individuals with ASD and may be useful in addressing the practical implications of their socioeconomic behavior.

Notes

Acknowledgments

The authors wish to extend their gratitude to the research team of the Medical Institute of Developmental Disabilities Research at Showa University for their assistance in data acquisition. This work was supported by grants-in-aid for scientific research A (24243061), C (17K10326), Young Scientists B (17K16398), and on Innovative Areas (23120009, 16H06572), from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT); SENSHIN Medical Research Foundation; and Takeda Science Foundation. A part of this study is the result of the Strategic Research Program for Brain Sciences by agency for medical research and development (JP18dm0107151), “Research and development of technology for enhancing functional recovery of elderly and disabled people based on non-invasive brain imaging and robotic assistive devices”, the Commissioned Research of National Institute of Information and Communications Technology, JAPAN, and the Joint Usage/Research Program of Medical Institute of Developmental Disabilities Research, Showa University. These agencies had no further role in the study design, the collection, analysis and interpretation of data, the writing of the report, or in the decision to submit the paper for publication.

Author Contributions

JF, ST, TI, YA, HO, RH, MN, NK, and HT designed research; JF, ST, TI, and CK participated in the data acquisition; JF, YA HO, CK, MK, MN, and NK were in charge of the clinical assessment. JF, and ST analyzed data; TI, YA HO, CK, MK, RH, MN, NK, and HT helped with interpretation of data. JF, ST, TI, YA, HO, CK, MK, RH, MN, NK, and HT wrote the paper. All authors have made intellectual contribution to the work and approved the final version of the manuscript for submission.

Compliance with Ethical Standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.

Supplementary material

10803_2018_3679_MOESM1_ESM.docx (54 kb)
Supplementary material 1 (DOCX 53 KB)

References

  1. Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140.Google Scholar
  2. Augenblick, N. (2016). The sunk-cost fallacy in penny auctions. The Review of Economic Studies, 83(1), 58–86.Google Scholar
  3. Baron-Cohen, S. (2000). Is asperger syndrome/high-functioning autism necessarily a disability? Development and Psychopathology, 12(3), 489–500.Google Scholar
  4. Blanchette, I., & Richards, A. (2010). The influence of affect on higher level cognition: A review of research on interpretation, judgement, decision making and reasoning. Cognition & Emotion, 24(4), 561–595.Google Scholar
  5. Bogdanov, M., Ruff, C. C., & Schwabe, L. (2017). Transcranial stimulation over the dorsolateral prefrontal cortex increases the impact of past expenses on decision-making. Cerebral Cortex, 27(2), 1094–1102.Google Scholar
  6. Boulter, C., Freeston, M., South, M., & Rodgers, J. (2014). Intolerance of uncertainty as a framework for understanding anxiety in children and adolescents with autism spectrum disorders. Journal of Autism and Developmental Disorders, 44(6), 1391–1402.Google Scholar
  7. Bray, S., Shimojo, S., & O’Doherty, J. P. (2010). Human medial orbitofrontal cortex is recruited during experience of imagined and real rewards. Journal of Neurophysiology, 103(5), 2506–2512.Google Scholar
  8. Brosnan, M., Ashwin, C., & Lewton, M. (2017). Brief report: Intuitive and reflective reasoning in autism spectrum disorder. Journal of Autism and Developmental Disorders, 47(8), 2595–2601.Google Scholar
  9. Brosnan, M., Chapman, E., & Ashwin, C. (2014). Adolescents with autism spectrum disorder show a circumspect reasoning bias rather than ‘jumping-to-conclusions’. Journal of Autism and Developmental Disorders, 44(3), 513–520.Google Scholar
  10. Brosnan, M., Lewton, M., & Ashwin, C. (2016). Reasoning on the autism spectrum: A dual process theory account. Journal of Autism and Developmental Disorders, 46(6), 2115–2125.Google Scholar
  11. Camerer, C. F. (2011). Behavioral game theory: Experiments in strategic interaction. Princeton: Princeton University Press.Google Scholar
  12. Conlon, D. E., & Garland, H. (1993). The role of project completion information in resource allocation decisions. Academy of Management Journal, 36(2), 402–413.Google Scholar
  13. De Martino, B., Harrison, N. A., Knafo, S., Bird, G., & Dolan, R. J. (2008). Explaining enhanced logical consistency during decision making in autism. The Journal of Neuroscience, 28(42), 10746–10750.Google Scholar
  14. Decety, J., & Lamm, C. (2007). The role of the right temporoparietal junction in social interaction: How low-level computational processes contribute to meta-cognition. Neuroscientist, 13(6), 580–593.Google Scholar
  15. Desir, M. P., & Karatekin, C. (2018). Parental reactions to parent-and sibling-directed aggression within a domestic violence context. Clinical Child Psychology and Psychiatry, 2018, 1359104518755219.Google Scholar
  16. Doya, K. (2008). Modulators of decision making. Nature Neuroscience, 11(4), 410–416.Google Scholar
  17. Evans, J. S. B. (2008). Dual-processing accounts of reasoning, judgment, and social cognition. Annual Review of Psychology, 59, 255–278.Google Scholar
  18. Farmer, G. D., Baron-Cohen, S., & Skylark, W. J. (2017). People with autism spectrum conditions make more consistent decisions. Psychological Science, 28(8), 1067–1076.Google Scholar
  19. Friedman, D., Pommerenke, K., Lukose, R., Milam, G., & Huberman, B. A. (2007). Searching for the sunk cost fallacy. Experimental Economics, 10(1), 79–104.Google Scholar
  20. Fujino, J., Fujimoto, S., Kodaka, F., Camerer, C. F., Kawada, R., Tsurumi, K., et al. (2016b). Neural mechanisms and personality correlates of the sunk cost effect. Scientific Reports, 6, 33171.Google Scholar
  21. Fujino, J., Hirose, K., Tei, S., Kawada, R., Tsurumi, K., Matsukawa, N., et al. (2016a). Ambiguity aversion in schizophrenia: An fMRI study of decision-making under risk and ambiguity. Schizophrenia Research, 178(1), 94–101.Google Scholar
  22. Fujino, J., Tei, S., Hashimoto, R. I., Itahashi, T., Ohta, H., Kanai, C., et al. (2017a). Attitudes toward risk and ambiguity in patients with autism spectrum disorder. Molecular Autism, 8(1), 45.Google Scholar
  23. Fujino, J., Tei, S., Jankowski, K. F., Kawada, R., Murai, T., & Takahashi, H. (2017b). Role of spontaneous brain activity in explicit and implicit aspects of cognitive flexibility under socially conflicting situations: A resting-state fMRI study using fractional amplitude of low-frequency fluctuations. Neuroscience, 367, 60–71.Google Scholar
  24. Haller, A., & Schwabe, L. (2014). Sunk costs in the human brain. Neuroimage, 97, 127–133.Google Scholar
  25. Hallsson, B. G., Siebner, H. R., & Hulme, O. J. (2018). Fairness, fast and slow: A review of dual process models of fairness. Neuroscience & Biobehavioral Reviews, 89, 49–60.Google Scholar
  26. Happé, F., & Frith, U. (2006). The weak coherence account: Detail-focused cognitive style in autism spectrum disorders. Journal of Autism and Developmental Disorders, 36(1), 5–25.Google Scholar
  27. Hartley, C. A., & Phelps, E. A. (2012). Anxiety and decision-making. Biological Psychiatry, 72(2), 113–118.Google Scholar
  28. Itahashi, T., Yamada, T., Watanabe, H., Nakamura, M., Ohta, H., Kanai, C., et al. (2015). Alterations of local spontaneous brain activity and connectivity in adults with high-functioning autism spectrum disorder. Molecular Autism, 6(1), 1.Google Scholar
  29. Jolliffe, T., & Baron-Cohen, S. (1997). Are people with autism and asperger syndrome faster than normal on the embedded figures test? Journal of Child Psychology and Psychiatry, 38(5), 527–534.Google Scholar
  30. Just, D. R., & Wansink, B. (2011). The flat-rate pricing paradox: Conflicting effects of “all-you-can-eat” buffet pricing. The Review of Economics and Statistics, 93(1), 193–200.Google Scholar
  31. Kang, M. J., Rangel, A., Camus, M., & Camerer, C. F. (2011). Hypothetical and real choice differentially activate common valuation areas. The Journal of Neuroscience, 31(2), 461–468.Google Scholar
  32. Krzyzanowska, K., Collins, P. J., & Hahn, U. (2017). Between a conditional’s antecedent and its consequent: Discourse coherence vs. probabilistic relevance. Cognition, 164, 199–205.Google Scholar
  33. Lieberman, M. D. (2007). Social cognitive neuroscience: A review of core processes. Annual Review of Psychology, 58, 259–289.Google Scholar
  34. Locey, M. L., Jones, B. A., & Rachlin, H. (2011). Real and hypothetical rewards. Judgment and Decision Making, 6(6), 552–564.Google Scholar
  35. Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule—generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30(3), 205–223.Google Scholar
  36. Luke, L., Clare, I. C., Ring, H., Redley, M., & Watson, P. (2012). Decision-making difficulties experienced by adults with autism spectrum conditions. Autism, 16(6), 612–621.Google Scholar
  37. Magalhães, P., White, K. G., Stewart, T., Beeby, E., & van der Vliet, W. (2012). Suboptimal choice in nonhuman animals: Rats commit the sunk cost error. Learning & Behavior, 40(2), 195–206.Google Scholar
  38. Mair, P., Schoenbrodt, F., & Wilcox, R. (2016). WRS2: Wilcox robust estimation and testing.Google Scholar
  39. Matsuoka, K., & Kim, Y. (2006). Japanese adult reading test (JART). Tokyo: Shinkou-Igaku Publishers.Google Scholar
  40. Matsuoka, K., Uno, M., Kasai, K., Koyama, K., & Kim, Y. (2006). Estimation of premorbid IQ in individuals with Alzheimer’s disease using Japanese ideographic script (Kanji) compound words: Japanese version of National Adult Reading Test. Psychiatry and Clinical Neurosciences, 60(3), 332–339.Google Scholar
  41. Minassian, A., Paulus, M., Lincoln, A., & Perry, W. (2007). Adults with autism show increased sensitivity to outcomes at low error rates during decision-making. Journal of Autism and Developmental Disorders, 37(7), 1279–1288.Google Scholar
  42. Mobbs, D., & Kim, J. J. (2015). Neuroethological studies of fear, anxiety, and risky decision-making in rodents and humans. Current Opinion in Behavioral Sciences, 5, 8–15.Google Scholar
  43. Mosner, M. G., Kinard, J. L., McWeeny, S., Shah, J. S., Markiewitz, N. D., Damiano-Goodwin, C. R., et al. (2017). Vicarious effort-based decision-making in autism spectrum disorders. Journal of Autism and Developmental Disorders, 47(10), 2992–3006.Google Scholar
  44. Mottron, L., & Belleville, S. (1993). A study of perceptual analysis in a high-level autistic subject with exceptional graphic abilities. Brain and Cognition, 23(2), 279–309.Google Scholar
  45. Mussey, J. L., Travers, B. G., Klinger, L. G., & Klinger, M. R. (2015). Decision-making skills in ASD: Performance on the Iowa Gambling Task. Autism Research, 8(1), 105–114.Google Scholar
  46. Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9(3), 353–383.Google Scholar
  47. Philip, R. C., Dauvermann, M. R., Whalley, H. C., Baynham, K., Lawrie, S. M., & Stanfield, A. C. (2012). A systematic review and meta-analysis of the fMRI investigation of autism spectrum disorders. Neuroscience & Biobehavioral Reviews, 36(2), 901–942.Google Scholar
  48. Plaisted, K., O’Riordan, M., & Baron-Cohen, S. (1998). Enhanced visual search for a conjunctive target in autism: A research note. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 39(5), 777–783.Google Scholar
  49. Plaisted, K., Saksida, L., Alcántara, J., & Weisblatt, E. (2003). Towards an understanding of the mechanisms of weak central coherence effects: Experiments in visual configural learning and auditory perception. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 358(1430), 375–386.Google Scholar
  50. Plaisted, K., Swettenham, J., & Rees, L. (1999). Children with autism show local precedence in a divided attention task and global precedence in a selective attention task. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 40(5), 733–742.Google Scholar
  51. Pulcu, E., Trotter, P., Thomas, E., McFarquhar, M., Juhász, G., Sahakian, B., et al. (2014). Temporal discounting in major depressive disorder. Psychological Medicine, 44(9), 1825–1834.Google Scholar
  52. R Core Team. (2016). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing.Google Scholar
  53. Robic, S., Sonie, S., Fonlupt, P., Henaff, M.-A., Touil, N., Coricelli, G., et al. (2015). Decision-making in a changing world: A study in autism spectrum disorders. Journal of Autism and Developmental Disorders, 45(6), 1603–1613.Google Scholar
  54. Rogers, R. D. (2011). The roles of dopamine and serotonin in decision making: Evidence from pharmacological experiments in humans. Neuropsychopharmacology, 36(1), 114–132.Google Scholar
  55. Roth, S., Robbert, T., & Straus, L. (2015). On the sunk-cost effect in economic decision-making: A meta-analytic review. Business Research, 8(1), 99–138.Google Scholar
  56. Ruderman, L., Ehrlich, D. B., Roy, A., Pietrzak, R. H., Harpaz-Rotem, I., & Levy, I. (2016). Posttraumatic stress symptoms and aversion to ambiguous losses in combat veterans. Depression and Anxiety, 33(7), 606–613.Google Scholar
  57. Shah, P., Catmur, C., & Bird, G. (2016). Emotional decision-making in autism spectrum disorder: The roles of interoception and alexithymia. Molecular Autism, 7(1), 43.Google Scholar
  58. Sharp, C., Monterosso, J., & Montague, P. R. (2012). Neuroeconomics: A bridge for translational research. Biological Psychiatry, 72(2), 87–92.Google Scholar
  59. Sleesman, D. J., Conlon, D. E., McNamara, G., & Miles, J. E. (2012). Cleaning up the big muddy: A meta-analytic review of the determinants of escalation of commitment. Academy of Management Journal, 55(3), 541–562.Google Scholar
  60. Stanovich, K. E., & Stanovich, P. J. (2010). A framework for critical thinking, rational thinking, and intelligence. In D. Preiss & R. J. Sternberg (Eds.), Innovations in educational psychology: Perspectives on learning, teaching and human development (pp. 195–237). New York: Springer.Google Scholar
  61. Strough, J., Mehta, C. M., McFall, J. P., & Schuller, K. L. (2008). Are older adults less subject to the sunk-cost fallacy than younger adults? Psychological Science, 19(7), 650–652.Google Scholar
  62. Takahashi, H. (2013). Molecular neuroimaging of emotional decision-making. Neuroscience Research, 75(4), 269–274.Google Scholar
  63. Tan, H.-T., & Yates, J. F. (1995). Sunk cost effects: The influences of instruction and future return estimates. Organizational Behavior and Human Decision Processes, 63(3), 311–319.Google Scholar
  64. Tei, S., Fujino, J., Hashimoto, R. I., Itahashi, T., Ohta, H., Kanai, C., et al. (2018). Inflexible daily behaviour is associated with the ability to control an automatic reaction in autism spectrum disorder. Scientific Reports, 8(1), 8082.Google Scholar
  65. Tei, S., Fujino, J., Kawada, R., Jankowski, K. F., Kauppi, J. P., van den Bos, W.,et al (2017). Collaborative roles of temporoparietal junction and dorsolateral prefrontal cortex in different types of behavioural flexibility. Scientific Reports, 7(1), 6415.Google Scholar
  66. Thaler, R. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior & Organization, 1(1), 39–60.Google Scholar
  67. Thaler, R. (1994). The winner’s curse: Paradoxes and anomalies of economic life. Princeton: Princeton University Press.Google Scholar
  68. Thiébaut, F. I., White, S. J., Walsh, A., Klargaard, S. K., Wu, H.-C., Rees, G., et al. (2016). Does faux pas detection in adult autism reflect differences in social cognition or decision-making abilities? Journal of Autism and Developmental Disorders, 46(1), 103–112.Google Scholar
  69. Tversky, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty, 5(4), 297–323.Google Scholar
  70. Zeng, J., Zhang, Q., Chen, C., Yu, R., & Gong, Q. (2013). An fMRI study on sunk cost effect. Brain Research, 1519, 63–70.Google Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  • Junya Fujino
    • 1
    • 2
  • Shisei Tei
    • 1
    • 2
    • 3
    • 4
  • Takashi Itahashi
    • 1
  • Yuta Aoki
    • 1
  • Haruhisa Ohta
    • 1
    • 5
  • Chieko Kanai
    • 1
  • Manabu Kubota
    • 1
    • 2
    • 6
  • Ryu-ichiro Hashimoto
    • 1
    • 7
  • Motoaki Nakamura
    • 1
    • 8
  • Nobumasa Kato
    • 1
  • Hidehiko Takahashi
    • 1
    • 2
  1. 1.Medical Institute of Developmental Disabilities ResearchShowa UniversityTokyoJapan
  2. 2.Department of Psychiatry, Graduate School of MedicineKyoto UniversityKyotoJapan
  3. 3.Institute of Applied Brain SciencesWaseda UniversityTokorozawaJapan
  4. 4.School of Human and Social SciencesTokyo International UniversityKawagoeJapan
  5. 5.Department of Psychiatry, School of MedicineShowa UniversityTokyoJapan
  6. 6.Department of Functional Brain Imaging Research, National Institute of Radiological SciencesNational Institutes for Quantum and Radiological Science and TechnologyChibaJapan
  7. 7.Department of Language Sciences, Graduate School of HumanitiesTokyo Metropolitan UniversityTokyoJapan
  8. 8.Kanagawa Psychiatric CenterYokohamaJapan

Personalised recommendations