The emotional intelligence of medical students: An exploratory cross-sectional study

Abstract

Background: Emotional intelligence (EI) may be related to student characteristics (such as conscientiousness and empathy), and performance at medical school, although few studies have so far been conducted.

Aim: To investigate the association of EI with students’ age, sex, ethnicity and stage of study at a London medical school.

Methods: All medical students were invited to complete an online EI instrument, the Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT) version 2, a 141-item measure of the ability to perceive, use, understand and manage emotions. An additional questionnaire to gather demographic data was linked to the MSCEIT.

Results: We analysed 263 responses from a population of 2114 medical students after three reminders (12.3% response rate). Aggregated EI scores were similar through the curriculum. Age, sex and ethnicity explained 9.2% of the variance in aggregated EI scores. In terms of managing emotions, 6.7% of the variance was explained by the stage of study, with significantly higher scores for students in their final year compared to those in the first two years.

Conclusion: This exploratory study provides preliminary data on EI scores for UK medical students identifies factors associated with higher and lower scores and suggests that aggregated EI scores remain stable during medical training.

Introduction

The term ‘emotional intelligence’ (EI) describes the ability ‘to monitor one's own and others’ feelings and emotions, to discriminate between them and to use the information to guide one's thinking and actions (Mayer et al. 1990). A recent definition includes the various ways in which emotions may be monitored, through the ability to perceive, use, facilitate and understand emotions, and to reflectively regulate emotions to promote emotional and intellectual growth (Mayer & Salovey 1997). Approaches to measuring EI have included self-report-, informant- and ability-based assessments (Conte 2005). The ability-based Mayer–Salovey–Caruso emotional intelligence test (MSCEIT, that we chose to use, demonstrates stronger discriminant validity than self-report measures which are more highly correlated than ability-based measures with personality constructs such as extroversion and neuroticism (Bracket & Mayer 2003).

Different methods have been used to study EI in a variety of educational settings. From a study of 373 Spanish undergraduate students working towards different degrees, using the perceived emotional intelligence (PEI) scale, a Spanish version of the trait meta-mood scale, perceived EI and high-perceived general self-efficacy were significantly linked (Duran et al. 2006). Those students with higher PEI scores were less threatened by examinations. A Belgian study conducted in a psychology undergraduate setting, using the trait emotional intelligence questionnaire (a self-report inventory), reported that trait EI significantly moderated the relationship between stress and self-reported health, where those with high EI appraised the examination period as less threatening. EI was found to predict better self-reported mental and physical health (Mikolajczak 2006). A study of secondary school students in England (Petrides & Furnham 2003) reported that trait EI was significantly related to scholastic achievement as measured by the trait EI questionnaire, key stage 3 assessments (KS3) and general certificate of secondary education (GCSE) marks. In two studies of Spanish high school students, both using a Spanish version of the MSCEIT, one reported that high EI was significantly associated with academic performance and pro-social behaviour (Marquez et al. 2006) and the other reported that strategic EI was significantly correlated with how many times female students were nominated as friends by their peers and with teacher ratings of male students’ academic behaviour and performance (Mestre et al. 2006). A study applying the MSCEIT to a sample of undergraduate business students in the United States found that EI was significantly related to public speaking effectiveness. Moreover, there was a high correlation between EI and conscientiousness and that these features interacted significantly with group behaviour and public speaking effectiveness, as well as academic performance (Rode et al. 2007).

Several authors report the impact of demographic variables on EI scores. In a study (using a Spanish version of the MSCEIT) of 946 college and high school students (Extremera et al. 2006), significantly higher EI scores were obtained by females and older respondents. This finding has been echoed in other studies (Mayer et al. 1999; Ciarrochi et al. 2000; Mayer et al. 2002; Palmer et al. 2005; Wong et al. 2007). Using the multifactor emotional intelligence scale (MEIS), a precursor of the MSCEIT, an American study of adults and adolescents found that EI ability levels increased significantly with age, and that females significantly outscored males (Mayer et al. 1999). A further study applying the MEIS to Australian psychology undergraduates (Ciarrochi et al. 2000) also reported that females significantly outscored male students, females tending to be better at perceiving emotions. In another Australian study, applying the MSCEIT to a sample of adolescents and adults drawn from the general population, females significantly outscored males on all measures (Palmer et al. 2005). The MSCEIT user manual (Mayer et al. 2002), presents evidence in support of younger respondents scoring significantly lower than their older counterparts in their ability to use, understand and manage emotions. In an exploratory study of undergraduate psychology and business management students in Singapore, Hong Kong and Taiwan, using Wong's emotional intelligence scale (WEIS), an EI measure developed for Chinese respondents, older age was found to be significantly related to higher EI scores (Wong et al. 2007). Few studies have investigated the relationship between ethnicity and EI. An American study of undergraduate psychology students using the emotional intelligence scale (EIS) found that Hispanic students significantly outscored their white counterparts (Van Rooy et al. 2005). A further study from the United States (Trinidad et al. 2004) found that Asian/Pacific Islander adolescents scored significantly higher on the MEIS than their white, Hispanic/Latino and multiethnic colleagues. Mayer et al. (2002) assert that EI scales may, to a certain extent, lack cross-cultural applicability and are developing, what they believe to be, a more sensitive measure of EI amongst ethnic minorities.

Although some information is available about the normative values of EI in student populations (Conte 2005), very little is known about EI in medical students. We have identified a small number of articles which have looked at EI and medical students. EI has been measured in first year medical students in Edinburgh (Austin et al. 2005), which reported a significant gender by cohort effect with male empathy scores increasing between years 1 and 2, and where as a whole, females scored significantly higher on all three measures, namely overall EI, empathy and the utilisation of emotions subscale. A recent review, containing some new data, from the Peninsula Medical School in UK and the University of Adelaide, has criticized the use of EI in a medical setting (Lewis et al. 2005), based on questions concerning the construct validity and psychometric properties of EI measures. A further study from the Peninsula Medical School focused on students’ views of the utility of this measurement (Lewis et al. 2004), reported that the majority of the students welcomed the opportunity to learn about the psychology of EI, and also valued identifying (using the emotional competency inventory) their competencies and areas for development. A study from Kentucky (Stratton et al. 2005) using the trait meta-mood scale and Davis’ interpersonal reactivity index, found empathic concern to be a significant predictor of students’ physical examination skills. Finally, a study from Ohio (Carrothers et al. 2000) using a semantic differential instrument for measuring medical students’ EI, reported that higher EI was related to being female and graduating from a university that values the social sciences and humanities.

The MSCEIT has recently emerged as the preferred measure (Spector 2005) and even a sceptical review of EI from Peninsula Medical School (Lewis et al. 2005) acknowledged that the MSCEIT possesses acceptable validity. Daus and Ashkenazy (2005), in their review article arguing the case for the MEIS/MSCEIT, report strong overall internal consistency reliability ranging from an r value of 0.90–0.96 with branch score reliabilities ranging from 0.76 for facilitating emotions to 0.98 for understanding and perceiving emotions. In preparation for undertaking further empirical work on factors associated with student progression, we decided to conduct a cross-sectional survey of EI among students in the King's College London School of Medicine at Guy's, King's and St Thomas’ Hospitals to obtain normative values in each year, to compare mean values in successive years and between various demographic groups, and to enable us to use these data as the basis for further work on EI and other factors likely to be associated with success and failure at medical school.

Table

Practice points

In medical education settings, EI has been shown to be moderately related to academic performance. Our preliminary findings suggest that females and older respondents tend to achieve higher EI scores and that aggregated EI scores remain stable throughout the undergraduate curriculum, although EI scores for the ability to manage and regulate emotions were higher in the final year of the course. Questions remain about whether EI has any relationship to clinical communication skills, quality of patient care or preservation of personal emotional well being. Future work in medical educational settings could evaluate the impact of an intervention programme as well as the longitudinal links between EI and academic performance in medical school, and the impact of interventions to develop EI. Further research is also required to develop a more user-friendly and less time-consuming version of the MSCEIT.

Methods

We used an online version of the MSCEIT, a 141-item instrument that measures the ability to perceive, facilitate, understand, and manage emotions in ourselves and others (Mayer et al. 2002). In this measure, perceiving (PEIQ) and facilitating (FEIQ) branches combine to form the experiential emotional intelligence (EEIQ) area, whilst the understanding (UEIQ) and managing emotions (MEIQ) branches combine to form the strategic emotional intelligence (SEIQ) area (Mayer et al. 2002). Table 1 provides a summary of these features and Table 2 outlines the psychometric properties of the MSCEIT. All four scores combine to form an aggregated EI score computed as an empirical percentile, where 100 is an average score with a standard deviation (SD) of 15.

Table 1.  Four-branch model of emotional intelligence

Area	Branch	Question types	How the ability may be used
Experiential (EEIQ): The degree to which one ‘takes in’ emotional experience, recognises it, compares it to other sensations, and understands how it interacts with thought	Perceiving emotions (PEIQ): Accurately identify emotions of people and elicited by objects	Identify emotions in faces, landscapes and designs	‘Read’ people's moods for feedback
Experiential area comprises perceiving and facilitating branches	Facilitating thought (FEIQ): Generate an emotion and solve problems with that emotion	How moods impact on thinking; relate feelings to thoughts	Create the right feeling to assist in problem solving, communicate a vision, and lead people
Strategic (SEIQ): The degree to which one can understand emotional meanings, their implications for relationships, and how to manage emotions in oneself and others	Understanding emotions (UEIQ): Understand the causes of emotions	Multiple choice emotion, vocabulary questions	Be able to predict how people will react emotionally
Strategic area comprises understanding and managing branches	Managing emotions (MEIQ): Stay open to emotions and integrate emotions with thinking	Indicates effectiveness of various solutions to problems	Integrate emotion and thought to make effective decisions

Source: Adapted from Mayer et al. (2002).

Table 2.  Psychometric properties of the MSCEIT

Study	Sample and setting	Findings
O’Connor and Little (2003)	90 psychology undergraduates at Grand Valley State University, USA	The MSCEIT correlated highly with indices of cognitive ability, but minimally with personality dimensions
Barchard (2003)	150 undergraduate students at the University of Nevada, USA	An overlap was found between EI measures and traditional measures of intelligence and personality
Bracket and Meyer (2003)	207 undergraduate psychology students at the University of New Hampshire, USA	The MSCEIT was discriminable from personality and well-being measures. Ability and self-report EI were weakly related and yielded different measurements of the same person
MacCann et al. (2004)	102 psychology undergraduates at the University of Sydney, Australia	Faces and designs (aspects of perceiving emotions branch in the MSCEIT) showed weak correlations with personality dimensions, crystallised intelligence and with visualisation abilities
Palmer et al. (2005)	431 adults from the general population in Victoria and New South Wales, Australia	There was a high level of convergence between the consensus and expert scoring methods. The reliability of the MSCEIT at total, area and branch levels was found to be good, although the reliability of most of the subscales was relatively low.
Marquez et al. (2006)	77 high school students in Cadiz, Spain	The MSCEIT was discriminable from well-established measures of personality (the big five personality traits, as well as a social competence inventory) and intelligence (factoral general intelligence)

All 2114 medical students were invited to complete the MSCEIT online via an institutional email circular, followed by three reminders. We linked a short questionnaire to the MSCEIT, asking respondents to provide details of their age, sex, ethnicity and stage of study. In order to test the influence of age on the MSCEIT scores independent of the stage of study, a dichotomous variable was created using a cut-off point of 25 years to distinguish between older students, for whom medicine was likely to be their second degree, and students at a later stage of the medical course. The cut-off point of 25 years of age has also been chosen by the developers of MSCEIT (Meyer et al. 2002), who report that young adults (under 25 years) score significantly lower than those aged over 25 years. Finally, ethnicity was defined using the UK National Statistics interim standard classification of ethnic groups: white, black or black British, Asian or Asian British, mixed, Chinese and other ethnic groups (Office for National Statistics 2008).

Statistical analysis

Data were analysed using SPSS version 15.0. The one-sample Kolmogorov–Smirnov test was used to check whether the continuous variables of the study were normally distributed. Independent sample t-tests were conducted to compare means of MSCEIT scales between the sexes, different age groups and groups created according to time taken to complete the test. The Kruskal–Wallis one-way analysis of variance (ANOVA) test was used to compare the time of completion between different ethnic groups, and the Spearman correlation procedure was used in order to measure the association between completion time and MSCEIT scores. One-way ANOVA tests were used to determine whether significant differences existed among the means of MSCEIT scales in different ethnic groups or stage of study, based on the post hoc Tukey's test. Finally, multivariate analysis was conducted for all MSCEIT scales using linear regressions, based on the Enter method, to identify factors predictive of MSCEIT scores. Although the descriptive statistics for all the MSCEIT scales have been calculated and presented for all the different ethnic groups (white, black or black British, Asian or Asian British, mixed, Chinese and other ethnic groups), in the multiple linear regression models, the ethnic groups with the lower numbers of representatives have been treated as one group (i.e., the results of Black or black British, ethnically mixed students, Chinese and other ethnic groups, have been analysed as one group). Assuming that the ethnic group membership (in one of the above three groups, white, Asian or Asian British and other) would add another 5% in the proportion of variance explained, our tests had adequate power (0.80) to detect the ethnic group differences in MSCEIT scores, at a significance level of 0.05.

Results

A total of 358 students completed the demographic questionnaire and 265 of these also completed the MSCEIT. Two students who completed the questionnaire in less than 10 min were eliminated from the analysis because they had scored more than 3 SDs below the mean on the aggregated scale, suggesting they had made random responses. A total of 93 students who completed the demographic questionnaire, but not the MSCEIT were also excluded from the study. Two-hundred and sixty-three responses (a response rate of 12.3%) were, therefore, included in the final analysis.

Completion times were not normally distributed. The mean completion time was 31.5 min (SD 15.5, range 11–135). No association was found between completion time and the scores of any of the subscales of MSCEIT, or with age group, sex, ethnicity or stage of study. Fifty-six percent of students (146/263) completed the MSCEIT in less than 30 min, and their mean aggregated EI score, did not differ significantly from the rest of the students (101.5 vs. 100.8, p > 0.05). 12.2% of students (32/263) completed the questionnaire in less than 20 min, and, similarly, their scores did not differ significantly from the rest of the students (99.3 vs. 101.5, p > 0.05).

Respondents

Table 3presents the students’ demographic details and their stage in the medical curriculum. Females, white students and ethnically mixed students were over-represented in our sample. Additionally, year 2 and final year students were under-represented and year 4 students were over-represented compared with the total number of students at each stage of the curriculum. As a check on possible sampling bias, means and standard errors for the principal outcomes (MSCEIT aggregate and branch scores), were recalculated using a weighting system (for sex, ethnicity and stage of study). They were found to be very close to the unweighted values which we used in our analysis.

Table 3.  Characteristics of the study sample and the total medical school population

Characteristic	Study Sample	Medical school population
Number of students	263 (12.3%)	2114 (100%)
Sex
Male	75 (28.5%)	824 (39.0%)
Female	188 (71.5%)	1290 (61.0%)
Age
<25 years	191 (72.6%)	1403* (66.5%)
≥25 years	72 (27.4%)	707* (33.5%)
Ethnicity
White	161 (61.2%)	938 (44.5%)
Asian or Asian British	43 (16.3%)	487 (23.1%)
Black or Black British	10 (3.8%)	138 (6.5%)
Chinese or other South-East Asian	22 (8.4%)	283 (13.4%)
Mixed	27 (10.3%)	87 (4.1%)
Unknown	0 (0%)	177 (8.4%)
Stage of study
Year 1	56 (21.3%)	442 (20.9%)
Year 2	37 (14.1%)	427 (20.2%)
Year 3	59 (22.4%)	403 (19.1%)
Year 4	72 (27.4%)	433 (20.5%)
Year 5	39 (14.8%)	409 (19.3%)

Note: *Four missing, transferred or dropped out.

MSCEIT scores

All MSCEIT scores in our sample were normally distributed. The mean aggregated MSCEIT score was 101.2 (SD 13.7, range 72–147), while the mean experiential area score was 98.8 (SD 14.6, range 66–139) and the mean strategic area score was 104 (SD 12.1, range 71–150). For the branch scores, the perceiving emotions mean score was 98.4 (SD 14.7, range 32–141), facilitating thought mean score was 100.9 (SD 15.2, range 65–152), understanding emotions mean score 107.3 (SD 12.5, range 76–145) and managing emotions mean score 98.1 (SD 12.2, range 66–143). Table 4shows the means and SDs according to age group, sex, ethnicity and stage of study for the aggregated score, the two-area scores and the four-branch scores.

Table 4.  Mean scores and SDs on all MSCEIT scales

Characteristics	Perceiving mean (SD)	Facilitating mean (SD)	Experiential mean (SD)	Understanding mean (SD)	Managing mean (SD)	Strategic mean (SD)	Aggregate mean (SD)
Sex
Male	94.6 (15.1)	98.4 (16.7)	94.6 (13.9)	106.8 (13.1)	96.1 (12.4)	102.6 (12.6)	97.5 (13.7
Female	99.9 (14.2)*	102 (14.4)	100.4 (14.6)*	107.5 (12.3)	98.9 (12.1)	104.5 (11.8)	102.6 (13.4)*
Age
<25 years	97.5 (14.9)	100.3 (15.2)	97.9 (14.8)	105.9 (12.5)	96.8 (12.0)	102.4 (12.1)	99.6 (13.6)
≥25 years	100.7 (13.8)	102.8 (15.1)	101.1 (13.9)	110.8 (11.9)*	101.5 (12.2)*	108.2 (11.0)**	105.4 (13.1)*
Ethnicity
White	99.9 (14.3)	101.6 (14.7)	100.2 (14.3)	109.2 (12.4)*	99.7 (12.6)	106.2 (12.2)*	103.3 (13.4)*
Asian or Asian British	93.1 (13.2)	98.9 (15.3)	93.6 (13.6)	101.9 (12.1)*	95.4 (10.9)	98.9 (10.9)*	95.1 (12.5)*
Black or Black British	99.1 (13.5)	100.6 (14.5)	98.9 (14.8)	110.4 (16.6)	95.4 (11.3)	104.2 (12.5)	101.7 (15.6)
Chinese or other South-East Asian	98.8 (14.6)	99.7 (16.8)	98.3 (14.7)	103.9 (11.4)	94.5 (10.9)	99.6 (10.6)	98 (10.5)
Mixed	97.3 (18.4)	101.5 (17.3)	99 (16.9)	105.9 (10.0)	96.6 (12.8)	102.4 (11.0)	100.8 (15.6)
Stage of study
Year 1	97 (13.9)	99.4 (13.9)	96.8 (13.8)	108 (9.7)	95.2 (12.6)*	102.6 (10.6)*	99.2 (12.6)
Year 2	95.3 (13.7)	100.3 (16.1)	96.2 (14.9)	103.2 (12.4)	93.9 (10.4)*	99 (11.1)*	96.3 (12.7)
Year 3	102.2 (16.4)	102 (13.9)	101.9 (14.9)	107 (13.9)	98.2 (12)	103.9 (12.9)	103.4 (15.1)
Year 4	98 (15.1)	102.7 (16.4)	99.6 (15.3)	107.6 (13.2)	100.1 (10.5)	105.6 (12.2)	102.9 (14.5)
Year 5	98.2 (12.4)	98.9 (15.8)	97.8 (13.4)	109.9 (12.0)	102.5 (14.9)*	107.6 (11.9)*	102.3 (11.0)

Note: *p < 0.05 and **p < 0.001.

Women scored slightly higher than men on all of the scales, with their aggregated score and scores in the experiential area and in perceiving emotions branch reaching statistical significance (95% CI:1.4–8.8, 2.1–9.7 and 1.3–9.3, respectively). Analysis of age differences in MSCEIT scores indicated that younger students (under 25 years) scored significantly lower in the strategic area and its two branches, understanding and managing emotions (95% CI: −8.9 to −2.7, −8.2 to −1.6 and −7.4 to −1.2, respectively). The effect of stage of study was significant in the managing emotions branch, with final year students scoring significantly higher compared to those in years 1 and 2 (95% CI: 0.8–13.7 and 1.2–15.5, respectively). However, no significant differences in aggregate EI scores across the year groups were noted. Students from the Asian or Asian British group scored significantly lower than those from the white group on the strategic area and on the understanding emotions branch (95% CI:−12.5 to −1.8 and −13.1 to −1.6, respectively).

Multiple regression models

Although the response rate was low, the size of the responders’ population was adequate to provide our study the power (0.78) to explain even small proportions of variance of MSCEIT (R² = 0.05) at a significance level of 0.05 (Cohen 1988.).

Three variables explained 9.2% of the variation in aggregated MSCEIT score. Being a female student, over 25 years and white (compared to Asian or Asian British) was predictive of higher aggregated scores. In the perceiving branch, sex and ethnicity (white compared to Asian or Asian British) accounted for 4.4% of the variance, while in the understanding branch, 5.4% of the variance was explained by ethnicity (white compared to Asian or Asian British). A total of 6.7% of the variance in the managing emotions branch is explained by the stage of study (higher in final year compared to both years 1 and 2). Our analysis did not explain any of the variation for the facilitating thoughts branch. The adjusted regression coefficients for our models are summarised in Table 5, showing the effect of each variable when adjusted for confounding variables.

Table 5.  Multiple regression models for all MSCEIT scales

	Adjusted regression coefficient	Standardised B coefficient (95% CI)
Total EIQ (adjusted R^2 = 0.092)
Constant	93.9 (86.3–101.5)
Ethnicity (Asian/Asian British)	−7.1 (−11.5 to −2.6)	−0.19
Sex (female)	4.5 (0.9–8.1)	0.15
Age group (over 25 years)	4.3 (0.5–8.0)	0.14
Strategic area (adjusted R^2 = 0.092)
Constant	88.8 (80.5–97.1)
Year of study (year 2)	5.1 (1.2–9.0)
Ethnicity (Asian/Asian British)	−5.8 (−10.7 to −0.90)	0.16
Age group (over 25 years)		−0.15
Strategic area (adjusted R^2 = 0.092)
Constant	104.3 (97.6–111)
Year of study (year 2)	−6.9 (−12.2 to −1.5)	−0.20
Ethnicity (Asian/Asian British)	−6.4 (−10.4 to −2.5)	−0.20
Age group (over 25 years)	3.7 (0.4 to 7.0)	0.14
Perceiving emotions branch (adjusted R^2 = 0.044)
Constant	90.6 (82.2–98.9)
Sex (female)	4.4 (0.5 to 8.4)	0.14
Ethnicity (Asian/Asian British)	−6 (−10.9 to −1)	−0.15
Understanding emotions branch (adjusted R^2 = 0.054)
Constant	108.9 (101.8–115.9)
Ethnicity (Asian/Asian British)	−6.7 (−10.8 to −2.5)	−0.20
Managing emotions branch (adjusted R^2 = 0.067)
Constant	97.4 (90.5–104.3)
Year of study (year 1)	−6.2 (−11.2 to −1.2)	−0.21
Year of study (year 2)	−7.3 (−12.7 to −1.8)	0.21

Discussion

Our exploratory study to the best of our knowledge is the first to report on the use of the MSCEIT in measuring EI in an undergraduate medical population, and provides information on the utility of the online version of the instrument and preliminary data on EI across a medical school population.

The use of an online instrument provided several potential benefits including speed and ease of data collection, and students’ ability to complete the test at a time and place convenient to them. However, our response rate was low and there may be several reasons for this, including lack of interest, problems with email communication, and lack of access to advice on completing the MSCEIT. Since 93 students completed the demographic questionnaire, and then started (but did not complete) the MSCEIT, this may suggest that the length of the MSCEIT was prohibitive and in future, studies using a shorter instrument could be considered. For example, one study (Tapia & Marsh 2006) discusses the validation of a 41-item scale, the emotional intelligence inventory (EII), based on the MSCEIT. More recently, a revision of the EII has been discussed (Kempenaers et al. 2008) suggesting that further work is required in the development of this instrument.

The overall MSCEIT scores for the entire sample are average and this may be of concern given that the medical profession requires professionals with well-developed EI. Our findings that female and older respondents tend to obtain higher EI scores are consistent with the literature (Mayer et al. 1999; Carrothers et al. 2000; Ciarrochi et al. 2000; Mayer et al. 2002; Palmer et al. 2005; Van Rooy 2005; Extremera et al. 2006; Wong et al. 2007). Since students in their final year of study significantly outperform year 1 and 2 students in managing emotions, this finding might suggest a positive impact of the curriculum, or at least an indication that students do not become less emotionally intelligent as they progress. Because we distinguished between students aged under 25 and those aged 25 years and above, we avoided the possibility of a confounding variable (i.e., age and increasing age associated with each progressive year), with older students in the earlier years outperforming their younger counterparts. As far as ethnic differences are concerned, it is possible that because of the linguistic and visual specificity of some of the questions, the current version of the MSCEIT may lack cross-cultural application. In a previous study, it has been suggested that language differences may explain the poorer objective structured clinical examination (Wass et al. 2003) performance of some ethnic minority, as opposed to ethnic majority students, and this may have been the case here. Items related to understanding emotions in the MSCEIT call for a subtle understanding of the differences between seemingly similar terms, indicating that the MSCEIT at certain points may be a measure of language comprehension rather than EI. Again, the question of the cross-cultural applicability of the MSCEIT is raised.

Our preliminary cross-sectional study examined a small sample from one medical school and it would be valuable to conduct more research in collaboration with other medical schools. There is a paucity of research measuring EI over time and a recognition of the importance of conducting longitudinal studies (Brotheridge 2006). Longitudinal studies could, therefore, investigate changes as students progress. Moreover, our study did not examine the link between EI and measures of academic performance and future studies could investigate this. Several studies report interventions that can successfully develop EI (Slaski & Cartwright 2003; Satterfield & Hughes 2007; Ulutas & Omeroglu 2007), but this begs the question about whether EI has any relationship to clinical communication skills, quality of patient care or preservation of personal emotional well being. Future work in medical educational settings could evaluate the impact of an intervention programme. Our findings provide new information on the use of the MSCEIT in a medical school setting, upon which subsequent research may be based.

Acknowledgements

We are very grateful to our department e-resources developer, Stevo Durbaba, and our department statistician, Paul Seed, for their valuable assistance with this study, and for the support given to us by the KCL Medical Education Committee, with particular thanks to professor John Rees.

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

Additional information

Notes on contributors

Mathew Todres

MATHEW TODRES is a research associate in medical education in the Department of General Practice and Primary Care, King's College London.

Zoi Tsimtsiou

ZOI TSIMTSIOU is an honorary research fellow in the Department of General Practice and Primary Care, King's College London. She is also a general practitioner currently working in the NHS in Greece.

Anne Stephenson

ANNE STEPHENSON is a senior lecturer and director of community education in the Department of General Practice and Primary Care, King's College London. She is also a general practitioner working in Southeast London.

Roger Jones

ROGER JONES is a Wolfson professor and head, Department of General Practice and Primary Care, King's College London. He is also a general practitioner working in Southeast London.