Exploring temperament and character traits in medical students; a new approach to increase the rural workforce

Abstract

Background: This study explored temperament and character traits in medical students to identify the possible predictive value of these traits to students with varying levels of intention to pursue rural medicine. This work is the precursor to a better understanding of personality traits associated with medical disciplines within specific environments such as rural medicine.

Aims: The long term aim is to increase the recruitment of students who are best suited, and choose to practice in rural locations.

Methods: Medical students (272) completed a demographic survey and the Temperament and Character Inventory (TCI–R 140) to identify levels of the seven basic dimensions of temperament and character. Multivariate statistics explored differences between students’ TCI levels based on gender, rural origin and level of intention to pursue rural medicine.

Results: Analysis showed only main effects and confirmed significant differences in certain TCI dimensions between students with a high compared to a low or medium intention to practice rural medicine and between males and females.

Conclusions: Preliminary findings suggest that certain temperament and character traits may be related to interest in rural medicine however the efficacy of assessing personality traits as an adjunct to medical school training and career counselling remains uncertain.

Introduction

Current medical student selection and admission procedures do not reliably or comprehensively select for all the desirable and predictive traits of a student best suited to practise medicine (Albanese et al. 2003; Benbassat & Baumal 2007; Groves et al. 2007; McManus & Powis 2007). Added to this, is the dilemma surrounding workforce shortages in rural and remote medicine and the difficulties in fostering recruitment and retention of doctors in these areas.

With pressure on medical schools to produce not only more doctors but supply them in geographical areas of need (Prideaux 2006), the call for selection procedures that tap into innate and modifiable characteristics as well as career interest becomes paramount. Current understanding is that the best predictors of taking up a rural career are rural origin plus early and repeated exposure to rural medicine (Dunbabin & Levitt 2003). However, the counselling and selection processes into medicine and the rural pathway in particular, is a line of investigation that has been neglected.

It may be timely to investigate the value of a standardised assessment that informs about personality traits to help counsel and train medical students best suited to particular disciplines and workplace locations. Studies have noted the importance of recognising personality traits in both the selection of students (Borges & Osman 2001; Ferguson et al. 2002; Knights & Kennedy 2006), in tailoring the curricula to better prepare them for their chosen discipline (Rabinowitz et al. 2001; Hays & Gupta 2003) and in providing guidance and career advice throughout undergraduate and postgraduate studies (Watmough et al. 2007). Although the assessment of personality traits is not generally part of this selection or counselling process it has been suggested that this might be considered (Ferguson et al. 2002; Lievens et al. 2002; Vaidya et al. 2004; Knights & Kennedy 2006).

Table

Practice points
Better understanding of traits dominant in a medical discipline could provide an adjunct to undergraduate career guidance. This study identified a different profile of temperament and character traits in students according to career intention. Future research may determine it appropriate to counsel students who have an interest in and portray traits best suited for a rural career.

Furthermore, no research to date has investigated the ‘rural doctor’, a profession that has been increasingly recognised as an entity in itself, comprising skills and traits different from urban generalist or family medicine practitioners (McManus et al. 1996; Ferguson et al. 2002; Lievens et al. 2002). The importance of personality traits in career choice has considerable support through studies that identify characteristics associated with medical specialty choices and doctors in all disciplines (McManus et al. 1996; Ferguson et al. 2002; Lievens et al. 2002; Vaidya et al. 2004; Stillwell et al. 2006). Our research question asks; do students who have a high intention of pursuing rural medicine as a chosen speciality possess a unique profile of temperament and character traits compared with students of low intention?

We specifically employed the Cloninger Temperament and Character Inventory [TCI] (Cloninger et al. 1994) because it provides a psychobiological model of personality that has the potential to provide insight into human personality at multiple levels of analysis. These include the genetics of personality, the neurobiological foundations of behaviour, behaviour development and the interaction of personality dimensions with development and environmental factors (Cloninger et al. 1993; Vaidya et al. 2004). Each temperament and character trait is multifaceted and high and low descriptors are summarised in Table 1.

Table 1.  Temperament and character descriptors

Temperament	High scorers	Low scorers
HA	Worrying & pessimistic	Relaxed & optimistic
Harm Avoidance	fearful & doubtful	bold & confident
	shy, fatigable	outgoing, vigorous
NS	Exploratory & curious	Indifferent, reflective
Novelty Seeking	impulsive, disorderly	frugal & detached
	extravagant & enthusiastic	orderly & regimented
RD	Sentimental & warm	Practical & cold
Reward Dependence	dedicated & attached	withdrawn & detached
	dependent	independent
PS	Industrious & diligent	Inactive & indolent
Persistence	hard-working	gives up easily
	ambitious & overachiever	modest & underachiever
	perseverant & perfectionist	quitting & pragmatist
Character
SD	Responsible & reliable	Blaming & unreliable
Self Directness	purposeful, self accepted	purposeless, self-striving
	resourceful & effective	inert & ineffective
	habits congruent with long term goals	habits congruent with short term goals
CO	Socially tolerant	Socially intolerant
Cooperativeness	empathic, helpful	critical, unhelpful
	compassionate & constructive	revengeful & destructive
	ethical & principled	opportunistic
ST	Wise & patient	Impatient
Self Transcendence	creative & self-forgetful	unimaginative & self-conscious
	united with universe	pride & lack of humility

Taken from Cloninger et al. 1994.

Temperament is defined as those components of personality that are mildly heritable, developmentally stable, emotion based and not influenced by socio-cultural learning (Cloninger et al. 1993). The four dimensions of temperament are described as Novelty Seeking: NS (exploratory impulsiveness versus stoic frugality), Harm Avoidance: HA (anxiety – proneness versus outgoing vigour and risk taking), Reward Dependence: RD (social attachment versus aloofness) and Persistence: PS (industry versus underachievement).

Character traits are a reflection of personal goals and values specified in terms of ‘subject-object’ relations, are moderately influenced by socio-cultural learning and mature progressively throughout life (Cloninger 2004). The three traits which represent the subject-object dichotomy are Self Directedness: SD i.e. self-concept (responsible versus blaming), Cooperativeness: CO i.e. concept of relationship with others (empathic versus insensitive) and Self-Transcendence: ST i.e. global concept of self and the world (idealistic versus practical).

The controversy around the use of non-cognitive measures in medical school selection is documented (Benbassat & Baumal 2007). Until longitudinal studies confirm their validity, their use as purely a selection tool should be treated with caution. Nevertheless, studies using the TCI (Kluger et al. 1999; Parker et al. 2003; Vaidya et al. 2004) suggest that each medical speciality has a unique profile and the assessment of personality should not be ignored in the counselling of medical students choosing a specialty or likewise in gaining further understanding of the personality traits dominant in a medical discipline (Borges & Savickas 2002; Joyce et al. 2007).

The ethos of any medical school admission process in conjunction with career counselling should be to provide students with the detailed and objective information, on every discipline, that they need to make an informed choice regarding their career pathway. Rural medicine is a discipline that is not well defined but recent work by the authors has begun to establish a temperament and character trait profile of rural doctors using the TCI (Eley et al. 2008).

We hypothesised that students who had a high intention to pursue rural medicine would embody a different profile of temperament and character traits compared to students with little or no intention of a rural career. This work is the precursor to identifying a profile of traits in medical students. In the future it may be appropriate to counsel students who not only have an interest in but portray the attributes best suited for rural practice with medical schools providing training to further nurture and develop these attributes.

Methods

Ethical approval was obtained from the appropriate Behavioural and Social Science Ethical Review Committee.

Participants

Participants were year one to year four medical students in a graduate entry Bachelor of Medicine Bachelor of Surgery (MBBS) degree from a medical school in Queensland, Australia. Initial invitations to participate were sent randomly (∼120 per year) by an email asking interested students to respond and request the research materials. Upon receipt of request, the study questionnaire, information sheet, consent form and a stamped addressed return envelope was posted to each student.

Materials

Research instruments comprised a two page demographic survey (2–3 minutes) and the TCI-R 140 (Cloninger et al. 1994) (20–30 minutes). The demographic survey included age, gender, marital status, year of study and rural or urban origin. Rural origin was defined as ‘spending a significant portion of your childhood or adolescent years in a rural or remote area’. There is no clear definition of what constitutes ‘rural origin’ and the literature indicates that identifying with a ‘sense of rural background’ is most appropriate (Somers et al. 2007). Therefore we did not impose a specific number of years spent rural to represent rural origin and worded the question as above to allow students’ to consider their own experience and identify as rural or urban.

Level of intent to pursue rural medicine as a career was assessed by the question, ‘What degree of intention do you have in practising medicine in a rural or remote area sometime in the future’, with a three point scale of 1 = Low (little or no intent), 2 = Medium and 3 = High (definite intent). This question has been used extensively as a measure of rural intention (Eley & Baker 2007). Measuring strength of intention in this way i.e. directed toward a single target allows for an obvious choice and is supported in the literature (Somers et al 2007).

The TCI-R 140 is the 140 question (five point Likert scale) version of the self-report questionnaire designed to assess three character and four temperament dimensions of personality. The seven dimensions are described in Table 1. Raw data from the TCI were scored externally by Washington University, St Louis, USA and individual participant dimension scores were returned. These scores and all corresponding demographic data were entered into SPSS (version 14) for analysis.

Statistical analysis

Descriptive statistics were used to summarise the data. All analyses used α = 0.05 with an accompanying 95% confidence level for measuring significant differences between variables. T-test was used to determine significant differences between students with regard to their TCI, based on gender, and rural background. ANOVA and two-way ANOVA was used to make multiple comparisons between TCI scores among students based on their level of intention, gender, year of study, age and marital status.

Results

Response rate

A total of 476 invitations were distributed and 272 requested and returned the questionnaires giving an overall response rate of 57%. Of those students who reported their year of study (n = 191), Year 3 comprised 49%. This over representation is due in part to the nature of the Year 3 curriculum where students rotate throughout the year in eight week clinical rotations. However, no demographic differences were detected between students according to their year of study or between those who responded early (within three weeks of receiving the questionnaire) or late (after a follow-up reminder).

Demographics

Table 2 summarises the demographic profile of our sample. A higher proportion (n = 177; 68%) of students were of urban origin compared to 32% (n = 82) who reported a rural origin (x² = 14.03. p < 0.001). Level of intention was positively correlated (r = 0.209; p < 0.01) with rural origin. Of those students who reported a rural origin (n = 82), 51 (62%) were in the high intention group compared to 15 (18%) and 16 (19%) in the low and medium groups respectively. However 66 students (38%) who identified with an urban origin (n = 177) also reported a high rural intention.

Table 2.  Distribution of respondents according their demographic characteristics

Demographic characteristics	Number	Percent
Year of MBBS:
First	37	19
Second	32	17
Third	93	49
Fourth	29	15
Total	191	100
Gender:
Male	109	40
Female	163	60
Total	272	100
Age:
20–25 years	126	48
26–30 years	91	34
Over 30 years	48	18
Total	265	100
Marital status:
Married	32	12
Single	176	65
Partnered	64	23
Total	272	100
Level of intent in rural medicine:
Low intent	70	26
Medium intent	73	27
High intent	129	47
Total	272	100
Rural background:
Yes	82	32
No	177	68
Total	259	100

TCI scores

The internal consistency (Cronbach alpha) of our sample ranged from 0.84 to 0.88 for the character and from 0.76 to 0.89 for the temperament dimensions. Tests of normality (Kolmogorov–Smirnov statistic and Normal Q-Q Plots) showed the TCI scores for this sample to be normally distributed.

Table 3 shows TCI scores based on level of intention to pursue rural medicine, gender and student origin. Post hoc analysis (Scheffe) showed that students who reported a high intention had lower levels of HA compared with low and medium intention students, higher levels of SD compared with low and medium intention students and higher levels of CO compared with low intention students. Female medical students were higher in HA, RD and CO compared with male students. No significant differences were detected between levels of TCI dimensions and students who reported a rural or urban origin.

Table 3.  Comparison of student TCI dimensions by origin, gender and level of intent to practise rural medicine

		NS Novelty seeking			HA Harm avoidance			RD Reward dependence			PS Persistence			SD Self directedness			CO Cooperativeness			ST Self transcendence
Origin	N	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value
Rural	82	55.43	9.52	0.02	53.13	11.47	0.46	67.87	9.12	1.70	72.84	10.01	0.89	77.52	9.99	0.85	81.48	8.43	0.20	42.06	12.36	0.82
Urban	176	55.45	8.20		53.87	12.00		70.13	10.36		73.97	9.17		76.33	10.15		81.24	8.61		43.31	10.88
Gender	N	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value	M	SD	T value
Male	109	56.42	7.19	1.41	50.64	11.71	3.30**	66.12	9.48	4.45**	72.83	10.05	−1.20	76.27	10.36	0.79	79.21	8.89	3.47**	42.83	12.33	−0.15
Female	162	55.04		8.31	55.36	11.43		71.46	9.84		74.26	9.24		77.25	9.76		82.80	7.95		43.05	10.77
Intent level	N	M	SD	F value	M	SD	F value	M	SD	F value	M	SD	F value	M	SD	F value	M	SD	F value	M	SD	F value
Low	70	55.90	6.91	0.38	56.07	12.21	5.27*	68.53	10.61	0.74	73.30	9.36	0.94	74.19	10.24	7.60**	79.20	8.28	5.33**	41.23	10.04	2.78
Medium	73	55.70	9.04		55.15	12.31		68.70	9.88		72.64	10.81		74.71	11.12		80.47	8.64		41.63	12.14
High	128	54.90	9.16		51.08	10.75		70.09	9.81		74.49	8.93		79.17†	9.11		83.12††	8.27		44.67	11.38

M = mean, SD = Standard deviation; *Significant at p ≤ 0.01; **Significant at p ≤ 0.001; †High intention students different from low and medium intention students; ††High intention students different from low intention students.

Two-way between groups ANOVA with post-hoc comparisons (Bonferonni and Tukey's HSD) were conducted to explore the impact of gender and intention on levels of each TCI dimension. There were no significant interaction effects detected.

Harm Avoidance (HA)

There was a significant main effect for gender [F (1, 264) = 10.75, p > 0.001], partial η² = 0.040 and intention [F (2, 264) = 6.38, p > 0.002], partial η² = 0.048 on HA. Post-hoc comparisons indicated that the mean score for the high intention group (M = 51.08, SD = 10.7) was significantly lower than the low and medium groups.

Reward Dependence (RD)

There was a main effect for gender [F (1, 264) = 17.55, p > 0.001] and a moderate effect size partial η² = 0.062 on RD.

Self Directedness (SD)

There was a main effect for level of intention [F (2, 264) = 7.57, p > 0.001], partial η² = 0.054 on SD. Post-hoc comparisons indicated the mean score for the high intention group (M = 79.17, SD = 9.13) was significantly higher than the low and medium groups.

Cooperativeness (CO)

There was main effect for gender [F (1, 264) = 13.39, p > 0.001], partial η² = 0.048 and intention [F (2, 264) = 6.41, p > 0.002], partial η² = 0.046 on CO. Post-hoc comparisons indicated that the mean score for the high intention group (M = 83.12, SD = 8.27) was significantly higher than the low and medium groups.

Comparison of our sample's overall TCI scores with Cloninger's (1994) normative data found no significant differences. No other significant differences were detected in TCI scores based on year of MBBS, age or marital status.

Discussion

Our findings concur with others suggesting that temperament and character may be associated with medical specialty choice (Kluger et al. 1999; Parker et al 2003; Vaidya et al 2004). Our hypothesis that students with a high intention to pursue rural medicine would have a different temperament and character profile compared with students with little or no intention was supported. Analysis confirmed significant differences in one temperament trait, notably HA and two character traits, SD and CO between students with a high intention compared with medium or low intention.

There are several limitations to this study. A full and accurate profile of personality is beyond the scope of this paper and our methods make no attempt to provide this. Furthermore, personality trait testing is highly controversial. It implies the existence of stable characteristics which are measurable, when it is possible that behaviour is context specific and not related to stable traits (Eva 2003). Our study represents a small sample of medical students from one university and our findings may not generalise to other schools or students. Stratification of students by level of intention to pursue rural medicine is based on current interest and may not stand up over time although the scale employed has been used extensively. Additionally there is a recognised potential for bias in our self-selected sampling procedure i.e. students with a greater interest in rural medicine may have opted to participate – in particular because one of the character dimensions we measure is Cooperativeness (CO). Additionally, the length of the questionnaire and the time to complete (by busy students) may have contributed to non-response. These limitations confirm this research as a precursor to future studies.

The profile of high intention students (with significantly higher levels of SD and CO) in comparison with low and medium intention students indicates they are likely to be more effective, goal-oriented, responsible, and have a greater ability to adapt their behaviour in accordance with their goals (SD). High SD is also thought to have adaptive advantages when a person is confronted with hardship and suffering. Higher Cooperativeness (CO) suggests they may also be more empathic, supportive, fair and principled.

High intention students in comparison with low and medium students were significantly lower in Harm Avoidance (HA). Persons low in HA are thought to portray greater adaptability, confidence in the face of uncertainty and optimism in situations that would worry most people. Lack of resources, location constraints and little or no back-up or opportunity for consulting is common in rural/remote medical practice and would involve a high degree of uncertainty, decision making and adaptability. Low harm avoidance in high rural intention students may indicate a willingness to accept these risks and a level of confidence that would support them in uncertain and stressful situations. Presumably this would be an advantage in the personality profile of rural and remote doctors. Preliminary work using the TCI (Eley et al., in press 2008) indicates that temperament trait differences between a sample of rural and urban doctors include significantly lower levels of HA in the rural sample.

The gender differences we detected in our sample confirms other research (Cloninger et al. 1994; Vaidya et al. 2004) that female medical students have higher levels of RD (more sensitive, dedicated, dependent, and sociable), and CO compared with their male counterparts. The fact that a slightly higher proportion of respondents were female may be another example of this. Likewise female students were higher in HA indicating they may be more cautious, careful and intolerant of uncertainty. Analysis found only main effects and although effect sizes were small to moderate, this suggests that differences in HA and CO between intention levels are probably not an artefact of an interaction between gender and intention.

As expected rural background was correlated with high intention but interestingly 38% of students who identified as urban origin also reported a high rural intention. These findings are important for two reasons. The first supports previous research suggesting rural origin students are twice as likely to end up in rural practice (Dunbabin & Levitt 2003). The second suggests it might be beneficial for medical schools to tailor their curricula to cultivate interest in rural medicine in non-rural students. This indicates that the provision of rural training and exposure to urban students, who may also possess the character and temperament suited for rural practice, is tantamount to complimentary strategies such as preferred entry for rural origin students and the Rural Clinical Schools initiative (DOHA 2002).

Our findings corroborate other results (Vaidya et al. 2004) that suggest the TCI may be useful in career counselling and helping students identify their interests. Identification of heritable/stable temperament traits and in particular modifiable/developmental character traits, could inform modifications to medical schools’ counselling and training processes. For example, students with very high HA may benefit from interventions to build confidence or have appropriate counselling regarding the demands of certain disciplines. Those with low HA and or high SD might be encouraged to pursue careers such as rural medicine that require a high degree of confidence in the face of uncertainty and self-directedness.

These preliminary findings imply that different profiles of specific traits can be identified in students according to rural career intention. With further investigation, this information could: (1) provide an adjunct to existing medical school career counselling procedures; (2) tailor training for rural experiences to encourage cohorts of students better equipped to cope with and be retained in a rural environment; and (3) provide comprehensive career advice to students with an interest in a rural medicine by gaining a better understanding of the personality traits dominant in this discipline.

Further studies

Similar work on a larger scale has described the temperament and character profiles (i.e. configurations of traits) of urban and rural doctors. This data will allow further comparison of those traits in medical students and may in time prove generalisable to other disciplines within the rural workforce. Longitudinal tracking of students will eventually determine where they undertake post graduate training and eventually practise medicine. Until such time, the efficacy of cross-sectional studies assessing personality traits as an adjunct to medical school training, career counselling and ultimately selection remains uncertain.

Acknowledgements

This research was supported by a grant from the Australian Research Council (ARC) and the Central and Southern Queensland Training Consortium (CSQTC).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Additional information

Notes on contributors

Diann Eley

DIANN ELEY, MSc, PhD, Senior Research Fellow, Director of Research, Rural Clinical School, School of Medicine University of Queensland.

Louise Young

LOUISE YOUNG, MPsychEd, PhD, Senior Lecturer, Deputy Director Centre for Medical Education, School of Medicine University of Queensland.

Thomas R. Przybeck

THOMAS R PRZYBECK, PhD, Research Assistant Professor, Department of Psychiatry, Washington University School of Medicine.