Article Text

PDF

Stoic beliefs and health: development and preliminary validation of the Pathak-Wieten Stoicism Ideology Scale
  1. Elizabeth B Pathak1,
  2. Sarah E Wieten2,
  3. Christopher W Wheldon3
  1. 1Department of Internal Medicine, Morsani College of Medicine, University of South Florida, Tampa, Florida, USA
  2. 2Departments of Philosophy and Internal Medicine, University of South Florida, Tampa, Florida, USA
  3. 3Department of Community and Family Health, College of Public Health, University of South Florida, Tampa, Florida, USA
  1. Correspondence to Dr Elizabeth B Pathak; dr.elizabeth.pathak{at}gmail.com

Abstract

Introduction We developed and validated a new parsimonious scale to measure stoic beliefs. Key domains of stoicism are imperviousness to strong emotions, indifference to death, taciturnity and self-sufficiency. In the context of illness and disease, a personal ideology of stoicism may create an internal resistance to objective needs, which can lead to negative consequences. Stoicism has been linked to help-seeking delays, inadequate pain treatment, caregiver strain and suicide after economic stress.

Methods During 2013–2014, 390 adults aged 18+ years completed a brief anonymous paper questionnaire containing the preliminary 24-item Pathak-Wieten Stoicism Ideology Scale (PW-SIS). Confirmatory factor analysis (CFA) was used to test an a priori multidomain theoretical model. Content validity and response distributions were examined. Sociodemographic predictors of strong endorsement of stoicism were explored with logistic regression.

Results The final PW-SIS contains four conceptual domains and 12 items. CFA showed very good model fit: root mean square error of approximation (RMSEA)=0.05 (95% CI 0.04 to 0.07), goodness-of-fit index=0.96 and Tucker-Lewis Index=0.93. Cronbach’s alpha was 0.78 and ranged from 0.64 to 0.71 for the subscales. Content validity analysis showed a statistically significant trend, with respondents who reported trying to be a stoic ‘all of the time’ having the highest PW-SIS scores. Men were over two times as likely as women to fall into the top quartile of responses (OR=2.30, 95% CI 1.44 to 3.68, P<0.001). ORs showing stronger endorsement of stoicism by Hispanics, Blacks and biracial persons were not statistically significant.

Discussion The PW-SIS is a valid and theoretically coherent scale which is brief and practical for integration into a wide range of health behaviour and outcomes research studies.

  • stoicism
  • help-seeking
  • health behaviour theories
  • masculinity
  • patient perspectives
  • gender

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/

Statistics from Altmetric.com

Strengths and limitations of this study

  • The Pathak-Wieten Stoicism Ideology Scale (PW-SIS) is a new, theoretically coherent, multidimensional scale which measures stoic beliefs and sense of self along four domains: stoic taciturnity, stoic endurance, stoic serenity and stoic death indifference.

  • The PW-SIS contains 12 items and demonstrates good psychometric properties and content validity in a large sample (n=390) of educated adults.

  • Mean stoicism ideology scores were higher for men than women, but for both genders the most frequent scores were neutral on stoic ideology, and the response distributions by gender overlapped almost completely.

  • Further validation of the PW-SIS in demographically and socioeconomically diverse populations will improve its generalisability.

Introduction

Stoicism is a school of philosophy which originated in ancient Greece.1–3 Core elements in the classical definition of stoicism were an idealisation of imperviousness to strong emotions, and an indifference to death.3 Major Asian philosophical systems of thought, such as Buddhism and Confucianism, also endorsed stoic principles and teachings.4 5 Beginning in the 19th century, academic and popular philosophers in Europe and the Americas were exposed to and influenced by Asian philosophy and religion. Therefore, it may not always be possible to distinguish whether particular strands of contemporary thought associated with stoicism originated in ancient Greece, ancient India or elsewhere. For example, using very different language and symbolism, both the Greek Stoics and the Buddha exhorted the student to live fully and completely in the present, while minimising concern about the future.

Contemporary meanings and connotations of stoicism have expanded beyond their ancient origins, to include ideals of taciturnity and self-sufficiency.6–8 Today, the philosophical principles of stoicism can be seen to closely align with some personal ideologies, values and behaviours which are commonplace across many industrial nations, and are evident in many non-Western cultures as well.9–12 For example, in the USA, the armed forces have explicitly embraced stoic ideology as a tool for mitigating combat stress.13 14

Previous research on stoicism and health

Much of the previous health-related research which mentions stoicism has invoked the term as a descriptor of particular patient groups or behaviours, without an explicit theoretical context.8 Stoicism is mentioned most frequently in studies related to pain (particularly cancer pain) and coping strategies; indeed stoicism has been labelled a ‘coping strategy’ in more than one study.6–8 15 16 Stoicism has also been invoked as a defining characteristic of masculinity and as a key explanatory factor for certain health behaviours and outcomes among men. There are several psychometric instruments that measure endorsement or adherence to social norms of masculinity, but these scales include only a few items which explicitly assess stoicism.17–19

Direct measurement of stoicism in previous health-related measures has implicitly defined stoicism as a pattern of behaviours, not as an ideology. The pain attitudes questionnaire (PAQ), published in 2001, has a brief subset of questions focused on stoic responses to physical pain.20–22 The stoicism items in this scale were designed to capture pain coping strategies of chronically ill or injured patients. Of the 29 items in the PAQ, most measured past actions (ie, pattern of behaviour) and only 2 were explicitly focused on ideology: #2 ‘When I am in pain I should keep it to myself,’ and #24 ‘Pain is something that should be ignored.’ The 20-item Liverpool Stoicism Scale (LSS) was first developed in 19923 and has not been widely used.24–27 The LSS predominantly (16 of 20 items) assesses a single theoretical domain (stoic taciturnity) of the four validated theoretical domains included in the final Pathak-Wieten Stoicism Ideology Scale (PW-SIS) (table 1).

Table 1

Liverpool Stoicism Scale items and correspondence to Pathak-Wieten Stoicism Ideology Scale conceptual domains

Furthermore, the majority of items in the LSS focus on behaviour or conduct, for example, ‘I tend not to express my emotions.’ However, there are three items that are ideological, for example, “One should keep a ‘stiff upper lip’.” Both the LSS and the PAQ contain statements that are aphorisms (ie, ‘Pain is something that should be ignored’) or proverbs (ie, ‘A problem shared is a problem halved’). We consider these formats problematic, because these statements do not refer explicitly to the respondent. Consequently, agreement cannot be interpreted as a reflection of self-identity. Furthermore, aphorisms and proverbs may invite endorsement to a great extent simply because of familiarity. In fact, Yong et al found that item #24, ‘Pain is something that should be ignored,’ on the PAQ had a low alpha and reduced the internal consistency of their scale.21

Theoretical context

In 1983, Kathy Charmaz published a very influential sociological study on the ‘loss of self’ suffered by people with chronic illnesses.28 Although stoicism per se was mentioned only briefly, the idea that the suffering caused by disease emerges as much (or more) from threats to a person’s identity and sense of self as from purely bodily experiences of pathophysiology is one of the theoretical underpinnings of our work.

In this report, we attempt to articulate an explicit theory of stoicism and its potential impact on health. We theorise that stoicism is a system for self-regulation rather than a behaviour or personality trait. As a guide to ideal self-conduct, it requires self-conscious implementation and regular enforcement; in other words, stoicism is an ideology (eg, a belief system which informs one’s attitudes and actions with the inherent potential for internal resistance and conflict). Personal ideologies create expectations for people about who they are, as well as how they should and should not behave. For example, we theorise that people who strongly endorse a personal ideology of stoicism may be more likely to avoid or delay seeking professional medical intervention for serious signs and symptoms of disease. This personal ideology of self will not mandate behaviour in a deterministic fashion; rather, stoicism will create expectations of ideal behaviour (which may not always be met). In order to test these theoretical propositions in future research, a validated measure of an individual’s endorsement of stoic ideologies is needed.

The purpose of our study was to develop a theoretically coherent multidimensional scale to assess endorsement of a personal ideology of stoicism, and to empirically validate this scale in a multiethnic sample of healthy community-dwelling adults. We present the results of confirmatory factor analysis (CFA) of the multidomain PW-SIS, and discuss the potential usefulness of this tool for predicting constraints in health-related help-seeking behaviours. The PW-SIS is a generalised scale which assesses stoic beliefs and sense of self but does not explicitly measure health behaviours or health outcomes. Therefore, the PW-SIS can be used in a wide range of empirical research studies.

In addition, in this report we conducted an exploratory assessment of the association between high endorsement of stoicism and participant age, gender, and race and ethnicity. We expect stoic ideologies to be embedded in a larger system of cultural beliefs that may be related to age, gender, race and ethnicity, and other social characteristics.

Methods

Conceptual development of the stoicism ideology scale

Drawing on multiple scholarly and popular sources,1–3 6 15 29–31 we developed the preliminary 24-item Stoicism Ideology Scale (PW-SIS) to capture endorsement of five dimensions of stoicism (see online supplementary table 1 in the Technical Supplement). Based on our literature review and expert knowledge of philosophy, we defined each domain as follows:

Supplementary file 1

Stoic taciturnity is the belief that one should conceal one’s problems and emotions from others.

Stoic endurance is the belief that one should endure physical suffering without complaining.

Stoic composure is the belief that one should control one’s emotions and behaviour under stress.

Stoic serenity is the belief that one should refrain from experiencing strong emotions.

Stoic death indifference is the belief that one should not fear or avoid death.

Each item in our scale was carefully worded to capture the respondents’ ideology, not their past behaviour, using a 5-point Likert response scale with the following responses: ‘disagree’, ‘somewhat disagree’, ‘not sure’, ‘somewhat agree’ and ‘agree’. Nine of the original 24 items were ‘reverse’ items that specified antistoic beliefs, that is, ‘I believe I should experience strong emotions.’ The participant version of the scale (pen-and-paper questionnaire) listed response codes of 0 (disagree) through 4 (agree). These responses were recoded during analysis to range from −2 (disagree) to +2 (agree). Consequently, an average score of 0 corresponds to a neutral stance—neither endorsement nor rejection of stoicism. Positive scores indicate endorsement of a stoic ideology, while negative scores indicate rejection of a stoic ideology.

Data collection

Data were collected over a period of 10 months during 2013–2014. All participants were university employees or students. Written consent forms were waived by the IRB to ensure respondent anonymity but all participants provided verbal informed consent. Each participant completed a brief paper-and-pencil questionnaire consisting of the 24-item preliminary PW-SIS, sociodemographic questions and a final single item ‘I try to be a stoic’ with a 7-item response scale ranging from ‘never’ to ‘all the time’. The study population consisted of a convenience sample of 390 adults aged 18 years and older who were recruited in person by the authors in public common areas of university facilities (eg, cafeterias) using walk-up tables. Monetary incentives were not provided to participants. A study response rate could not be calculated due to the data collection methods.

Data analyses

Data analysis proceeded in five steps. During step 1, we examined univariate response distributions for each of the 24-scale items. A simple correlation matrix was examined to identify redundant items. Finally, we assessed content validity based on agreement with the statement ‘I try to be a stoic.’ As a result of step 1 analyses, six items were dropped from further analyses—including the entire stoic composure domain. Further details of this scale reduction step are included in the Technical Supplement (online supplementary table 2).

During step 2, we conducted a CFA of the reduced 18-item PW-SIS. CFA is the appropriate analytic choice to test scales that have an a priori, theoretically explicit subdomain structure.32–36 We used proc calis in SAS V.9.4 for the CFA. Based on the results of the first CFA, we eliminated two items with poor factor loadings (see online supplementary file 1, Technical Supplement, for details).

During step 3, we repeated the CFA on the reduced 16-item PW-SIS. Finally, for the purpose of parsimony we further reduced the total number of scale items to 12 (3 items in each of 4 domains) and conducted a CFA on the final 12-item version of the PW-SIS (step 4; see online supplementary table 3). Additional details and rationale for analytic steps 1–4 are included in online supplementary tables 1–3: Technical Supplement.

Step 5 of our analysis consisted of preliminary content validation, examination of response distributions for the overall and domain scores, and exploratory logistic regression modelling of sociodemographic predictors of strong endorsement of stoicism. For the logistic regression analysis, we categorised the outcome using the top quartile of the overall distribution of responses to represent strong endorsement of stoicism.

Results

The size of our study population (n=390) provided more than 15 respondents for each question in the preliminary scale, which exceeds the widely accepted norm of at least 10 respondents per question.37 Although skewed towards younger adults (78% of respondents were <25 years old), the study population was in other respects diverse (table 2). A majority self-identified as female (57%) and white (55%). Hispanics (15%) and Blacks (14%) were the second and third largest racial/ethnic groups, followed by Asians (9%) and biracial or other ethnicity (6%). A substantial minority of respondents (19%) were born outside the USA or Puerto Rico.

Table 2

Characteristics of the study population (n=390)

The final four-domain, 12-item PW-SIS is shown in table 3. CFA of the final scale showed very good model fit with individual item factor loadings ranging from 0.48 to 0.76, root mean square error of approximation (RMSEA) =0.05 (95% CI 0.04 to 0.07), goodness-of-fit index=0.96 and Tucker-Lewis Index=0.93.

Table 3

Pathak-Wieten Stoicism Ideology Scale

Relationships among the PW-SIS and its four conceptual domains are shown in table 4. Cronbach’s alpha ranged from 0.64 to 0.71 for the subscales and was 0.78 for the 12-item PW-SIS. Scores for stoic taciturnity were strongly correlated with scores for both stoic endurance and stoic serenity, but stoic endurance and stoic serenity were not highly correlated with each other. Stoic death indifference had the highest (most stoic) mean scores among the four domains, and it was least correlated with the other three domains.

Table 4

Conceptual domains of the Pathak-Wieten Stoicism Ideology Scale (PW-SIS)

Figure 1 depicts mean PW-SIS scores by response to the statement ‘I try to be a stoic.’ There was a clear monotonic and statistically significant trend, with respondents who reported trying to be a stoic ‘all of the time’ having the highest stoicism scores, and respondents who reported trying to be a stoic ‘never’ having the lowest stoicism scores. Most respondents chose one of the three intermediate categories. Respondents who chose ‘I don’t know’ as their response had stoicism scores similar to those who said they ‘sometimes’ tried to be a stoic.

Figure 1

Content validity of the Pathak-Wieten Stoicism Ideology Scale: mean scores by response to the statement ‘I try to be a stoic.’

The distributions of mean scores for the four conceptual domain subscales are shown in figure 2. Domain scores comprised the mean score for the three questions in the domain. In this study population, respondents were least likely to endorse stoic serenity and most likely to endorse stoic death indifference.

Figure 2

Pathak-Wieten Stoicism Ideology Scale: distribution of domain scores.

The full distribution of respondent scores is shown separately for women and men in figure 3. The distributions overlapped almost completely, but there were no men with the least stoic scores, and no women with the most stoic scores. Response distributions were skewed to the left for women (less stoic) and to the right for men (more stoic), consistent with a statistically significant difference in the mean scores for women (−0.31, 95% CI −0.40 to −0.22) and men (+0.04, 95% CI −0.05 to +0.14).

Figure 3

Pathak-Wieten Stoicism Ideology Scale: distribution of overall scores by gender.

Results of an exploratory analysis of sociodemographic predictors of high endorsement of stoicism are shown in table 5. There is no a priori cut point designated as ‘highly stoic’ in the PW-SIS; in this analysis, the cut point used was a mean score greater than the 75th percentile of the overall response distribution. The top quartile of the distribution of all respondents (n=390) ranged from +0.33 to +1.67. Among women, 18.9% strongly endorsed stoicism, compared with 32.8% of men. After multivariate adjustment, men were over two times as likely as women to fall into the top quartile of responses (OR=2.30, 95% CI 1.44 to 3.68, P<0.001). Adults born in the USA or Puerto Rico were also twice as likely as adults born elsewhere to strongly endorse stoicism (OR=1.97, 95% CI 1.01 to 3.84, P=0.048). ORs showing stronger endorsement of stoicism by Hispanics, Blacks, biracial persons and adults 25 years and older were not statistically significant.

Table 5

Sociodemographic predictors of a mean Pathak-Wieten Stoicism Ideology Scale score in the top quartile (>0.167)

Discussion

The PW-SIS is a theoretically coherent, multidimensional scale which demonstrates good psychometric properties and content validity based on initial validation in a large sample of educated adults. The PW-SIS is also brief and practical for integration into a wide range of empirical research studies. In our study population of mostly younger adults, endorsement of stoicism varied by conceptual domain, with the weakest endorsement of the classical domain stoic serenity (aversion to strong emotions). Exploratory logistic regression analysis identified male gender and US birth as significant predictors of strong endorsement of stoicism. Finally, point estimates suggested higher endorsement of stoicism for Blacks, Hispanics and biracial persons compared with Whites, but these results were not statistically significant.

Integration of our theory of a stoic ideology of the self into existing health behaviour models could help explain the formation of beliefs and attitudes towards criterion-specific help-seeking behaviours. Reasoned action approaches—such as the integrative model of behaviour prediction—poorly define background factors that underlie belief formation.38 Measurement of self-concepts, such as stoicism ideologies, may help explain this population variability. Expanding health behaviour theory to include aspects of the self could also help inform health education messaging and risk-based communication.

Ironically, a personal ideology of stoicism almost guarantees failure to live up to one’s personal ideal. Experiences of illness and disease often involve transient weakness and functional limitations. With ageing, these experiences will increase in frequency, duration and severity for most people. Simply put, experiences of illness and disease tend to require aid—whether from health professionals in a formal context, or from family members or friends in an informal context. An ideology of stoicism creates an internal resistance to external objective needs, which can lead to negative consequences.8–12

Gender and stoicism

Stoicism is widely viewed as a defining attribute of masculinity. Instruments designed to assess endorsement of hegemonic masculine ideologies have included specific questions that touch on stoicism. However, the conceptual and measurement overlap between these instruments and the four-domain PW-SIS is minor.17 For example, in the widely used Personal Attributes Questionnaire, only 2 of 24 items relate to a single domain of the PW-SIS. The Conformance to Masculine Norms scale assesses 11 distinct domains of masculinity, of which only 2 (emotional control and self-reliance) partially overlap with domains of the PW-SIS.18 19 In our study, the results are notable because for both genders the most frequent scores were in the middle of the distribution (neutral on stoic ideology), and the response distributions for women and men overlapped almost completely. Despite the fact that men were twice as likely as women to strongly endorse stoic ideology, our results suggest that gendered stereotypes about stoicism (‘stoic men’ and ‘emotional women’) are overblown. Because the PW-SIS is agnostic to respondents’ genders, it is ideally suited to investigate the empirical reality of stoicism among both women and men. Furthermore, our finding that a minority of women strongly endorsed stoic ideology may be particularly important. For example, a study of major strain among family caretakers of elderly patients with dementia found that those who used stoicism as a coping strategy suffered burnout, while those who sought social support did not.39

Study limitations

In any questionnaire-based scale, validity of the individual items and the total scale against the concept of interest is of paramount concern. Unlike many psychometric instruments, the PW-SIS does not purport to measure a latent, inherent trait such as personality, or a clinically definable disorder such as depression or anxiety. Rather, we attempt to measure an explicit set of beliefs, which by definition are neither inborn nor immutable. Therefore, a robust assessment of the content validity of our scale items must come after publication and evaluation by multiple experts and researchers. We included a single questionnaire item ‘I try to be a stoic’ to assess content validity, but future validation and outcome studies could expand on this approach or include a qualitative component.

A related question pertains to the predictive validity of the PW-SIS. In other words, to what extent does strong endorsement of stoic ideology predict actual stoic behaviours? Predictive validity can only be rigorously addressed through prospective study designs.

Our study population, similar to many scale validation studies, was university based. Therefore, validity and generalisability to very different populations should not be assumed, but instead tested in future studies. In particular, validation of the PW-SIS among the elderly and persons of lower educational attainment would be valuable for health-related research.

Strengths of the PW-SIS

Our scale has several strengths. First, all items refer explicitly to the respondent; there are no aphorisms or proverbs. Second, each item refers to an expectation or belief about ideal self-conduct, rather than to a simple description of past behaviour. So, for example, Q5 states ‘I expect myself to manage my physical discomfort without complaining’ rather than ‘I always manage my physical discomfort without complaining.’ This distinction is critical to the theoretical underpinnings of the scale. Third, we deliberately chose not to mention disease or illness in the scale items, so that the scale would be appropriate for a wide range of study populations, including currently healthy individuals. (Although some items do explicitly mention ‘physical pain’ and ‘everyday aches and pains’.) Our intention was to capture the respondents’ global endorsement of stoicism as a code of ideal conduct. Finally, the PW-SIS does not reference gender norms, so it can serve as a tool to empirically investigate gender differences in stoic ideology.

Directions for future research

The PW-SIS should be validated in multiple study populations with a range of socioeconomic and demographic characteristics. Our theory that ideologies of stoicism will result in constraints on health-related behaviours needs to be empirically tested, ideally in rigorously designed prospective studies. Given the rise of patient-centred healthcare,40 41 understanding patients’ motivations and perspectives has never been more important. The current health education paradigm holds that improving patients’ knowledge of symptoms and signs will result in more timely help-seeking behaviour.38 42–44 Each year, thousands of individuals suffer needlessly and many die because of extended delays in seeking professional aid for acute medical conditions (eg, myocardial infarctions, strokes, diabetic emergencies, cancer complications and pain, and acute exacerbations of congestive heart failure).45–52 Numerous studies have been conducted to attempt to elucidate the reasons behind patient delays,46–51 53 with the ultimate goal of designing education programmes and interventions that will result in timely help-seeking. Significant risk factors for help-seeking reluctance have been identified (eg, Black race52 54 55) but much of the variation remains unexplained and we still lack a complete understanding of why certain patients and not others delay seeking aid.

A distinction of our theory is movement of the focus of inquiry away from the disease and the patient’s relationship to the disease (eg, health knowledge, symptom awareness, ability to comply with self-care regimens) and onto patients’ sense of self—their self-concepts and self-identity.56 We hypothesise that illness behaviours may become ‘noncompliant’ or ‘irrational’ or ‘self-harming’ when specific courses of action would create an internal conflict with patients’ ideas of who they are. Specifically, we posit that people who strongly believe that they should manage their problems on their own, not show emotions, and not complain about physical discomfort will experience an internal cognitive conflict when faced with a situation that could require help from others. This internal conflict will lead to delays in or avoidance of help seeking, with potentially life-threatening consequences. For example, empirical studies of increasing rates of male suicide in rural Australia have identified hegemonic masculine norms of stoicism as an important causal factor in the context of severe economic stress.57 58 Understanding the influences of race, ethnicity, socioeconomic status, religion and other cultural factors on stoic ideologies may help explain past research findings on delays in help seeking. Finally, there may also be positive health consequences of stoic ideologies for individuals,15 which careful prospective research could confirm.

Acknowledgments

The authors are grateful to the respondents for their voluntary participation in this study. We would particularly like to thank the four peer reviewers of this manuscript, whose detailed and thoughtful readings resulted in a substantially improved final paper.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
View Abstract

Footnotes

  • Contributors EBP conceived the study. EBP and SW developed the preliminary PW-SIS. All authors enrolled participants and collected questionnaire data. CWW contributed statistical expertise to the confirmatory factor analysis. EBP analysed the data. All authors interpreted the results and outlined the paper. EBP drafted the manuscript. All authors contributed to literature review and substantive revisions of the paper.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Institutional Review Board of the University of South Florida.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.