Research Report
Can Empathy, Other Personality Attributes, and Level of Positive Social Influence in Medical School Identify Potential Leaders in Medicine? Mohammadreza Hojat, PhD, Barret Michalec, PhD, J. Jon Veloski, MS, and Mark L. Tykocinski, MD
Abstract Purpose To test the hypotheses that medical students recognized by peers as the most positive social influencers would score (1) high on measures of engaging personality attributes that are conducive to relationship building (empathy, sociability, activity, self-esteem), and (2) low on disengaging personality attributes that are detrimental to interpersonal relationships (loneliness, neuroticism, aggressionhostility, impulsive sensation seeking). Method The study included 666 Jefferson Medical College students who graduated in 2011–2013. Students used a peer nomination instrument to identify
P
ositive social influence, according to leadership scholars, is a core feature of effective leadership.1–3 Leadership emerges from interpersonal interactions that occur when individuals work together to achieve a common goal.3 Eberly and colleagues1 have defined leadership as an exertion of social influence that can be examined through social network analysis. This approach has also been used to identify the emergence of potential leaders.4–6 Another approach to identifying effective leaders is studying leaders’ personality attributes.1 For example, certain cognitive abilities, empathy, emotional intelligence, Please see the end of this article for information about the authors. Correspondence should be addressed to Dr. Hojat, Sidney Kimmel Medical College at Thomas Jefferson University, 1015 Walnut St., Curtis Building, Suite 320, Philadelphia, PA 19107-5083; telephone: (215) 955-9549; e-mail: Mohammadreza.Hojat@ Jefferson.edu. Acad Med. 2015;90:505–510. First published online January 27, 2015 doi: 10.1097/ACM.0000000000000652
Academic Medicine, Vol. 90, No. 4 / April 2015
classmates who had a positive influence on their professional and personal development. At matriculation, these students had completed a survey that included the Jefferson Scale of Empathy and Zuckerman–Kuhlman Personality Questionnaire short form and abridged versions of the Rosenberg Self-Esteem Scale and UCLA Loneliness Scale. In multivariate analyses of variance, the method of contrasted groups was used to compare the personality attributes of students nominated most frequently by their peers as positive influencers (top influencers [top 25% in their class distribution], n = 176) with those of students nominated least frequently (bottom influencers [bottom 25%], n = 171).
Results The top influencers scored significantly higher on empathy, sociability, and activity and significantly lower on loneliness compared with the bottom influencers. However, the effect size estimates of the differences were moderate at best.
sociability, tolerance of ambiguity, and social appraisal skills have been reported to foster effective leadership across a variety of situations.7–12
(b) to focus on identifying the engaging, relationship-oriented personality attributes of future medical leaders.
In particular, “relationship-oriented” personality attributes such as empathy,7–9,12,13 active engagement,14 and self-confidence14 have been reported as being conducive to effective leadership.11,15–19 In contrast, there are other personality attributes that are at odds with the “engaging style” of leadership.20 For example, isolation, loneliness, neuroticism, impulsiveness, and aggression have been reported as being detrimental to social relationships21,22 and, by extension, to effective leadership.1–3 Much of the previous research on leadership has focused on the personality profile, or the individualistic model of leadership.7–11 However, with the increasing shift toward team-based, interprofessional, and collaborative health care delivery, researchers need (a) to examine a paradigm of engaging, relationship-oriented leadership that will facilitate teamwork and
Conclusions The research hypotheses were partially confirmed. Positive social influencers appear to possess personality attributes conducive to relationship building, which is an important feature of effective leadership. The findings have implications for identifying and training potential leaders in medicine.
On the basis of the assumptions that leadership is an exertion of social influence1–3 and that social influence is a function of pertinent personality attributes,7–19 we designed this study to test the following hypotheses: (1) Medical students who are recognized by their peers as having positive social influence in medical school will score high on measures of “engaging” personality attributes that are conducive to relationship building (e.g., empathy, sociability, activity, and selfesteem); and (2) these students will score low on “disengaging” personality attributes that are detrimental to interpersonal relationships (e.g., loneliness, neuroticism, aggression-hostility, and impulsive sensation seeking). Method
Participants Participants in this study included 666 graduating students at Jefferson Medical
505
Research Report
College (now, Sidney Kimmel Medical College at Thomas Jefferson University). The participants graduated in 2011 (n = 236), 2012 (n = 216), and 2013 (n = 214). Instrument to measure positive social influence We used a social networking assessment approach, similar to the peer nomination inventory described by Wiggins and Winder,23 to determine the degree of students’ positive social influence among their peers. (See Procedures, below, for a description of our peer nomination instrument.) Instruments to measure engaging (relationship-building) personality attributes Empathy. We used the Jefferson Scale of Empathy (JSE), a 20-item validated instrument specifically developed to measure empathy in the context of patient care in medical and other health professions students and practitioners. We used the S version of the JSE, which was developed for administration to medical students.24 Evidence in support of the JSE’s validity25–29 and reliability25,29 has been reported, and the instrument has been translated into 43 languages and used in more than 60 countries.30 The possible score range is 20 to 140; a higher score on this scale indicates a greater orientation toward empathic engagement in patient care. The typical Cronbach alpha coefficient for this instrument, which has been reported in many studies, hovers around 0.75.24,25,29,30 A sample item on this scale is “It is difficult for a physician to view things from patients’ perspectives.”
measures a tendency to be active and to prefer challenging work.31 Evidence in support of the validity and reliability of this scale in male (α = 0.67) and female (α = 0.72) college students has been reported.31 A higher score on this scale indicates a higher degree of preference for challenging work. A sample item from this scale is “I like complicated jobs that require a lot of effort and concentration.” Self-esteem. We used an abridged, five-item version of the Rosenberg SelfEsteem Scale,33 which is a measure of the self-acceptance aspect of self-esteem.34 This abridged scale has been used with medical and other health professions students.35–37 The reliability coefficient of this abridged scale among health professions students has been reported as 0.72.36 A higher score on this scale indicates a higher degree of self-esteem. A sample item from this scale is “I feel that I am a person of worth, at least on an equal basis with others.” Instruments to measure disengaging personality attributes Loneliness. We used an abridged, fiveitem version of the UCLA Loneliness Scale, which is a global measure of loneliness experiences.38 The abridged version has been used previously with medical and other health professions students,35,36 and its psychometric support in medical students has been reported.37 The reliability coefficient of the abridged scale among health professions students has been reported as 0.87.36 A higher score on this scale indicates a greater experience of loneliness and a lack of satisfaction with social relationships. A sample item from this scale is “I feel isolated from others.”
Sociability. We used a seven-item scale from the short form of the Zuckerman– Kuhlman Personality Questionnaire (ZKPQ) to measure sociability.31,32 (The ZKPQ was developed to measure five basic factors of personality that have a strong biological–evolutionary basis.31) Evidence in support of the validity and reliability of this scale in male (α = 0.78) and female (α = 0.79) college students has been reported.31 A higher score on this scale indicates a more sociable personality. A sample item from this scale is “I tend to start conversations at parties.”
Neuroticism. We used a seven-item scale from the short form of the ZKPQ that measures a tendency to be tense, to worry, to be overly sensitive to criticism, to be easily upset, and to be obsessively indecisive.31 Evidence in support of validity and reliability of this scale in male (α = 0.70) and female (α = 0.72) college students has been reported.31 A higher score on this scale indicates a more neurotic personality. A sample item from this scale is “I often worry about things that other people think are unimportant.”
Activity. We used a seven-item scale from the short form of the ZKPQ that
Aggression-hostility. We used a sevenitem scale from the short form of the
506
ZKPQ that measures a tendency to express verbal aggression and to show rudeness, thoughtlessness, vengefulness, spitefulness, a quick temper, and impatient behavior.31 Evidence in support of the validity and reliability of this scale in male (α = 0.66) and female (α = 0.67) college students has been reported.31 A higher score on this scale indicates a higher degree of aggression and hostility. A sample item from this scale is “If people annoy me, I do not hesitate to tell them so.” Impulsive sensation seeking. We used a seven-item scale from the short form of the ZKPQ that measures a tendency to act quickly on impulse without planning, often in response to a need for thrills and excitement, change, and novelty.31 Evidence in support of validity and reliability of this scale in male (α = 0.62) and female (α = 0.71) college students has been reported.31 A higher score on this scale indicates a higher degree of impulsiveness and thrill-seeking behavior. A sample item from this scale is “I often do things on impulse.” Procedures This study, determined to be exempt from review by the Thomas Jefferson University institutional review board, was conducted with graduating medical students between 2011 and 2013 as part of the Jefferson Longitudinal Study of Medical Education.39,40 At the beginning of medical school, the participating students had completed a survey measuring personality attributes (ZKPQ short form)31 that are less amenable to change under normal circumstances; the survey also included the S version of the JSE24–29 and abridged versions of the Rosenberg Self-Esteem Scale34 and the UCLA Loneliness Scale.38 These personality attributes data and selected demographic data were retrieved from the Jefferson Longitudinal Study database and merged with peer nomination data (i.e., the number of nominations each student received from peers). To identify students’ degree of positive social influence among their peers, we asked students in each graduating class to complete a peer nomination instrument, which included the names of all students in the class listed in alphabetical order. Students were told that these nominations would be used in a study
Academic Medicine, Vol. 90, No. 4 / April 2015
Research Report
impulsive sensation seeking). We also calculated Cohen d as an estimate of the effect size of the differences.41,42
designed to enhance understanding of personal connections. The instrument’s instructions asked students to think back on their medical school experiences and respond to the following question: “Which of your classmates had significant positive influences on your professional and personal development?” The instructions specified that students were to consider all of their classmates who had positive influences on them and to check as many names as they deemed necessary. Participation was voluntary, and no incentive was offered. Students were not asked to sign their names on the peer nomination sheets. These peer nomination data were collected annually in 2011, 2012, and 2013 at the Match meeting when the results of the National Resident Matching Program were announced.
Results
The 666 participating students included 328 (49%) men and 338 (51%) women and represented 87% of the students in the three graduating classes (n = 763; 379 men, 384 women). These students were included as study participants because complete data on the JSE and measures of personality attributes were available for them. We compared the demographics of participants and nonparticipants to ensure the representativeness of the sample: Participants were representative of their entire classes with respect to gender (P < .15), age (P < .82), and ethnicity (P < .62). Frequency distributions of participating students by gender and average age at entrance to medical school by level of positive influence are reported in Table 1. There were no differences in the proportions of men and women among the top, middle, and bottom influencers (χ2(2) = 0.43, P = .80). Results of univariate analysis of variance indicated that the top influencers were slightly older than the rest of the participants (F(2,663) = 6.6, P < .01).
Statistical analyses We examined the frequency distribution of the number of positive influence nominations by classmates in each class (range: 2–82 nominations per participating student; median = 30). We designated students in the top 25% of the distribution for each class as the top influencers (n = 176) and those in the bottom 25% (n = 171) as the bottom influencers. For the purpose of the statistical analyses, we used the method of “contrasted groups” for a more pronounced distinction between the top and bottom influencers; thus, the middle group (middle 50%, n = 319) was excluded from the statistical analyses for a sharper contrast between the two extreme groups. With the exception of the empathy scores from the JSE, we transformed the scores of the personality attributes to a standard distribution with a mean of 100 and a standard deviation of 10 to assess the magnitudes of group differences on a unique scale for all measures of personality attributes. No such transformation was made of the JSE scores to allow direct comparisons of the results with the published data on this scale. We used multivariate analysis of variance, followed by univariate analyses of variance and Duncan post hoc multiple range tests, to examine the significance of the difference between the top and bottom influencers (the independent variable) on each of the dependent variables (empathy, sociability, activity, self-esteem, loneliness, neuroticism, aggression-hostility, and
Academic Medicine, Vol. 90, No. 4 / April 2015
Table 2 reports the mean scores and standard deviations on measures of engaging (relationship-oriented) and disengaging personality attributes for the top and bottom influencers as well as summary results of the statistical analyses. The multivariate F ratio was
statistically significant (F(8,335) = 5.1, P < .01), indicating statistically significant differences between the top and bottom influencers on some measures of personality attributes. Results of univariate analysis of variance showed that the top influencers scored significantly higher on three engaging (relationship-oriented) personality attributes: empathy (F(1,342) = 6.3, P
Can Empathy, Other Personality Attributes, and Level of Positive Social Influence in Medical School Identify Potential Leaders in Medicine? Mohammadreza Hojat, PhD, Barret Michalec, PhD, J. Jon Veloski, MS, and Mark L. Tykocinski, MD
Abstract Purpose To test the hypotheses that medical students recognized by peers as the most positive social influencers would score (1) high on measures of engaging personality attributes that are conducive to relationship building (empathy, sociability, activity, self-esteem), and (2) low on disengaging personality attributes that are detrimental to interpersonal relationships (loneliness, neuroticism, aggressionhostility, impulsive sensation seeking). Method The study included 666 Jefferson Medical College students who graduated in 2011–2013. Students used a peer nomination instrument to identify
P
ositive social influence, according to leadership scholars, is a core feature of effective leadership.1–3 Leadership emerges from interpersonal interactions that occur when individuals work together to achieve a common goal.3 Eberly and colleagues1 have defined leadership as an exertion of social influence that can be examined through social network analysis. This approach has also been used to identify the emergence of potential leaders.4–6 Another approach to identifying effective leaders is studying leaders’ personality attributes.1 For example, certain cognitive abilities, empathy, emotional intelligence, Please see the end of this article for information about the authors. Correspondence should be addressed to Dr. Hojat, Sidney Kimmel Medical College at Thomas Jefferson University, 1015 Walnut St., Curtis Building, Suite 320, Philadelphia, PA 19107-5083; telephone: (215) 955-9549; e-mail: Mohammadreza.Hojat@ Jefferson.edu. Acad Med. 2015;90:505–510. First published online January 27, 2015 doi: 10.1097/ACM.0000000000000652
Academic Medicine, Vol. 90, No. 4 / April 2015
classmates who had a positive influence on their professional and personal development. At matriculation, these students had completed a survey that included the Jefferson Scale of Empathy and Zuckerman–Kuhlman Personality Questionnaire short form and abridged versions of the Rosenberg Self-Esteem Scale and UCLA Loneliness Scale. In multivariate analyses of variance, the method of contrasted groups was used to compare the personality attributes of students nominated most frequently by their peers as positive influencers (top influencers [top 25% in their class distribution], n = 176) with those of students nominated least frequently (bottom influencers [bottom 25%], n = 171).
Results The top influencers scored significantly higher on empathy, sociability, and activity and significantly lower on loneliness compared with the bottom influencers. However, the effect size estimates of the differences were moderate at best.
sociability, tolerance of ambiguity, and social appraisal skills have been reported to foster effective leadership across a variety of situations.7–12
(b) to focus on identifying the engaging, relationship-oriented personality attributes of future medical leaders.
In particular, “relationship-oriented” personality attributes such as empathy,7–9,12,13 active engagement,14 and self-confidence14 have been reported as being conducive to effective leadership.11,15–19 In contrast, there are other personality attributes that are at odds with the “engaging style” of leadership.20 For example, isolation, loneliness, neuroticism, impulsiveness, and aggression have been reported as being detrimental to social relationships21,22 and, by extension, to effective leadership.1–3 Much of the previous research on leadership has focused on the personality profile, or the individualistic model of leadership.7–11 However, with the increasing shift toward team-based, interprofessional, and collaborative health care delivery, researchers need (a) to examine a paradigm of engaging, relationship-oriented leadership that will facilitate teamwork and
Conclusions The research hypotheses were partially confirmed. Positive social influencers appear to possess personality attributes conducive to relationship building, which is an important feature of effective leadership. The findings have implications for identifying and training potential leaders in medicine.
On the basis of the assumptions that leadership is an exertion of social influence1–3 and that social influence is a function of pertinent personality attributes,7–19 we designed this study to test the following hypotheses: (1) Medical students who are recognized by their peers as having positive social influence in medical school will score high on measures of “engaging” personality attributes that are conducive to relationship building (e.g., empathy, sociability, activity, and selfesteem); and (2) these students will score low on “disengaging” personality attributes that are detrimental to interpersonal relationships (e.g., loneliness, neuroticism, aggression-hostility, and impulsive sensation seeking). Method
Participants Participants in this study included 666 graduating students at Jefferson Medical
505
Research Report
College (now, Sidney Kimmel Medical College at Thomas Jefferson University). The participants graduated in 2011 (n = 236), 2012 (n = 216), and 2013 (n = 214). Instrument to measure positive social influence We used a social networking assessment approach, similar to the peer nomination inventory described by Wiggins and Winder,23 to determine the degree of students’ positive social influence among their peers. (See Procedures, below, for a description of our peer nomination instrument.) Instruments to measure engaging (relationship-building) personality attributes Empathy. We used the Jefferson Scale of Empathy (JSE), a 20-item validated instrument specifically developed to measure empathy in the context of patient care in medical and other health professions students and practitioners. We used the S version of the JSE, which was developed for administration to medical students.24 Evidence in support of the JSE’s validity25–29 and reliability25,29 has been reported, and the instrument has been translated into 43 languages and used in more than 60 countries.30 The possible score range is 20 to 140; a higher score on this scale indicates a greater orientation toward empathic engagement in patient care. The typical Cronbach alpha coefficient for this instrument, which has been reported in many studies, hovers around 0.75.24,25,29,30 A sample item on this scale is “It is difficult for a physician to view things from patients’ perspectives.”
measures a tendency to be active and to prefer challenging work.31 Evidence in support of the validity and reliability of this scale in male (α = 0.67) and female (α = 0.72) college students has been reported.31 A higher score on this scale indicates a higher degree of preference for challenging work. A sample item from this scale is “I like complicated jobs that require a lot of effort and concentration.” Self-esteem. We used an abridged, five-item version of the Rosenberg SelfEsteem Scale,33 which is a measure of the self-acceptance aspect of self-esteem.34 This abridged scale has been used with medical and other health professions students.35–37 The reliability coefficient of this abridged scale among health professions students has been reported as 0.72.36 A higher score on this scale indicates a higher degree of self-esteem. A sample item from this scale is “I feel that I am a person of worth, at least on an equal basis with others.” Instruments to measure disengaging personality attributes Loneliness. We used an abridged, fiveitem version of the UCLA Loneliness Scale, which is a global measure of loneliness experiences.38 The abridged version has been used previously with medical and other health professions students,35,36 and its psychometric support in medical students has been reported.37 The reliability coefficient of the abridged scale among health professions students has been reported as 0.87.36 A higher score on this scale indicates a greater experience of loneliness and a lack of satisfaction with social relationships. A sample item from this scale is “I feel isolated from others.”
Sociability. We used a seven-item scale from the short form of the Zuckerman– Kuhlman Personality Questionnaire (ZKPQ) to measure sociability.31,32 (The ZKPQ was developed to measure five basic factors of personality that have a strong biological–evolutionary basis.31) Evidence in support of the validity and reliability of this scale in male (α = 0.78) and female (α = 0.79) college students has been reported.31 A higher score on this scale indicates a more sociable personality. A sample item from this scale is “I tend to start conversations at parties.”
Neuroticism. We used a seven-item scale from the short form of the ZKPQ that measures a tendency to be tense, to worry, to be overly sensitive to criticism, to be easily upset, and to be obsessively indecisive.31 Evidence in support of validity and reliability of this scale in male (α = 0.70) and female (α = 0.72) college students has been reported.31 A higher score on this scale indicates a more neurotic personality. A sample item from this scale is “I often worry about things that other people think are unimportant.”
Activity. We used a seven-item scale from the short form of the ZKPQ that
Aggression-hostility. We used a sevenitem scale from the short form of the
506
ZKPQ that measures a tendency to express verbal aggression and to show rudeness, thoughtlessness, vengefulness, spitefulness, a quick temper, and impatient behavior.31 Evidence in support of the validity and reliability of this scale in male (α = 0.66) and female (α = 0.67) college students has been reported.31 A higher score on this scale indicates a higher degree of aggression and hostility. A sample item from this scale is “If people annoy me, I do not hesitate to tell them so.” Impulsive sensation seeking. We used a seven-item scale from the short form of the ZKPQ that measures a tendency to act quickly on impulse without planning, often in response to a need for thrills and excitement, change, and novelty.31 Evidence in support of validity and reliability of this scale in male (α = 0.62) and female (α = 0.71) college students has been reported.31 A higher score on this scale indicates a higher degree of impulsiveness and thrill-seeking behavior. A sample item from this scale is “I often do things on impulse.” Procedures This study, determined to be exempt from review by the Thomas Jefferson University institutional review board, was conducted with graduating medical students between 2011 and 2013 as part of the Jefferson Longitudinal Study of Medical Education.39,40 At the beginning of medical school, the participating students had completed a survey measuring personality attributes (ZKPQ short form)31 that are less amenable to change under normal circumstances; the survey also included the S version of the JSE24–29 and abridged versions of the Rosenberg Self-Esteem Scale34 and the UCLA Loneliness Scale.38 These personality attributes data and selected demographic data were retrieved from the Jefferson Longitudinal Study database and merged with peer nomination data (i.e., the number of nominations each student received from peers). To identify students’ degree of positive social influence among their peers, we asked students in each graduating class to complete a peer nomination instrument, which included the names of all students in the class listed in alphabetical order. Students were told that these nominations would be used in a study
Academic Medicine, Vol. 90, No. 4 / April 2015
Research Report
impulsive sensation seeking). We also calculated Cohen d as an estimate of the effect size of the differences.41,42
designed to enhance understanding of personal connections. The instrument’s instructions asked students to think back on their medical school experiences and respond to the following question: “Which of your classmates had significant positive influences on your professional and personal development?” The instructions specified that students were to consider all of their classmates who had positive influences on them and to check as many names as they deemed necessary. Participation was voluntary, and no incentive was offered. Students were not asked to sign their names on the peer nomination sheets. These peer nomination data were collected annually in 2011, 2012, and 2013 at the Match meeting when the results of the National Resident Matching Program were announced.
Results
The 666 participating students included 328 (49%) men and 338 (51%) women and represented 87% of the students in the three graduating classes (n = 763; 379 men, 384 women). These students were included as study participants because complete data on the JSE and measures of personality attributes were available for them. We compared the demographics of participants and nonparticipants to ensure the representativeness of the sample: Participants were representative of their entire classes with respect to gender (P < .15), age (P < .82), and ethnicity (P < .62). Frequency distributions of participating students by gender and average age at entrance to medical school by level of positive influence are reported in Table 1. There were no differences in the proportions of men and women among the top, middle, and bottom influencers (χ2(2) = 0.43, P = .80). Results of univariate analysis of variance indicated that the top influencers were slightly older than the rest of the participants (F(2,663) = 6.6, P < .01).
Statistical analyses We examined the frequency distribution of the number of positive influence nominations by classmates in each class (range: 2–82 nominations per participating student; median = 30). We designated students in the top 25% of the distribution for each class as the top influencers (n = 176) and those in the bottom 25% (n = 171) as the bottom influencers. For the purpose of the statistical analyses, we used the method of “contrasted groups” for a more pronounced distinction between the top and bottom influencers; thus, the middle group (middle 50%, n = 319) was excluded from the statistical analyses for a sharper contrast between the two extreme groups. With the exception of the empathy scores from the JSE, we transformed the scores of the personality attributes to a standard distribution with a mean of 100 and a standard deviation of 10 to assess the magnitudes of group differences on a unique scale for all measures of personality attributes. No such transformation was made of the JSE scores to allow direct comparisons of the results with the published data on this scale. We used multivariate analysis of variance, followed by univariate analyses of variance and Duncan post hoc multiple range tests, to examine the significance of the difference between the top and bottom influencers (the independent variable) on each of the dependent variables (empathy, sociability, activity, self-esteem, loneliness, neuroticism, aggression-hostility, and
Academic Medicine, Vol. 90, No. 4 / April 2015
Table 2 reports the mean scores and standard deviations on measures of engaging (relationship-oriented) and disengaging personality attributes for the top and bottom influencers as well as summary results of the statistical analyses. The multivariate F ratio was
statistically significant (F(8,335) = 5.1, P < .01), indicating statistically significant differences between the top and bottom influencers on some measures of personality attributes. Results of univariate analysis of variance showed that the top influencers scored significantly higher on three engaging (relationship-oriented) personality attributes: empathy (F(1,342) = 6.3, P
