Saturday, April 17, 2021

Many practitioners in the US are unaccustomed to using probability in diagnosis and clinical practice; widespread overestimates of the probability of disease likely contribute to overdiagnosis and overuse

Accuracy of Practitioner Estimates of Probability of Diagnosis Before and After Testing. Daniel J. Morgan et al. JAMA Intern Med., April 5, 2021. DOI 10.1001/jamainternmed.2021.0269

Key Points

Question  Do practitioners understand the probability of common clinical diagnoses?

Findings  In this survey study of 553 practitioners performing primary care, respondents overestimated the probability of diagnosis before and after testing. This posttest overestimation was associated with consistent overestimates of pretest probability and overestimates of disease after specific diagnostic test results.

Meaning  These findings suggest that many practitioners are unaccustomed to using probability in diagnosis and clinical practice. Widespread overestimates of the probability of disease likely contribute to overdiagnosis and overuse.


Abstract

Importance  Accurate diagnosis is essential to proper patient care.

Objective  To explore practitioner understanding of diagnostic reasoning.

Design, Setting, and Participants  In this survey study, 723 practitioners at outpatient clinics in 8 US states were asked to estimate the probability of disease for 4 scenarios common in primary care (pneumonia, cardiac ischemia, breast cancer screening, and urinary tract infection) and the association of positive and negative test results with disease probability from June 1, 2018, to November 26, 2019. Of these practitioners, 585 responded to the survey, and 553 answered all of the questions. An expert panel developed the survey and determined correct responses based on literature review.

Results  A total of 553 (290 resident physicians, 202 attending physicians, and 61 nurse practitioners and physician assistants) of 723 practitioners (76.5%) fully completed the survey (median age, 32 years; interquartile range, 29-44 years; 293 female [53.0%]; 296 [53.5%] White). Pretest probability was overestimated in all scenarios. Probabilities of disease after positive results were overestimated as follows: pneumonia after positive radiology results, 95% (evidence range, 46%-65%; comparison P < .001); breast cancer after positive mammography results, 50% (evidence range, 3%-9%; P < .001); cardiac ischemia after positive stress test result, 70% (evidence range, 2%-11%; P < .001); and urinary tract infection after positive urine culture result, 80% (evidence range, 0%-8.3%; P < .001). Overestimates of probability of disease with negative results were also observed as follows: pneumonia after negative radiography results, 50% (evidence range, 10%-19%; P < .001); breast cancer after negative mammography results, 5% (evidence range, <0.05%; P < .001); cardiac ischemia after negative stress test result, 5% (evidence range, 0.43%-2.5%; P < .001); and urinary tract infection after negative urine culture result, 5% (evidence range, 0%-0.11%; P < .001). Probability adjustments in response to test results varied from accurate to overestimates of risk by type of test (imputed median positive and negative likelihood ratios [LRs] for practitioners for chest radiography for pneumonia: positive LR, 4.8; evidence, 2.6; negative LR, 0.3; evidence, 0.3; mammography for breast cancer: positive LR, 44.3; evidence range, 13.0-33.0; negative LR, 1.0; evidence range, 0.05-0.24; exercise stress test for cardiac ischemia: positive LR, 21.0; evidence range, 2.0-2.7; negative LR, 0.6; evidence range, 0.5-0.6; urine culture for urinary tract infection: positive LR, 9.0; evidence, 9.0; negative LR, 0.1; evidence, 0.1).

Conclusions and Relevance  This survey study suggests that for common diseases and tests, practitioners overestimate the probability of disease before and after testing. Pretest probability was overestimated in all scenarios, whereas adjustment in probability after a positive or negative result varied by test. Widespread overestimates of the probability of disease likely contribute to overdiagnosis and overuse.

Discussion

In this survey study, in scenarios commonly encountered in primary care practice, practitioners overestimated the probability of disease by 2 to 10 times compared with the scientific evidence, both before and after testing. This result was mostly associated with overestimates of pretest probability, which were observed across all scenarios. Adjustments to probability in response to test results varied from accurate to overestimates of risk by type of test. There was variation in accuracy between type of practitioner that was small compared with the magnitude of difference between practitioners and the scientific evidence. Many practitioners reported that they would treat patients for disease for which likelihood had been overestimated.

The most striking finding from this study was that practitioners consistently and significantly overestimate the likelihood of disease. Small studies with limited generalizability have had similar findings, often asking practitioners to perform one isolated aspect of diagnosis, such as interpreting a test result. However, past studies8-11 have not explored a range of questions or clarified estimates at different steps in the diagnostic pathway. The reason for inaccurate estimates of probability are not clear, although anecdotes reported during the current study imply that practitioners often do not think in terms of probability. One participant stated that estimating probability of disease “isn’t how you do medicine.” This attitude is consistent with a previous study20 of diagnostic strategies that describe an initial pattern recognition phase of care with only 10% of practitioners engaging in a secondary phase of probabilistic reasoning.

This study found that probability estimates were consistently biased toward overestimation, as has been seen in other contexts, such as expectations of high stock returns among investors.21 This overestimation is consistent with cognitive biases, including base rate neglect, anchoring bias, and confirmation bias.14 These biases drive overestimation because true base rates are usually lower than expected and anchoring tends to reflect experiences that represent improbable events or those in which a diagnosis was missed. Such cognitive biases have been associated with diagnostic errors that may occur from errors in estimating risk.5,22,23 Notably, practitioners in this survey were often residents or academic physicians who often practice with populations with higher prevalence of disease. This experience may have also contributed to higher estimates of disease.

Pretest probabilities were consistently overestimated for all questions, but overestimates were particularly apparent for the pneumonia and UTI scenarios. Estimates of pretest probability generally reflect clinical knowledge. Reasons for overestimates for these infectious diseases may relate to the fact that antibiotics are often appropriately given even when the likelihood of infection is moderate. In the UTI scenario, estimates of high pretest probability may reflect the evolution of the definition of asymptomatic bacteriuria as a separate entity from UTI.24

In contrast to past literature,8-10,19 practitioners accurately adjusted estimates of disease based on the results of some tests, as demonstrated by the imputed likelihood ratios. This adjustment could be artifactual because of inability to adjust probability for tests that had high pretest estimates (ie, pneumonia and UTI). In other cases, practitioners markedly overestimated the probability of disease after testing, specifically after a positive or negative mammography result or a positive exercise stress test result. Practitioners are known to overestimate chance of disease when completing a theoretical estimate of likelihood of disease after a positive test result when pretest probability was 1 in 1000 tests.9,10 The current study included the identical question with an identical response, with participants estimating the likelihood of disease at 95% when the correct answer was 2%.5,8-10,19 The findings regarding real-life examples are consistent with evidence from limited past studies,8-11 for example, physician interpretation of a positive mammography result in a typical woman as conveying 81% probability of breast cancer.8

The assessment of test results in this study was simplified to positive or negative. This dichotomization reflects the literature on the sensitivity and specificity of testing.5,6 However, in clinical medicine, these tests often present a range of descriptions for a positive result from mild positives, such as well-circumscribed density on a mammogram, to a strongly positive result, such as inducible ischemia on a stress test or spiculated mass on a mammogram. A more strongly positive or abnormal result would be less sensitive but more specific for disease. This study did not evaluate interpretation of more complex test results.

There are important implications of the finding of a gap between practitioner estimates and scientific estimates of the probability of disease. Practitioners who overestimate the probability of disease would be expected to use that overestimation when deciding whether to initiate therapy, which could lead to overuse of medications and procedures with associated patient harms. Practitioners in the study reported that they would initiate treatment based on estimates of disease, including 78.2% who would treat cardiac ischemia and 71.0% who would treat a UTI when a positive test result would place their patient at 11% or less chance of disease. These errors would similarly corrupt shared decision-making with patients, which relies on practitioner understanding and communication of the likelihood of various outcomes.25-27 Training in shared decision-making has focused on communication skills,28 not on understanding the probability of disease,29 but the findings suggest another important educational target.

More focus on diagnostic reasoning in medical education is important. The finding of a primary problem with pretest probability estimates may be more amenable to intervention than the more commonly discussed bayesian adjustment to probability from test results.30 Pretest probability is commonly discussed in medical education, but a standard method for estimating pretest probability has not been described.30 Ideally, such estimates incorporate knowledge of disease prevalence and the predictive value of components of the history and physical examination, but for many conditions this information is difficult to find. The fact that estimates are so far from scientific evidence identifies a pressing need for improvement. There are a limited number of well-characterized diseases with pretest probability calculators, notably cardiac ischemia.31,32 Despite the fact that respondents in this study had no access to external aids while completing the survey, pretest estimates of cardiac ischemia were more accurate than for other clinical scenarios, implying that access to these calculators may improve knowledge and impact clinical reasoning. There is also a need to improve bayesian adjustment in probability from test results, which requires readily accessible references for clinical sensitivity and specificity. Computer visual decision aids that guide estimates of probability may also have a role.5,33 Alternative approaches, such as natural frequencies and naturalistic decision-making or use of heuristics, may improve decisions.34

Limitations

This study has limitations. One is that the small fraction of respondents who did not complete the survey were more likely to be female, nurse practitioners, or physician assistants or to have been in practice for more than 10 years. However, the overall response rate was high. The format of survey questions required participants to estimate pretest probability before giving interpretation of positive or negative test results, which may not reflect their natural practice. Finally, although validity was extensively assessed via a multidisciplinary expert panel, reliability of our novel survey was not assessed.

Strikingly, compared with younger adults, older people were more willing to put in effort for others and exerted equal force for themselves and others

Aging Increases Prosocial Motivation for Effort. Patricia L. Lockwood et al. Psychological Science, April 16, 2021. https://doi.org/10.1177/0956797620975781

Abstract: Social cohesion relies on prosociality in increasingly aging populations. Helping other people requires effort, yet how willing people are to exert effort to benefit themselves and others, and whether such behaviors shift across the life span, is poorly understood. Using computational modeling, we tested the willingness of 95 younger adults (18–36 years old) and 92 older adults (55–84 years old) to put physical effort into self- and other-benefiting acts. Participants chose whether to work and exert force (30%–70% of maximum grip strength) for rewards (2–10 credits) accrued for themselves or, prosocially, for another. Younger adults were somewhat selfish, choosing to work more at higher effort levels for themselves, and exerted less force in prosocial work. Strikingly, compared with younger adults, older people were more willing to put in effort for others and exerted equal force for themselves and others. Increased prosociality in older people has important implications for human behavior and societal structure.

Keywords: prosocial behavior, aging, effort, motivation, reward, computational modeling, open data

Many prosocial behaviors require the motivation to exert effort. Here, we showed that older people, compared with younger people, are more prosocially motivated in two crucial aspects of behavior. First, computational modeling and mixed-effects models show that older adults discount rewards by effort less when benefiting others, and thus they are more willing to choose highly effortful prosocial acts. Second, whereas younger adults show a self-bias, pursuing highly effortful actions that benefited themselves more than others, older adults do not. Thus, greater prosociality was demonstrated not only in older adults’ decisions but also in how much energy they allocated to self- and other-benefitting acts. Finally, we observed individual differences in the relationship between discounting in the two groups and their feelings of positivity at helping themselves and others. Positive feelings toward rewarding others were correlated with the willingness to put in effort for others in both younger and older adults, consistent with a maintained sense of “warm glow” across the life span, but only in younger adults did the willingness to put in effort for themselves correlate with how positive the rewards made them feel. Overall, we found, across several indices, that older adults are more prosocial than younger adults and have a lower self-favoring bias in their effort-based decision-making. Therefore, prosocial behavior could fundamentally shift across the life span.

Studies examining life-span changes in prosocial behavior have been mixed. Here, we showed that older adults might be more prosocial in social interactions than younger adults, as suggested by some studies using economic games (Sze et al., 2012). However, our approach was able to show that this effect is not because older adults value money differently per se, as the cost was not money but effort. Moreover, this effort cost was adjusted to each person’s capacity and was manipulated independently from reward in separate self and other conditions, so we were able to identify changes in sensitivity to a cost between a self-benefiting and a prosocial act. Importantly, both in choice behavior and in the energization of actions, there were significant differences between young and older adults’ sensitivity to the effort cost that differed between the self and other conditions. These findings highlight the necessity to examine effort and self- and other-oriented motivation independently, in order to understand specific life-span changes in prosocial behaviors. In addition, these results highlight the importance of comparing people’s willingness to put effort into different types of behavior and not treat motivation as a unidimensional construct. Indeed, some studies in the cognitive domain have found that older adults are more averse to effort than younger adults when it comes to cognitive effort (Hess & Ennis, 2012Westbrook et al., 2013) and also that cognitive and physical efforts are valued differently (Chong et al., 2017). Dissecting the different components of effort-based decision-making in various contexts will be crucial for accurately quantifying and unpacking the mechanisms underlying multiple facets of people’s motivation (Ang et al., 2017Cameron et al., 2019Chong et al., 2017Inzlicht & Hutcherson, 2017Kool & Botvinick, 2018Lockwood, Ang, et al., 2017).

Why might older adults be more prosocial when deciding to put in effort and energize their actions? There are several possible explanations both at the biological and sociocultural level. Socioemotional-selectivity theory posits that as people grow older, their time horizon shrinks, leading to changes in motivational goals and shifts in priority driven by changes in emotional needs (Beadle et al., 2013Carstensen, 2006). Evidence in support of this is provided by the observation that antisocial and aggressive behaviors significantly decrease across the life span. Young adults (16–24 years old) have the highest rates of homicide (Office for National Statistics, 2019), and several studies have suggested that criminal activity increases during adolescence and declines in older adulthood (Liberman, 2008). As levels of antisocial behavior and criminality lessen across the life span, it is plausible that such changes would, in parallel, be associated with increased prosociality. However, we did not find much evidence that changes between age groups are linked to higher emotional reactivity. In both groups, how willing someone was to put in effort for another person was positively correlated with how positive they felt when winning points for the other person, and there were no significant difference in the strength of correlation. This would not be entirely consistent with a socioemotional-selectivity account, which would posit that there is a stronger prioritization of this emotional response in older adults. Intriguingly, these results do show that the warm glow linked to how much a person will help others is maintained across the life span, with the caveat that ratings of positivity might be susceptible to experimenter demand effects.

Such findings, as well as the reduced difference between participants’ motivation for themselves and others in both choices and force exerted, suggest that older adults may have lost an emotionally driven self-bias that could lead to their putting in more effort for others compared with themselves, relative to younger adults. There is considerable evidence that young adults show a self-bias in many aspects of cognition and behavior; they prioritize self-relevant over other-relevant information. This includes effort, as shown here, but also other factors. Young adults show a self-bias when learning which of their actions earn rewards for themselves and which arbitrary stimuli belong to them, and they also demonstrate bias in many forms of memory and attention (Lockwood et al., 20162018). Existing studies of changes in self-bias with increased age have been somewhat mixed. One study found an increased emotional-egocentricity bias in older adults (Riva et al., 2016), as measured by the incongruency of self and other emotional states. A study that employed an associative-matching task suggested a reduced self-bias in older compared with younger adults (Sui & Humphreys, 2017). Here, by independently manipulating costs and benefits on self and other trials, we found that when it comes to motivation to exert effort, older adults become less self-biased. Future work should begin to distinguish what aspects of the self-bias increase and which decline.

In this study, we specifically focused on willingness to exert physical effort that benefits others—effort that may relate to everyday real-world prosocial acts. Prosocial acts also include behaviors such as doing charitable work or donating money to charity. However, volunteer work can be affected by the amount of time people have available to sacrifice, and monetary donations depend on wealth; both are key issues in aging research on prosocial behavior (Mayr & Freund, 2020). In our task, one major strength was that putting in effort to give rewards to other people had no impact whatsoever on the participant’s own payment at the end. Nevertheless, in future studies, researchers could try to link prosocial effort to everyday prosocial acts, perhaps through measures such as experience sampling, to translate these findings outside the lab. Moreover, researchers could include a measure of perceived wealth to see whether any differences explain variance in how much participants value the monetary rewards on offer. It would also be intriguing to link willingness to exert effort to measures that may quantify social isolation in older adults, such as their social-network size, to examine whether those adults who choose to put in more effort to help others have larger or smaller social networks than younger adults.

Willingness to be prosocial can be affected by social norms such as reciprocity and acceptance (Gintis et al., 2003). We specifically designed our study to minimize these effects: Participants never met face to face, and they were told that they would leave the building at different times and that their identities would never be revealed. However, it could be that social norms are internalized differently across different ages and cultures. It would be interesting for researchers to try to manipulate different social norms in future studies to examine the effect on prosocial choice and force exerted. A strength of the task is that both people’s explicit choices and their implicit energization of action can be measured to provide complimentary insights into prosocial motivation. It would also be important for researchers to examine whether the nature of the receiver changes people’s prosociality, depending perhaps on their age, their closeness, or whether they are perceived as part of an in-group or an out-group. Researchers could also examine whether possible increases in empathy between age groups are linked to differences in willingness to help others: Previous research has suggested that older adults have greater empathic concern for people in need compared with younger adults, although they do not show a benefit from imagining helping others in the same way as younger adults (Sawczak et al., 2019). That also dovetails with research showing an important link between empathy and motivation (Cameron et al., 2019Lockwood, Ang, et al., 2017). Finally, we note that our results are from a single, albeit well-powered, study, and researchers should seek to replicate our effects in future work.

Overall, we showed that older adults are more prosocial than younger adults in two core components of motivation. Moreover, different emotional considerations may drive decisions in younger and older adults to invest effort to help themselves and others. Understanding the trajectory of social behavior across the life span can inform theoretical accounts of the nature of human prosociality as well as theories of healthy aging—and ultimately, in the long term, help to develop strategies for scaffolding lifelong health and well-being.

Friday, April 16, 2021

Mindfulness Decreases Prosocial Behavior for Those with Independent Self-construals

Poulin, Michael, Lauren Ministero, Shira Gabriel, Carrie Morrison, and Esha Naidu. 2021. “Minding Your Own Business? Mindfulness Decreases Prosocial Behavior for Those with Independent Self-construals.” PsyArXiv. April 9. doi:10.31234/osf.io/xhyua

Abstract: Mindfulness appears to promote individual well-being, but its interpersonal effects are less clear. Two studies in adult populations tested whether the effects of mindfulness on prosocial behavior differ by self-construals. In Study 1 (N = 366), a brief mindfulness induction, compared to a meditation control, led to decreased prosocial behavior among people with relatively independent self-construals, but had the opposite effect among those with relatively interdependent self-construals. In Study 2 (N = 325), a mindfulness induction led to decreased prosocial behavior among those primed with independence, but had the opposite effect among those primed with interdependence. The effects of mindfulness on prosocial behavior appear to depend on individuals' broader social goals. This may have implications for the increasing popularity of mindfulness training around the world.


Moral panic about fake news? It is problematic to establish causal relationships between fake news and the effects it has been said to produce

Misinformation about fake news: A systematic critical review of empirical studies on the phenomenon and its status as a ‘threat’. Fernando Miro-Llinares, Jesus C. Aguerri. European Journal of Criminology, April 15, 2021. https://doi.org/10.1177/1477370821994059

Abstract: After the 2016 US presidential elections, the term ‘fake news’ became synonymous with disinformation and a catch-all term for the problems that social networks were bringing to communication. Four years later, there are dozens of empirical studies that have attempted to describe and analyse an issue that, despite still being in the process of definition, has been identified as one of the key COVID-19 cyberthreats by Interpol, is considered a threat to democracy by many states and supranational institutions and, as a consequence, is subject to regulation or even criminalization. These legislative and criminal policy interventions form part of the first stage in the construction of a moral panic that may lead to the restriction of freedom of expression and information. By analysing empirical research that attempts to measure the extent of the issue and its impact, the present article aims to provide critical reflection on the process of constructing fake news as a threat. Via a systematic review of the literature, we observe, firstly, that the concept of fake news used in empirical research is limited and should be refocused because it has not been constructed according to scientific criteria and can fail to include relevant elements and actors, such as governments and traditional media. Secondly, the article analyses what is known scientifically about the extent, consumption and impact of fake news and argues that it is problematic to establish causal relationships between the issue and the effects it has been said to produce. This conclusion requires us to conduct further research and to reconsider the position of fake news as a threat as well as the resulting regulation and criminalization.

Keywords: Criminalization, fake news, misinformation, social networks, threat


Life satisfaction predicted by different "recipes" (or sets) of personal values between 5 regions of the world in massive study of over 100,000 people; community, voluntary, satisfaction with finances, and exercise were common

Alternative Recipes for Life Satisfaction: Evidence from Five World Regions. Bruce Headey, Gisela Trommsdorff & Gert G. Wagner. Applied Research in Quality of Life, Mar 25 2021. https://link.springer.com/article/10.1007/s11482-021-09937-3

Abstract: In most cross-national research on Life Satisfaction (LS) an implicit assumption appears to be that the correlates of LS are the same the world over; ‘one size fits all’. Using data from the World Values Survey (1999–2014), we question this assumption by assessing the effects of differing personal values/life priorities on LS in five world regions: the West, Latin America, the Asian-Confucian region, ex-Communist Eastern Europe, and the Communist countries of China and Vietnam. We indicate that differing values - traditional family values, friendship and leisure values, materialistic values, political values, prosocial and environmental values, and religious values – are endorsed to varying degrees in different parts of the world, and vary in whether they have positive or negative effects on LS. Personal values provide the basis for alternative ‘recipes’ affecting LS. By ‘recipes’ we mean linked set of values, attitudes, behavioural choices and domain satisfactions that have a positive or negative effect on LS. We estimate structural equation models which indicate that differing values-based recipes help to account for large, unexpected differences between mean levels of LS in the five world regions, compared with the levels ‘predicted’ by GDP per capita. In particular, the high priority given to traditional family and religious recipes in Latin America helps to account for unexpectedly high LS in that region. Deficits in prosocial attitudes and behaviours partly account for low LS in ex-Communist Eastern Europe.


Traditional Family Values

It is well known that married/partnered people are on average happier than unmarried/unpartnered people, and that a cohesive family and satisfaction with family life are closely related to high LS (Diener et al. 1999; Argyle 2001). It is a fairly obvious next step to show that strong commitment to family values is linked to above average LS (Inglehart et al. 2008; Schwarz 2012; Headey and Wagner 2018, 2019).


Friendship and Leisure Values

A well established finding in LS research is that people with good social networks and high levels of social interaction/participation in activities with friends and acquaintances are happier than average (Bradburn 1969; Diener et al. 1999; Argyle 2001; Headey et al. 2010a). In this paper we extend this line of inquiry by investigating links between endorsing friendship and leisure values, related attitudes and choices, and LS.

Materialistic Values

Diener and Seligman (2002) and Nickerson et al. (2003) reported that individuals who prioritise materialistic values - financial and career success - are less happy than their less materialistic countrymen/women. We replicated their results, analysing Australian, British and German panel data (Headey 2008; Headey et al. 2010b; Headey and Wagner 2018, 2019). We also found that materialists are less rather more satisfied than average with their income and financial situation. Ng and Diener (2014) reported that in low income countries people place high priority on material goals, whereas in high income countries material and non-material goals are about equally prioritised.

Political, Prosocial and Environmental Values1

Dunn et al. (2008), analysing experimental data, showed that prosocial, altruistic people who spent money that had been donated to them on other people, rather than themselves, gained greater satisfaction from their expenditure (see also Aknin et al. 2019).

Studies of volunteering – a clear form of prosocial behaviour – have shown that volunteers have above average levels of LS (Harlow and Cantor 1996; Thoits and Hewitt 2001). Our previous papers, based on panel data, have confirmed that people who prioritise prosocial values record well above average LS (Headey 2008; Headey et al. 2010a; Headey and Wagner 2018, 2019).

Religious Values

There has been extensive investigation of the hypothesis that religious people are more satisfied with life than non-religious people (Koenig and McCullogh 1998; Friedman and Martin 2011; Headey et al. 2010b). The evidence is not unambivalent, but on balance most studies show that the devoutly religious, especially if they attend church (mosque, synagogue etc) regularly, are more satisfied than average, and also live longer (Koenig and McCullogh 1998; Friedman and Martin 2011; Headey et al. 2014). The relationship between LS and longevity is almost certainly partly due to commitment to traditional family values, and also to a relatively healthy lifestyle with below average rates of smoking and alcohol consumption (Friedman and Martin 2011).


No Clear Values/Life Priorities

Emmons (1986, 1988, 1992) found that individuals who give relatively low ratings to all values have low LS. He inferred that just having values promotes LS by giving people a sense of purpose. Diener and Fujita (1995) investigated links between values/life goals and resources, finding that people have higher LS if they prioritise values/goals for which they have appropriate resources.


Japan: Narcissism encourages guilt and therefore inhibits lying behavior

Daiku Y, Serota KB, Levine TR (2021) A few prolific liars in Japan: Replication and the effects of Dark Triad personality traits. PLoS ONE 16(4): e0249815. https://doi.org/10.1371/journal.pone.0249815

Abstract: Truth-Default Theory (TDT) predicts that across countries and cultures, a few people tell most of the lies, while a majority of people lie less frequently than average. This prediction, referred to as “a few prolific liars,” is tested in Japan. The study further investigated the extent to which the Dark Triad personality traits predict the frequency of lying. University students (N = 305) reported how many times they lied in the past 24 hours and answered personality questions. Results indicate that the few prolific liars pattern is evident in Japan thereby advancing TDT. Results also show that Japanese frequent liars tend to have Dark Triad personality traits, but the nature of the findings may be unique to Japan. Results of the generalized linear model suggest that the Dark Triad components of Machiavellianism and psychopathy exacerbate lying behavior by reducing the guilt associated with lying. However, narcissism encourages guilt and therefore inhibits lying behavior with both direct and indirect effects. These narcissism findings appear to contradict prior studies but stem from use of a more appropriate statistical analysis or the Japanese context.

Discussion

The purpose of this study was to test the few prolific liars predictions in Japan and to examine these prolific liars’ personality traits. Consistent with TDT predictions, the results documented the existence of the few prolific liars pattern in the current sample of Japanese students. Moreover, the results demonstrate that people high in Machiavellianism and psychopathy reported more lying, mediated by lowering guilt, while people high in narcissism reported less lying through both direct and indirect paths. Although we cannot fully establish the causal relationships with only this study, the results suggest that people high in Machiavellianism or psychopathy may be inclined to tell more lies due to reduced feelings of guilt and that people high in narcissism may tell fewer lies due to increased guilt. The reverse causal order alternative is that the act of lying reduces guilt causing Machiavellianism scores to increase. While it is possible that people who lie frequently come to experience less guilt over time, and as a consequence, rate themselves as higher on Machiavellianism and psychopathy, this seems less plausible than personality being the antecedent.

Consistent with prior studies, the distribution of self-reported lies is extremely skewed, indicating the existence of a few prolific liars in our sample. The average lying frequency was similar to that reported by prior studies, such as DePaulo et al. [1], Murai [2], and Serota and Levine [8]. Most participants reported five or fewer lies in the past 24 hours and only a few people reported six or more lies. Importantly, prior results demonstrate that the few prolific liar phenomenon is not an artifact of the self-reporting methodology. Halevy et al. [11] showed that the self-reported number of lies correlates with behavioral indices of dishonesty in a laboratory and in our data, eliminating low-confidence participants does not change the overall finding. Therefore, the self-reported results appear to represent a reliable index and the universality of the “few prolific liars” module of TDT.

TDT seeks to provide a pan-cultural account of human deceptive communication. Because TDT predictions are not culturally bound, it is critical to test TDT in a variety of cultures. Only by testing TDT in various countries can the robust nature of TDT’s predictions be ascertained. Although TDT studies have previously been conducted in North America, South America, Europe, Asia, and the Middle East, this research is the first to test TDT in Japan. The current findings add to the cultural span of TDT by replicating effects documented elsewhere.

Investigating the personality traits of the prolific liars using GLM yielded a more complex outcome than prior results. These results showed that Machiavellianism and psychopathy are associated with more lying, similar to prior studies [1114]. This suggests the two effects are robust enough to endure more rigorous statistical analysis. In addition, this study revealed that the effects are mediated by reduced feeling of guilt. Those high on Machiavellianism and psychopathy are thought to have lower guilt than ordinal people do, and this lower inhibition contributes to telling more lies.

These results, that the few prolific liars are Machiavellian and psychopathic people, may shed light on the fundamental question, “why is the distribution so skewed?” from an evolutionary perspective. Previous research found that people who have Dark Triad personality traits take the fast life strategy characterized by short-term mating, selfishness, and other antisocial manifestations [1526] and that they account for only a small part of the entire population [27]. Considering these findings, one possible explanation for the skewed distribution of lying is that the few prolific liars are people who adopted the fast life strategy. In modern society, the traits are seen as undesirable because most people do not adopt this strategy [28] but prolific lying may help those who adopt the fast life strategy to survive and reproduce. This evolutionary system may be the reason why we see the few prolific liars across cultures. This hypothesis is speculative but warrants further investigation.

However, somewhat surprisingly, narcissism had a negative effect on the frequency of lying. That is, results show people high in narcissism tell fewer lies. This result is contradictory to prior studies, which may result from the choice of statistical analyses. Jonason et al. [14] calculated the correlation coefficients and partial regression coefficients, finding a slightly positive correlation between narcissism and the number of lies. Similarly, Zvi and Elaad [12] found a positive relationship between narcissism and lying behavior. However, without accounting for the extremely skewed distribution of lie frequency, calculating Pearson correlations may yield misleading results, especially Type I errors [29]. As this and prior studies [7811] indicated, approximately 40–60% of people asked about lying frequency report no lies during any specific 24-hour period. Therefore, the distribution for lying frequency will be positively skewed and substantial (Skewness > 1.0 is considered substantial; for the Japan data Skewness = 12.67, SE of Skewness = 0.14). This inclination is not only an extreme deviation from the assumption of normality, it is wholly unsuitable for calculating Pearson’s correlations, which assume linear relationships between two variables. In addition, just a few prolific liars might exorbitantly increase the correlation, as Pearson’s correlation is very sensitive to outliers. For these reasons, Pearson’s correlations with lie frequency may be unreliable when the skewed distribution is considered. Spearman’s rank correlation suppresses the effect of outliers.

Moreover, we found the negative effect for narcissism (i.e., narcissists tell fewer lies) when controlling Machiavellianism and psychopathy. While the zero-order correlations of narcissism include the effects of Machiavellianism and psychopathy, the result of the negative binomial regression partials out the effects of them when assessing the effect of narcissism. Thus, it may be safe to say that the negative coefficient of narcissism is the pure effect of narcissism on lying frequency. This may be the reason why we had the negative coefficient while we had a positive correlation between lying frequency and narcissism in Spearman’s rank correlation.

This negative effect of narcissism on lying is interpretable from three perspectives. The first is narcissism’s relative brightness. Narcissism is considered the least dark trait among the three [30]. Narcissism has weaker relationships with anti-social behavior [153132] and the ability to lie [33] than do either Machiavellianism or psychopathy. Considering these findings, perhaps it is not so surprising that narcissism had a different effect from Machiavellianism and psychopathy in our study. Narcissism is characterized by entitlement, superiority, and dominance [14]. The narcissist’s priority is keeping self-image positive, and frequent lying may hurt self-image. If so, it may be a reason why those higher on narcissism tell fewer lies.

The second consideration is lying types. Our study did not classify lying types, so all kinds of lies are included in the analysis. Narcissists are thought to tell lies mostly about themselves to make a good impression on others. In fact, Jonason et al. [14] revealed that narcissism had its strongest relationship with the number of self-gain lies. Future research might benefit by classifying lie types as well as motives to lie.

The third possibility is cultural differences. Narcissism scores may differ across countries. Foster et al. [34] found that narcissism was higher in an individualistic culture than in a collectivistic culture; the United States, especially, produced the highest levels of reported narcissism. According to their study, Japan’s narcissism is predicted to be lower than that of the United States. Moreover, Japan is thought to have a shame culture rather than a guilt culture [35], suggesting that in Japan, social behavior might be determined by feelings of shame rather than guilt. Replicating the current study in a western country could facilitate a comparative cultural analysis.

Further research on the subtypes of narcissism also might be useful for interpreting this result. Narcissism can be divided into vulnerable narcissism—associated with introversion, defensiveness, anxiety and vulnerability to life’s traumas—and grandiose narcissism—associated with extraversion, self-assurance, exhibitionism, and aggression [36]. Previous research has revealed that grandiose narcissism is more strongly related to unethical behaviors than vulnerable narcissism [16]. The Dark Triad Dirty Dozen, which we used in the current study, does not measure the two types separately. Consequently, there is a possibility that the DTDD is primarily measuring vulnerable narcissism and that this form of narcissism, which is associated with a positive self-image, is more likely to inhibit lying.

The current study has three limitations to consider. First, our analysis did not control for the frequency of social interaction. The Dark Triad personality traits are positively correlated with extraversion among the Big Five personality traits [13]. Thus, an alternative explanation for high lie frequency could be that prolific liars have more social interactions in a day rather than having an anti-social personality. However, studies that have controlled for frequency of interaction [137] found prolific liars even with a known rate of interaction. Future research may resolve this point by controlling for interaction rate.

Second, the results of this study are based solely on lies reported by college students. To improve the generalizability of the results, a study obtaining lie reports from a broader sample could be conducted. Fortunately, research in other countries is informative about how student samples are similar and different from more broadly representative samples. Research has documented the few prolific liars pattern (i.e., positive skew) in studies of both students and adult samples [7810]. The primary difference is that students tend to tell more lies on average. It is reasonable to expect that we would find a similarly skewed distribution among Japanese adults even though they may tell fewer lies, overall.

Third, the measurement of the Dark Triad used in this study may be insufficient. The Japanese version of the DTDD has differences from the original English version (e.g., lower reliability of psychopathy). The differences are most evident in Machiavellianism and psychopathy, but due to the strict translation procedures they are not substantial. It appears unlikely that the divergence for narcissism may have resulted from a translation problem.

Future research might examine other TDT propositions in Japan and other countries in Asia. Truth-bias has been documented in Korea [6] and Murai [2] found that Japanese participants reported (knowingly) receiving far few lies each day than they told. Both prior findings are consistent with TDT’s applicability in Asian countries. Future research might provide a more direct test of the truth-default using the method developed by Clare and Levine [5] thus investigating if thoughts of deception come to mind unprompted. Given known cultural differences (e.g., collectivism versus individualism; power distance), TDT’s predications regarding pan-cultural deception motives and the projected motive model also need to be tested across Asia.

Overall, this research clearly indicates the existence of a few prolific liars in a student sample in Japan. As observed in other parts of the world, most Japanese people tell few or no lies on a given day and a small number of people, prolific liars, tell the majority of lies. Additionally, the study found that lying frequency increased with higher Machiavellianism and psychopathy scores, and that these factors are mediated by feelings of guilt. Documenting the mediating effects of guilt expands our knowledge about lying and its prediction. This mediating effect suggests that people with certain personality traits such as Machiavellianism may feel less guilty about lying and consequently have fewer inhibitions about lying. Practically, it may be effective to activate people’s feelings of guilt to suppress lying in real world. We further observed an unexpected effect of narcissism, which inhibited lying frequency. How narcissism affects lying should be investigated further.