Nature or Nurture? Have Health Care Professionals Always Been Good People or Do They Only Become So During Their Professional Career? An Empirical Study of Personality Styles in Health Care Professionals and the General Population

doi:10.21203/rs.3.rs-7164956/v1

Download PDF

Research Article

Nature or Nurture? Have Health Care Professionals Always Been Good People or Do They Only Become So During Their Professional Career? An Empirical Study of Personality Styles in Health Care Professionals and the General Population

https://doi.org/10.21203/rs.3.rs-7164956/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Healthcare professionals (HCPs) need certain personality traits to cope with the emotional and interpersonal challenges of their profession. This study investigated the personality styles of HCPs using the Personality Style and Disorder Inventory (PSDI). It focused on functional and dysfunctional patterns. Participants were 6,803 experienced and young HCPs (aged 18 to 30 years) as well as students and a normative population from German-speaking countries. Robust statistical analyses revealed significant style differences between HCPs and the general population. HCPs scored lower on the styles "willful/paranoid", "spontaneous/borderline", "reserved/schizoid" and "ambitious/narcissistic". HCPs exhibit greater emotional stability, empathy and relationship orientation. Young HCPs showed styles that were almost identical to those of experienced professionals, while students showed significantly less structured styles. The cluster analysis identified three distinct functional clusters characterized by a) resilient, socially competent, b) impulsive-dysregulated and c) inhibited-internalizing personality styles. These results indicate that personality traits of HCPs develop through early professional socialization and not solely through age. This underscores the importance of personality traits for professional development in healthcare and the quality of patient care.

Health care professionals (HCP)

physicians

dentists

psychologists

psychotherapists

hypnotists

nurses

occupational therapists and physiotherapists

Personality Style and Disorder Inventory (PSDI)

factor analysis

cluster analysis

robust statisticial methods

Healthcare professions focus on the diagnosis, treatment, and prevention of diseases and disorders with the aim of maintaining and/or restoring health. In this paper, we summarize persons and occupational groups responsible for these tasks as health care professionals (HCP). They include doctors, dentists, psychologists, nurses and professionals from paramedical professions such as occupational therapy and physiotherapy (WHO, 2013).

The University of Minnesota (2023) lists ten key personality traits in an information brochure for prospective health care professionals. At the top of the list are characteristics such as helpfulness, empathy and compassion. These are followed by commitment, strong communication skills and other competencies that indicate positive character aptitude, which, among other things, contribute to job satisfaction (Richardson et al., 2009) or protect against burnout (Betts et al., 2024).

1.1 Personality traits in healthcare professions

In the past, research into the relationship between personality and occupation in general focused primarily on whether people choose their occupation based on existing personality traits (person-job fit hypothesis; Holland, 1997; Schneider, 1987) or whether people adapt to their occupation, i.e., the occupation shapes their personality (socialization hypothesis; Bühler et al., 2024). Research has also been conducted into whether people with similar personality traits choose similar occupations and whether the length of time spent in an occupation leads to personality traits converging with those of experienced colleagues (Anni et al., 2025). This is confirmed by Rossetti et al. (2025) using data from 11,000 people over a period of 12 years: People tend to develop more homogeneous personality traits within an occupation than between different occupations due to attraction, selection, occupational change and socialization. In the following, we examine these general findings specifically for people with occupations in the healthcare sector.

Several studies have investigated the importance of personal characteristics for successful performance in the healthcare sector. Most studies used the so-called Big Five model with the five personality factors openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism (Costa & McCrae, 1992).

A systematic review by van der Wal et al. (2022) found, for example, that successful anesthetists have lower scores in neuroticism and higher scores in extraversion and conscientiousness. Vermeulen et al. (2024) found significant differences between OR nurses and norm samples, particularly lower neuroticism scores. Other studies have shown correlations between empathy and successful treatment (Hojat et al., 2011) or between personality factors and patient satisfaction (Apedzi & Apedzi, 2024).

While there are several studies on physicians and nurses (Ammi et al., 2023; Louwen et al., 2023), the number of studies on dentists is more limited. However, some studies (z.B. Asokan et al., 2023; Chamberlain et al., 2005) also showed low neuroticism, high conscientiousness and agreeableness. They are predictors of academic and career success. Further research pointed to the importance of empathic communication, particularly in dental anxiety (Furnes et al., 2025; Jones & Huggins, 2014).

1.2 Personality and psychotherapy success

The relevance of personality is particularly evident in the psychotherapeutic context. In some studies, the so-called "therapist effect" explained between 5% and 10% of the variance in treatment outcomes (Baldwin & Imel, 2013). This revealed large differences in performance: the most effective therapists achieved significantly better results than averagely successful therapists and had lower dropout rates (Delgadillo et al., 2020; Saxon, Barkham, Foster, & Parry, 2017). Some of these differences can already be detected during education (Schwartz et al., 2025). Important predictors are interpersonal skills such as relationship and communication skills, empathy, acceptance and warmth (Elliott et al., 2018). They are more significant for therapy success than age, gender or therapeutic school (Andersen et al., 2025).

1.3 Personality models: Big Five versus PSDI

Research on therapists has so far concentrated primarily on descriptive traits such as the Big Five personality traits (Costa & McCrae, 1992). To complement and extend the above-mentioned studies, which primarily tested personality traits, we focused on personality styles and used the PSDI (Personality Styles and Disorders Inventory) developed by Kuhl and Kazén (2009, 2024) for this purpose. It is based on Kuhl's PSI theory(2000, 2001), which sees personality as an interplay of cognitive-affective and executive macrosystems:

1. Intention memory (storage and maintenance of intentions, plans and goals; analytical, sequential, conscious)

2. Extension memory (holistic integration of life experiences about one's own "self" and its environment; holistic, parallel, largely unconscious)

3. Intuitive behavior control (execution of automated actions and implementation of intentions; associative, effortless, unconscious)

4. Object recognition system (recognition of dangers and errors, detailed perception; sequential, strenuous, conscious)

The balance and flexibility between these systems, which are modulated by positive or negative affects, determine a person's ability to control themselves. Disturbances in the interaction of these systems can lead to dysfunctional personality styles. The PSDI measures the extent to which certain styles are pronounced that may indicate imbalances in the four macrosystems mentioned above. The styles are not analyzed dimensionally, as is the case with the Big Five through factor analyses, for example, but rather embody individual patterns of experience in the confrontation with social and ecological conditions. They shape how people react to typical life demands with individual behavioral patterns and therefore are functional patterns of how a person deals with certain psychological demands. Styles thus show how emotional arousal affects action and how processual procedures take place in the above-mentioned macrosystems, i.e., how regulation, self-control, integration of motivation and affect are organized. They therefore have significance for mental health or disorders. Styles are thus functionally embedded in a complex system of self-regulation processes (including volitional initiation, affect modulation, self-control of cognitive and emotional processes and behavioral action programs). In addition to positive sequences of such processes, they can also escalate into personality disorders and lead to dysfunctional personality styles. The PSDI allows the characteristics of personality styles and their changes to be recorded, which indicate imbalances in these systems. Table 1 shows the main differences between PSDI personality styles and personality traits (Big Five).

As shown in Table 2, the PSDI scales are named bipolar. This does not mean that there are two separate styles, but rather a continuum of styles. Whether low values of this continuum reflect more positive style characteristics and the other pole characterizes negative styles pointing to the pathological depends on the content-related, clinically relevant meaning of the respective style.

The strength of these styles is characterized by values that correspond to a T-distribution. T-values below 40 indicate a below-average style. T-values between 40-60 represent average characteristics and therefore do not indicate any conspicuous styles. Finally, T-values above 60 represent above-average characteristics that indicate a tendency towards pathological styles.

One advantage of the PSDI lies in its postulated clinical relevance: It allows a more differentiated view of functional and dysfunctional personality styles. However, its lower prevalence, limited standardization and the partially moderate internal consistency of individual scales, and particularly an inconsistent distribution of style expression in relation to the normal T-based distribution form of the styles (see below) are criticized.

1.4 Personality styles in healthcare professions: Our studies

Our research group has used the PSDI several times to examine the personality profiles of healthcare professionals (HCPs). The extent to which the styles of these individuals differ significantly from the normal population was examined (see Figure 1). In a study of 1027 psychotherapists from the DACH countries (Germany, Austria and Switzerland), significant differences were found between HCPs and the norm sample (Peter et al., 2017). Very low scores were particularly pronounced in the styles PN willfull/paranoid, BL spontaneous/borderline, SZ reserved/schizoid and NAR ambitious/narcissistic. The low scores in these four styles were interpreted as emotional stability, empathic ability, low egocentricity and relationship orientation – central attitudes that are required, for example, in the client-centered approach according to Rogers (1957). The first three of these styles, PN, BL and SZ, had already been shown with the same low values in a previous study of hypnotherapists from the DACH countries (Peter et al., 2012) and all four again in a subsequent study for dentists (Peter & Wolf, 2022). For dentists in this and other studies (Peter & Wolf, 2021, 2022; Wolf, Baumgärtner, & Peter, 2022) the increased values for conscientious/compulsive (ZW) were striking.

Dentists working with hypnosis (Wolf et al., 2022) and hypnotherapists (Peter et al., 2012) also showed increased values in the styles SL unselfish/self-sacrificing, ST intuitive/schizotypal, HI charming/histrionic and in RH optimistic/rhapsodic.

Despite different samples, all these studies provide a consistent picture of functional personality styles that reflect the professional requirements of psychotherapeutic and hypnotherapeutic, dental and hypnosis-dental work.

1.5 Training and personality development

As already indicated above, it is an open question whether such personality styles are a prerequisite or rather the result of professional activity. Previous studies on HCPs show mixed findings: Gumz et al. (2024) found lower interpersonal skills in psychology students than in psychotherapy trainees. Demisch and Kuchinke (2022) reported that older and more experienced therapists were less neurotic but open to new experiences, but also less conscientious than younger ones.

Our own studies with students of psychology, dentistry and STEM subjects consistently showed different profile trajectories to those of HCPs. Student profiles meander strongly around the mean value of the normal population, while those of professionals are much more structured and consistently show the same profile progression of styles from e.g. PN willful/paranoid to ZW conscientious/compulsive in all our studies, as can be seen in Figure 1 as an example for DACH psychotherapists (Peter & Böbel, 2020; Peter et al., 2017). The same meandering profile progression of students had already been shown in an earlier study of psychology and STEM students (Bochter et al., 2014). In a publication by Peter and Böbel (2020, cf. Figure 2), data is published on psychotherapy training candidates whose personality profile no longer resembles the meandering profile of the students, but is already approaching the profile of the professionals, which indicates the selection and/or adaptation processes mentioned above.

The findings to date – both from the literature and from our own studies – show a consistent picture: People working in healthcare professions predominantly have personality styles that support their work and protect them as a person. Their key resources are empathy, agreeableness, openness and emotional stability. The studies we have conducted to date essentially confirm these observations in different subgroups of health care professionals.

The aim of the present study is to substantiate the overall picture of previous studies using broad groups of healthcare professionals – and to examine whether healthcare professionals differ from the normal population in German-speaking countries in terms of specific characteristics. In contrast to previous studies on personality traits of HCPs, mainly based on the Five-Factor Inventory (Costa & McCrae, 1992), we supplement these by examining 14 personality styles using the Personality Styles and Disorders Inventory (Kuhl & Kazén, 2009), whose functions are embedded in a broad psychological personality theory. The gender- and age-specific style characteristics and profiles of the HCPs are examined in comparison to a norm sample, as well as the characteristics of superordinate style factors and style clusters, which were obtained from these styles with the help of factor and cluster analyses. Based on these factors and clusters, we want to test hypotheses as to which individuals are better suited to the profession of HCP, and which individuals may be less well suited. Embedded in this analysis is also an investigation into whether the differences in styles between HCPs and the norm population only become apparent when working as an HCP or already during training as an HCP and thus either represent a prerequisite for taking up a heath care profession or rather become effective because of a heath care profession. In addition, we no longer use t-tests and classical analyses of variance in this study to investigate differences in means between groups and variables, as renowned statisticians have repeatedly shown that these methods are not sensitive to violations of the normal distribution and heteroscedasticity of data in either small or large studies, and can therefore distort statistical results (Blanca et al., 2017; Delacre et al., 2017; Wilcox, 2012). As an alternative, this study therefore exclusively uses robust methods (robust MANOVAs and robust ANOVAs or the so-called Welch-ANOVA) as well as permutation tests, which have proven to be largely insensitive to such violations.

3.1 Survey instrument

The Personality Styles and Disorders Inventory in short version (PSDI-S; Kuhl and Kazén, 2009) was used to record personality styles. With 14 scales and 56 items, it records individual patterns of thinking, feeling, and behavior that represent adaptive or tendentially dysfunctional personality styles. Each scale consists of four statements that are answered using a 4-point Likert scale (not at all / some / much / completely). The reliability of the scales is between α = .63 and α = .79. The long version of the PSDI has been shown to correlate with Big Five dimensions (Costa & McCrae, 1992) and 16PF dimensions (Cattell et al., 1993); the short form was used for reasons of economic survey duration.

3.2 Sample

The current overall study of HCPs includes 3,805 adults from the German speaking DACH countries (Germany, Austria and Switzerland) who we were able to recruit between 2017 and 2024 from a data pool of around 19,000 members of various psychological, medical and dental societies and participants in their conferences as well as in the context of academic theses at German universities. The participants were informed by email about the conduct and content of this study and asked to participate via the internet portal SoSci Survey for conducting online studies (Leiner, 2024). In addition to other questions not covered in this article, the items of the PSDI-S were provided, supplemented by some professional and socio-demographic information such as the age and gender of the participants.

At the beginning of their participation in the survey, participants were informed that participation was voluntary and could be discontinued at any time without negative consequences. At the end of the survey, participants were informed that all data had been stored anonymously and were asked for their consent to use their data for research purposes.

Responses from 3,411 HCPs were analyzed for the current study. 2,486 persons were women and 925 men with an average age of 46 years for women and 51.2 years for men. The youngest person was 18 for both men and women and the oldest was 83 for women and 89 for men. The participants all came from healthcare professions or were still studying psychology (N = 159) or dentistry (N = 249). The subjects included 1,524 psychological and 154 medical psychotherapists, 540 dentists, 217 psychologists, alternative practitioners or doctors without psychotherapeutic accreditation. The remaining 406 people worked with state-recognized training in physiotherapy, occupational therapy, or healthcare and nursing in various branches of the healthcare system.

To contrast the results of the current study, we used the previously unpublished norm sample of 3,392 people already used in Peter et al. (2017), which was provided by Kazén 2016. It can be regarded as representative of the German-speaking population as a whole in terms of gender, age, level of education, professional activities, and family circumstances and as comparable to the previous norm samples specified in the two test manuals by Kuhl and Kazén (2009, 2024).

3.3 Data analysis

The statistical analyses of the 14 personality styles and their relationships to the variables age and gender as well as other variables derived from these basic variables were carried out with the statistical program Jamovi (version 2.6.17.0, The jamovi project, 2024) and various method packages of Python version 3.12.2, which were installed via the Anaconda distribution. All calculations were performed on a multi-core processor with 40 GB RAM, so that even computationally intensive procedures such as permutation tests could be realized quickly.

The following methods were used to calculate the differences between the PSDI values and the factors derived from them, as well as their dependence on different study groups and other influencing variables:

Robust multivariate analysis of variance (MANOVA) according to Pillai's trace (Hampel et al., 1986; Levene, 1960; Tabachnick & Fidell, 2019)
Permutation tests (n = 5,000 replicates), (Good, 2005, 2013)
Bootstrapping for the estimation of confidence intervals (n = 5,000 replicates), (Efron & Tibshirani, 1993)
Univariate analyses with robust ANOVA (Wilcox, 2012), Welch-ANOVA and Games-Howell test, (Games & Howell, 1976; Welch, 1947)
Kruskal-Wallis test for non-parametric comparisons, (Kruskal & Wallis, 1952)
Effect sizes: Cohen's d, Hedges' g, (Field et al., 2012)
Exploratory (EFA) and confirmatory factor analysis (CFA), (Field et al., 2012)
Hierarchical and k-means cluster analysis, (Field et al., 2012)
Multinomial logistic regression to analyze extreme value distributions.

In the following, we will first briefly outline the publication by Peter et al. (2017), which to a certain extent represents the start of the current article, and then carry out a detailed analysis of the PSDI test and describe the 14 styles in terms of content and statistics and examine the relationship of the styles to the results of the norm sample by Kazén (2016).

4.1 Results of the original DACH study by Peter et al. (2017)

In an earlier study, Peter et al. (2017) used the PSDI-S for the first time to investigate whether psychotherapists from the DACH countries (N = 1,027) differ in their personality styles from a norm sample (N = 3,392). The PSDI scores of this study can be found in Fig. 1 as a simple black continuous line labeled PB 2017. The result showed a significant difference between the two groups in four personality styles (PN willful/paranoid, BL spontaneous/borderline, SZ reserved/schizoid and NAR ambitious/narcissistic), which can contribute to relationship skills such as empathy and appreciation, openness to the patient's emotional experience and building a trusting relationship. Moderate but also statistically significant differences were found in seven personality styles (AB loyal/dependent to AS assertive/antisocial), which were equally indicative of the psychotherapists' professional social skills, i.e. could be characterized as neither submissive nor passive, not overly helpful, but also not overly self-confident. There were hardly any or no differences in the HI charming/histrionic, RH optimistic/rhapsodic and ZW conscientious/compulsive styles, i.e. general friendliness, optimism and conscientiousness. The exact statistical data of the style differences between this DACH-study (PB 2017) and the norm group (NM) can be found in Table 3. This observation has since been confirmed in many subsequent studies by our working group (Peter & Böbel, 2020; Peter et al., 2017; Peter et al., 2012; Peter & Wolf, 2021, 2022; Wolf et al., 2022) (see Fig. 2).

The following study, in which we have combined the groups of PB 2017 (DACH), PB 2020, PW 2021 and ND 2024 (ND = New Data) to form the overall HCP group, aims to replicate this finding in a large group of healthcare workers. In addition, several aspects will be examined as to how this deviation of healthcare workers from the normal population can be described and which styles in particular set this group so massively apart from the normal population.

4.2 Results of the PSDI of the HCP group vs. the normal population

Checking the data for normal distribution: The data set of our HCP group (N = 3,003) and students (N = 408) exhibits several problems that are not unusual for empirical studies of this type but can have a negative impact on the quality of the statistical analyses. These include missing values and non-normally distributed data due to left- or right-skewed (screwness) and very sharp curvature (kurtosis). Missing values were only present in less than 4% of the data in this dataset. They were corrected by a simple mean value imputation from the existing data. In addition, the study groups had widely differing group sizes. While the PSDI variables PN willful/paranoid, SZ reserved/schizoid, HI charming/histrionic, RH optimistic/rhapsodic, ZW conscientious/compulsive show no strong skewness or conspicuous kurtosis or heteroscedasticity, several variables (BL spontaneous/borderline (1.168), NT critical/negativistic (1.092) and DP passive/depressive (1.103)) are strongly skewed to the right and a number of variables are also slightly skewed to the left. NT critical/negativistic (1.314) and DP passive/depressive (1.141) are also slightly acutely curved. More significant, however, is the fact that the Levene test for testing heteroscedasticity or variance homogeneity for seven variables (BL spontaneous/borderline (p = 0.005), NAR ambitious/narcissistic (p = 0.002), AB loyal/dependent (p = 0.017), SL unselfish/self-sacrificing (p = 0.009), SU self-critical/avoidant (p = 0.001), DP passive/depressive (p = 0.012) and AS assertive/antisocial (p < 0.001) produced strong to very strong fluctuations in the variance between the groups. This can cause problems when using classic MANOVA/ANOVA methods, so that more robust methods or non-parametric tests and permutation tests were chosen to validate the results. The broad black curve in Fig. 1 illustrates the curve progression (profile) of the overall group of HCPs from style PN willful/paranoid (left) to style ZW conscientious/compulsive (right) in comparison to the profile of the norm sample (NM).

The profile of the HCP shows a clear deviation of all styles from the norm group (NM). While the latter meander with a mean value of around 50, the values of the HCP group rise continuously from the PN willful/paranoid style (42.7) to the ZW conscientious/compulsive style and are just above the mean value of the norm sample of 50 for both styles RH optimistic/rhapsodic at 51.1 and ZW conscientious/compulsive at 51.7. The average style values of the HCP group are always below those of the norm group. The range of values for all styles extends from 25 for the PN willful/paranoid style to 89 for the NT critical/negativistic style. This result thus replicates the observations from Peter et al. in 2017 and has since been confirmed in many subsequent studies by our working group (see Fig. 1). Table 4 shows the corresponding style values of the HCP group and the NM norm group.

The differences in PSDI scores between the two groups HCP and NM were tested with a robust multivariate analysis of variance (MANOVA) and further follow-up tests. Pillai's Trace was preferred as the test statistic, as this test statistic has been shown to be particularly robust to violations of normal distribution and variance homogeneity (Field, 2018; Tabachnick & Fidell, 2019).

With a Pillai's Trace of 0.264 with F(14/6788) = 173.553 and a significant p-value of < 2.2e-16 = < .0001), Table 5 confirms that there are strong multivariate differences between the groups. In addition, the stability of the MANOVA result was tested with bootstrapping from 5,000 replicates (Good, 2005; Wilcox, 2012). The mean Wilks' lambda value of the bootstrapping was 0.9959 and thus within the 95% confidence interval of 0.9935 and 0.9978. The bootstrapping thus confirms very stable differences between the style characteristics of the two groups. An additional permutation test with 5,000 repetitions came to the same result and confirmed the results of the multivariate group difference with highly significant p-values of < 0.0001 and a range of Wilks' lambda values (min-max of 0.9917 to 0.9986).

The subsequent group comparison of the 14 PSDI styles using the Welch test (Table 5) revealed highly significant differences between the two groups for all 14 individual variables, with the HCPs always achieving lower values than the norm sample, except for the RH optimistic/rhapsodic and ZW conscientious/compulsive styles. In the HCP group, the RH and ZW styles were significantly higher than in the norm sample. For all comparisons, the associated p-values are consistently significantly small (< 0.001). A subsequent permutation test also confirmed these results. These results were corroborated by individual comparisons between the two groups, also using the Games-Howell post-hoc test and effect sizes according to Hedges' g. The results are shown in Table 5. Hedges' g is like Cohen's d but corrects for unequal group sizes. The value range varies between negative (HCP lower than NM) to positive (HCP higher than NM) and, analogous to Cohen's d, means a small effect for a g|| ≈ 0.2, |g| ≈ 0.5 a medium effect and |g| ≈ 0.8 a large effect.

The results thus provide multiple statistically verified differences in the style characteristics between the groups. While the HCP groups, except for two styles, have lower values in most styles in the lower half of the norm range between 40 and 50, the NM norm sample shows a constant mean norm value of 50 for all styles.

For the results of the group comparisons HCP vs. NM, large effects were determined for the following four styles: PN willful/paranoid (Ω² = 0.126, g = -0.763), BL spontaneous/borderline (Ω² = 0.10, g = -0.674), SZ reserved/schizoid (Ω² = 0.077, g = -0.581) and NAR ambitious/narcissistic (Ω² = 0.083, g = -0.606). Medium-sized effects were found for the variables AB loyal/dependent (Ω² = 0.051, g = -0.467) and NT critical/negativistic (Ω² = 0.035, g = -0.383) and small effects were shown by the styles ST intuitive/schizotypal, SL unselfish/self-sacrificing, SU self-critical/avoidant, DP passive/depressive, AS assertive/antisocial and HI charming/histrionic (g ≈ 0.1 to 0.3).

The statistical group comparisons and the effect sizes thus indicate in a significant way that people who work in the healthcare sector obviously deviate significantly from the norm sample in the personality styles and also replicate the results of an earlier DACH study (Peter et al., 2012).

4.3 Analysis of extreme T-values of the PSDI styles

However, when looking at the distribution of styles again, it was also noticeable that many group members of the HCP and norm group achieved style values that lie outside the value range of T = 40 to 60, which is characterized as normal. In the case of styles with values less than 40 or greater than 60, both test manuals by Kuhl and Kazén (2009, 2024) speak of conspicuous style values. If the value falls below 40, the styles tend to be characterized as deficient and if the values exceed 60, they tend to be characterized as pathological. These distributions are therefore analyzed in more detail below. Table 6 shows the distribution of these more extreme value ranges for both groups.

As can be seen in Table 6, the values of most styles in both groups are, as expected, in the normal range between 40 and 60. In the case of HCP, this is 2,023 people across all styles and in the norm group 2,218 people, i.e. 67.4% and 65.4% of the people in these groups. However, around a third of HCPs are outside the normal range, with almost 25% at the lower extreme and around 9% at the upper extreme; in contrast, the figure for NMs is roughly the same at both extremes, averaging 17%.

As can already be expected from Fig. 1, the greatest differences between the HCP and NM groups lie in the first four styles PN willful/paranoid, BL spontaneous/borderline, SZ reserved/schizoid and NAR ambitious/narcissistic, which make up the left end of the PSDI profile. They are mainly dominated by people in the HCP group. Conversely, people in the NM group mainly dominate the styles in the normal range and above 60.

The Kruskal-Wallis test supports this distribution. For the range below 40, a highly significant chi-square χ² of 19.6 with df = 2, p < 0.001, ɛ = 0.491 was found with a significantly greater presence of people in the HCP group than people in the norm group. For the 40–60 range, the χ² test with 35.9 with a p < 0.001 and df = 2 or ɛ of 0.889 showed more people in the NM group and also for the range above 60, with a χ² of 32.9, df = 3 and p < 0.001 or ɛ of 0.823, highly significant group differences with clear dominance of the norm group. These results were confirmed with robust post-hoc tests and mean comparisons using bootstrapping with 5,000 runs to test confidence interval (CI)-based effect sizes (F(1/16.4) = 392, p < 0.001, ɛ =0.828; bootstrap CI lower/upper: 0.795 : 0.867).

In the following, we will deepen this observation that people with different personality styles are more frequently represented in one of the three areas with the help of a multimodal logistic regression. This method provides information on the probability of a personality style being observed more frequently or less frequently in one of the extreme ranges < 40 or > 60 compared to the normal range 40–60, depending on membership of the HCP or NM group. The dependent variable is the subdivided value range < 40, 40–60 and > 60, with membership of the respective group acting as a predictor. The probability is examined using a so-called omnibus likelihood ratio test χ². Although the multimodal logistic regression was calculated for all 14 styles, only styles that are characteristic of one of the two groups are shown here – i.e. either occur more frequently or less frequently than "normal" (Table 7). Styles in which one of the two groups shows a marked accumulation in the extreme areas are particularly relevant. Such characteristics are psychologically considered potentially relevant for behavior and stress – in both a positive and negative sense. These results are summarized in Table 8.

These results also show that the groups of healthcare professionals (HCP) and the general population (NM) differ systematically in several personality styles – both in the lower range of style values < 40 and in the range above values > 60 of the PSDI style scales. The odds ratio (OR) indicates how strong the probability is that a person from one of the two groups is represented in a certain value range in contrast to the other group. An OR > 1 indicates that people from one of the two groups are overrepresented in one area. For the BL spontaneous/borderline style in the upper range (> 60), this results in an OR of 7.84. This means that people from the general population are almost eight times more likely to be in this extreme value range than people from healthcare professions. In the lower range (< 40), the style PN willful/paranoid shows an OR of 0.31. This means that HCPs are more than three times more likely to be in this range – which indicates that HCPs have lower levels of this style.

4.4 Comparison with students

Like all psychological characteristics, PSDI styles are subject to social, ecological, economic and, finally, political influences. On the one hand, this affects the fact that responses to the PSDI items can change over time due to such influences. Another influence may be that the styles of the HCPs change during professional experience. With this last aspect in mind, we examined below whether the styles of young people in the HCP group differ from older members of this group and whether young HCP members and older HCP members differ significantly in their styles from people who, as students of dentistry and psychology, may become potential HCPs in the future, but who, as students, are still on the path to a healthcare profession. Are there significant differences between these three groups that indicate that the PSDI styles of students behave in the same way as those of the normal population, but that young HCP members are already deviating from the style characteristics of the normal population in the direction of older HCP individuals without having fully reached their characteristics? This could then suggest that the PSDI styles of HCP individuals only develop gradually over the course of their careers and deviate more from the normal population over time. If the styles of students were already similar to those of young HCPs, one would have to assume that such styles are a basic prerequisite for choosing a profession in the healthcare sector and that it is above all those people who choose a profession there who already exhibit styles that differ from those of the normal population before actually practicing the profession. However, if the styles of students are statistically congruent with the style characteristics of the normal population, it would be very unlikely that predominantly predisposed people in the healthcare sector would establish themselves in the profession in the long term (cf. Holland, 1997; Schneider, 1987; Bühler et al., 2024; Rossetti et al., 2025).

Five groups were compared with each other for this purpose: Dental students (ZM students), psychology students (Psy students), young health care professionals (Y-HCP), older health care professionals (HCP, n = 2,883, i.e. persons of the 3,003 HCP members who are older than 30 years and provided information on age) and the norm sample (NM). A robust multivariate analysis of variance (MANOVA) with Pillai's Trace as test criterion was used, with the 14 PSSI styles (PN willful/paranoid to ZW conscientious/compulsive) as dependent variables and group membership as independent variable.

The graphical comparison of mean values across all 14 scales (see Fig. 2) shows a similar development of the Y-HCP and HCP profiles. In contrast, the two student groups show a clearly divergent profile, which is characterized by higher values for several scales (including PN willful/paranoid, BL spontaneous/borderline, AB loyal/dependent). The norm sample lies between the HCPs and the students in many areas.

Means and standard deviations per group for each scale are shown in Table 8. Subsequent univariate robust ANOVAs for all styles revealed significant differences between groups (all p < .001).

Pillai's Trace resulted in a test score of 0.359 with an F-value of F(56, 12808) = 7.31, p < .001). The results indicate significant differences in the style profile between the five groups.

The largest effect sizes (d > 0.5) were found in the comparisons of students with Y-HCP (Table 9) in the PN willful/paranoid, BL spontaneous/borderline and NT critical/negativistic scales. Overall, students have "stronger" style profiles. The differences between Y-HCP and HCP were consistently very small (all d < 0.2), indicating an almost identical profile. Y-HCP also differed from NM on several scales with medium effect sizes (d ≈ 0.4–0.5). A subsequent permutation test across all 14 PSSI styles yielded a mean F-statistic of 148.41 with a p-value of p < .001 based on 1,000 permutations. The likelihood of such a group difference occurring purely by chance is low. The differences in the personality style profile between the five groups are therefore highly significant and robust, even when the requirements of classic MANOVA are violated (e.g. normal distribution, homoscedasticity).

The results show that the personality style profile of young HCPs (Y-HCP) is already very similar to that of experienced HCPs. This suggests that HCP-relevant styles already appear to emerge during the first years of working in the healthcare sector. Young HCPs (aged 18–30) are already in the profession and typically already have several years of clinical or patient-related experience. In contrast, students of the same age show a significantly different profile, which in several styles is more like that of the norm sample. This leads to the conclusion that the typical HCP profile is not fully explained by a dispositional predisposition, but at least partially develops during professional socialization in the healthcare sector, which supports the current positions of Bühler et al. (2024) and Rossetti et al. (2025).

4.5 Exploratory and confirmatory factor analysis

In the following, the 14 personality styles are examined with the help of an exploratory factor analysis (EFA). The aim was to identify latent structures of the PSDI styles and to reduce the dimensional data structure in addition to the previous analyses of the PSDI. The analysis was intended to clarify whether the 14 PSDI styles can be summarized into a smaller number of meaningfully interpretable latent, superordinate styles due to strong intercorrelations between individual styles and are more likely to have beneficial characteristics or are more likely to be found in excess areas of the style values and thus tend to identify negative deviations. The factor structure obtained was then subjected to a confirmatory factor analysis.

For the analysis of both factor models (EFA and CFA), the data of all persons in the HCP group and norm sample NM (N = 3,003 HCP + N = 408 students + 3,392 NM = a total of 6,803 persons) were used. The EFA was calculated with subjects whose ID numbers were even (even, N = 3,401) and the CFA was calculated with subjects whose ID numbers were odd (odd, N = 3,402). Both factor analyses are therefore independent groups and observations and thus a validation test of the latent factors identified with the help of the EFA.

Before conducting the EFA, the essential prerequisites for the factorizability of the data were checked using a) the Kaiser-Meyer-Olkin (KMO) criterion, the MSA and Bartlett's test. The total KMO value was 0.813, which indicates good suitability of the data for factor analysis (cut-off: >0.60). The individual MSA values of the variables were between 0.615 (AS) and 0.899 (NT), which confirms the factorizability of the data. The Bartlett's test for sphericity was also significant (χ²(78) = 15.407, p < 0.001), which also supports the suitability of the correlation matrix for a factor analysis.

The factors were extracted using the maximum likelihood method, combined with an Oblimin rotation, which allows the factors to be correlated. About the eigenvalue criterion, four of the EFA-identified factors had eigenvalues greater than 1. The Cattell scree plot test confirmed the choice of four factors.

The four factors together explained 53.9% of the total variance (Table 10). Factor one explained 25.3%, factor two 11.63%, factor three 9.65% and factor four 7.31% of the variance in the data. In psychological studies, EFA results from 50% variance clarification are considered appropriate compression and summarization of the data (Field et al., 2012).

The following CFA specified the following four latent factors after adjustment to the model of the EFA results:

Factor 1 (socially competent style): DP passive/depressive, SU self-critical/avoidant, BL spontaneous/borderline, SL unselfish/self-sacrificing, NT critical/negativistic, AB loyal/dependent
Factor 2 (socially sensitive and intuitive style): RH optimistic/rhapsodic, ST intuitive/schizotypal, HI charming/histrionic
Factor 3 (socially problematic, relationship-insecure style): SZ reserved/schizoid, PN willful/paranoid, ST intuitive/schizotypal (ST was removed from factor 2 and moved to factor 3)
Factor 4 (socially dysregulated style): AS assertive/antisocial, NAR ambitious/narcissistic, PN (PN was also moved from factor 3 to factor 4)
Error covariances between: SL & NT, RH & ST, BL & HI, SZ & AS

Factor 1 summarizes with the DP passive/depressive, SU self-critical/avoidant, BL spontaneous/borderline, SL reserved/schizoid, NT critical/negativistic, AB loyal/dependent those styles that, by low levels, reflect socially competent traits such as integrity awareness, and express loyalty, criticality, altruism and self-criticism. Their pathological forms, i.e. in high expression, embody dependent, affectively unstable, conflictual and negativistic behaviors.

Factor 2 combines styles that can be described in a positive way as socially sensitive and intuitive with the styles intuitive (ST), charming (HI) and optimistic (RH) and can reveal themselves in their pathological eccentric-dramatic manifestation through schizotypal, histrionic and rapturous characteristics.

Factor 3 characterizes personality styles that can be characterized in their positive form as willful and in their negative forms as suspicious aloofness with paranoid (PN) and schizoid (SZ) behavioral reactions and are therefore socially problematic.

Finally, factor 4 summarizes styles whose positive manifestations are characterized by an ambitious sense of duty and conscientiousness and whose negative pathological form can appear as dominant egocentrism paired with antisocial (AS) and narcissistic (NAR) traits. We refer to them as socially dysregulated.

Communality describes the squared loading of a variable on a factor and indicates how much of the variance of a variable is explained by this factor. The correlations between the factors show low to moderate correlations: factor 1 and factor 2: r = -0.130; factor 1 and factor 3: r = 0.260; factor 1 and factor 4: r = 0.1008; factor 2 and factor 3: r = -0.250; factor 2 and factor 4: r = 0.1567; factor 3 and factor 4: r = 0.0336. These results indicate that the factors are not completely independent but can still be clearly separated from each other.

The model fit reflected the following quality criteria: the Root Mean Square Error of Approximation (RMSEA) with a value of 0.0613, 90% CI = [0.058, 0.065] (cut-off < 0.08), as well as the Tucker-Lewis Index (TLI) with a value of 0.935 indicate a good model fit (> 0.90). The Standardized Root Mean Squared Residual (SRMR) of 0.07 is also acceptably below 0.08 and confirms a good model fit. The BIC (Bayesian Information Criterion) also indicates a good model fit with a low value of 180. Only the chi-square test of χ²(78) = 15407, p < 0.001 shows a significant deviation from the perfect model, but high values of this measure are common for large samples > 60. They are therefore generally not considered a serious objection to the other positive criteria.

The exploratory factor analysis thus resulted in a theoretically well interpretable stable solution with four factors. The model quality indicators showed an adequate to very good fit of the model to the data. The factor loadings were significant and showed a good discriminatory power of the indicators.

A robust MANOVA and subsequent robust univariate tests were used to investigate whether these four factors differed significantly between the HCP and NM groups. However, neither the MANOVA nor the univariate tests revealed any significant difference between the HCP and norm groups.

4.6 Cluster analysis

In addition to the EFA/CFA, we used cluster analyses to analyze whether the PSDI profiles of the HCP subjects and the norm sample differed systematically. The 14 PSDI scales of all study subjects were z-standardized to enable an equally weighted analysis of all variables. The cluster analysis of the 14 PSDI styles with all 6,803 individuals began with a hierarchical cluster analysis based on the Ward.D2 method and the z-standardized scale means (Euclidean distances between the individuals) to determine the optimal number of clusters. The dendrogram (see Fig. 3) showed a pronounced "jump" in the reduction from three to two clusters, so that a 3-cluster solution was empirically recommended and justified.

This was followed by a k-means cluster analysis with k = 3 (Fig. 4), the results of which show a high degree of homogeneity within the clusters and a clear separation between the clusters, reflected in an explained total variance of 25.7% (between-cluster SS = 24.464; total SS = 95.228). This value can be considered a decent value for psychological personality scales with high individual variance.

Cluster 1: Individuals in this cluster (n = 2,195, black line) show significantly higher values for half of the styles (NAR ambitious/narcissistic, ST intuitive/schizotypal, HI charming/histrionic, RH optimistic/rhapsodic, AS assertive/antisocial), which indicates a resilient, socially competent and self-confident personality style. The style can be characterized as psychologically favorable and well-regulated.

Cluster 2: The second cluster (n = 1,718) (dashed line) is strongly characterized by above-average values for 6 styles (DP passive/depressive, BL spontaneous/borderline, NT critical/negativistic, SU self-critical/avoidant, SL unselfish/self-sacrificing and AB loyal/dependent), with simultaneously reduced values on socially competent scales. It is characterized as a center of impulsive-dysregulated, conflictual and externalizing style elements.

Cluster 3: Finally, cluster 3 (dotted line) represents styles that can be interpreted as inhibited-internalizing. With 2,890 cases, this cluster shows lower scores on almost all scales, especially for SZ reserved/schizoid, PN willful/paranoid, BL spontaneous/borderline, ST intuitive/schizotypal, NAR ambitious/narcissistic, suggesting a controlled, reserved personality style. We therefore describe it as a controlled-internalizing, socially adapted and introverted style.

The present results confirm and extend previous findings on the personality of health care professionals (HCP), as described both in international research (e.g. van der Wal et al., 2022; Vermeulen et al., 2024) and in earlier studies by our research group (Peter et al., 2012, 2017, 2022). It was already explained in the introduction that personality traits such as emotional stability, empathy and low egocentricity are considered key resources that can support HCPs in their professional role and protect them from excessive demands. The present study is the first to use a very large sample of HCPs to examine whether healthcare professionals differ in their personality styles from a normal population in German-speaking countries. The key findings are discussed below and placed in the context of the research literature.

5.1 Personality profiles of healthcare professionals compared to the normal population

The results of the robust multivariate and univariate analyses of variance show significant differences in PSDI styles between health care professionals (HCP) and the norm sample (NM). Particularly striking are the lower scores of the HCPs in the styles PN willful/paranoid, BL spontaneous/borderline and SZ reserved/schizoid. These findings are consistent with previous studies by our research group (Peter et al., 2017; Peter & Wolf, 2022) and confirm the consistent pattern of functional personality styles among healthcare professionals.

The low levels of these styles can be interpreted as emotional stability, empathy and relationship orientation – central attitudes that are required, for example, in the client-centered approach according to Rogers (1957). These characteristics are not only relevant for psychotherapists but appear to be an overarching characteristic of healthcare professionals. The consistency of these findings across different professional groups (psychotherapists, dentists, other healthcare professions) underlines the robustness of this personality profile.

Interestingly, these differences are not only evident in individual styles but manifest themselves in a characteristic profile trajectory that differs significantly from that of the normal population. While the profiles of the HCPs show a clear structure with pronounced minima and maxima, the norm sample shows a barely differentiated profile. This indicates that the observed differences are not isolated characteristics, but rather a fixed pattern of functional personality styles that reflect job-related requirements in the healthcare sector.

5.2 Role of extreme manifestations of the PSDI personality styles

In addition to the mean value analysis, extreme values (< 40 or > 60 T-points) were also considered. This showed that around a third of the HCPs were outside the normal range, primarily in the lower value range. While the norm group more frequently exhibited high (> 60) and potentially pathological scores, HCPs were significantly more frequently represented with scores in styles that are considered positive for social interaction and communication – especially in the four styles mentioned above. These findings shed new light on the often uncritically interpreted under-expression of personality styles as "functional". It remains to be seen whether low scores always reflect desirable resources or possibly also limitations (e.g. reduced assertiveness).

The logistic regression analyses confirmed these tendencies: For example, members of the NM group were almost eight times more likely to be in the higher value range of the BL spontaneous/borderline style (> 60) than HCPs, while the latter were more than three times more likely to be in the extreme lower value range (< 40) for the PN willful/paranoid style, for example.

Overall, these results also support the hypothesis that healthcare professionals have a specific personality profile that differs significantly from that of the general population. This profile is characterized by low levels of maladaptive, self-confident and communication-critical styles and could be an expression of professional selection, social desirability or adaptive development processes during training (cf. Demisch & Kuchinke, 2022; Schwartz et al., 2025).

The clinical diagnostic embedding of the PSDI styles allows a more differentiated assessment of such profiles compared to classic descriptive approaches such as the Big Five. Thus, the present findings not only offer empirical replication of earlier studies but also provide indications for potential aptitude diagnostics and the development of job-related personality profiles.

5.3 Development of personality styles during a career

One particularly revealing aspect of our study is the comparison between young HCPs, experienced professionals and students. The results show that the personality style profile of young HCPs is already very similar to that of experienced HCPs, while students of the same age have a significantly different profile. This suggests that HCP-relevant styles are already evident during the first years of working in the healthcare sector.

The young HCPs (aged 18–30) are already in the profession and have several years of clinical or patient-related experience very early on. In contrast, students of the same age show a profile that is more like that of the norm sample in several styles. This suggests that the typical HCP profile is not fully explained by a dispositional predisposition, but rather develops at least in part during professional socialization in the healthcare sector.

This observation raises the question of the underlying mechanisms: Is it a selection process, in which individuals with certain personality traits are more likely to remain in healthcare professions, or an adaptation process, in which personality changes in response to occupational demands? Our data suggest a combination of both factors. Regardless of possible self-selection processes based on personality traits developed in childhood to early adolescence (Caspi et al., 2005; Roberts et al., 2007), we still know too little about the specific interaction of selection and adaptation processes in the subsequent early years of career choice. The fact that already young HCPs show a characteristic profile may speak for early selection (Holland, 1997) as well as for adaptation (Bühler et al., 2024), but probably for a subtle interaction between the different processes that Rossetti et al. (2025) found: Attraction, selection, occupational change and socialization within an occupation tend to create more homogeneous personalities. The similarity of the profiles of young and experienced HCPs also suggests that these characteristics remain relatively stable once established.

These findings are also consistent with some studies specific to personality development in healthcare professionals. For example, Gumz et al. (2024) found lower interpersonal skills in psychology students than in psychotherapy training candidates, suggesting a developmental process during training. The findings of Demisch and Kuchinke (2022) that older and more experienced therapists are less neurotic but more open to new experiences than younger ones also support the idea of job-related personality development.

5.4 Complementary findings from factor analysis and cluster analysis

An interesting methodological aspect of our study concerns the different results of the factor analysis and the cluster analysis. The exploratory factor analysis (EFA) of the 14 PSDI styles yielded a robust four-factor solution that captured content-consistent and psychologically interpretable latent factors of Kuhl's personality styles (Kuhl, 2001). Together, these four factors explained 53.9% of the total variance and can be interpreted as superordinate personality axes in clinically relevant styles. The factors included dimensions such as (1) socially competent, (2) interpersonally sensitive, (3) socially problematic and (4) dysregulated-impulsive. Our EFA largely corresponds to the factor analysis already presented by Kuhl and Kazén (2009, p. 38) with the same styles on factor 1 and a very similar distribution of styles on the other factors. The subsequent confirmatory factor analysis (CFA) validated this structure, and the fit indices indicated an overall good model fit, suggesting the structural validity of the PSDI factor structure across different groups. While the exploratory and confirmatory factor analysis identified the four-factor structure of the PSDI scales, there were no significant differences between health care professionals and the norm sample in terms of the mean scores of these factors. This indicates that the two groups are similarly organized in their basic personality style structure when viewed at the level of higher-level latent factors.

In contrast, the cluster analysis revealed clear differences in the distribution of people across specific personality style profiles. Three clearly differentiated clusters were identified (functional-resilient and socially competent, impulsive-dysregulated, inhibited-introverted), the frequency of which differed significantly between the groups. Health care professionals were significantly more frequently represented in the functional, socially competent cluster 1, while the norm sample was more strongly represented in the dysregulated cluster 2.

This apparent discrepancy with the factor analysis can be explained methodologically: Factor analysis measures higher-level dimensions that may be similarly pronounced in different individuals, although the specific combination of individual PSDI styles differs. Cluster analysis, on the other hand, captures inter-individual differences in the profile progression of the scales – i.e. not only how strongly individual personality styles are expressed, but also in what combination they occur.

The results of both analyses thus complement each other: the factor analysis shows that the basic structure of the personality styles is similar in HCPs and the normal population, while the cluster analysis makes it clear that the specific patterns of expression differ systematically. This underlines the importance of multiple methodological approaches in personality research (see Table 11).

A central methodological point concerns the relationship between the latent structure identified by the factor analysis and the cluster solution. At first glance, both methods appear to provide similar information – after all, both are based on patterns within the PSDI scale profiles. However, they analyze different levels: Factor analysis describes correlations at the variable level and identifies scales that regularly vary together across a larger number of individuals. The cluster analysis, on the other hand, groups people based on their specific patterns on all scales.

It is noteworthy that the HCP group was predominantly represented in cluster 1, while students more frequently fell into clusters 2 or 3. This could be an indication that a more stable and socially integrated personality profile emerges during professional development (e.g. through experience, selection or resilience building). At the same time, the results emphasize the need to consider psychosocial resources and not just professional skills when selecting and promoting junior staff in the healthcare sector.

In our analysis, we found a high degree of convergence between the factor structure and the cluster structure (Table 11), which indicates that certain personality dimensions are not only present as an abstract latent structure, but also manifest themselves in typical, inter-individually similar personality patterns. This strengthens the validity of the results and indicates that the identified factor structure also has inter-individual relevance in terms of psychologically relevant type formation.

5.5 Limitations and methodological restrictions

Despite the robust findings, the present study has some limitations that should be considered when interpreting the results. A central limitation concerns the interpretation of the PSDI styles, particularly in the lower value range. The PSDI model assumes a bipolar structure of styles, with extremely low T-scores (< 40) reflecting deficient style characteristics, while extremely high T-scores (> 60) indicate dysfunctional or pathological characteristics.

However, this conceptual assumption is not fully empirically validated. In our sample, we found that values below T = 30 hardly ever occur, which indicates that the PSDI is not normally distributed in the lower value range.

In addition, the content-related relationship between the bipolar designations of the PSDI styles is not always clear. For example, "willful" does not appear as a direct contrast to "paranoid" and "spontaneous" does not appear as a direct contrast to "borderline". This conceptual ambiguity makes it difficult to interpret very low scale values. When interpreting the style characteristics, it should be borne in mind that the PSDI was primarily developed as a clinical measurement instrument and is often used as such. Low values in PN willful/paranoid are to be assessed as positive in a clinical-diagnostic sense, whereas high values in PN are negative. The situation is different with HI charming/histrionic or RH optimistic/rhapsodic, where moderately high values above T = 50 tend to be seen as positive, especially for HCP, whereas low values are negative. HI or RH values exceeding T = 60, on the other hand, would possibly be pathological and therefore classified as negative.

Another limitation concerns the cross-sectional structure of the study. Although we compared different age groups and career stages, the design does not allow us to draw any direct conclusions about causal development processes. Longitudinal studies would be necessary to clarify whether and how personality styles change during professional socialization in the healthcare sector.

Finally, it should be noted that, despite its size, our sample is not fully representative of all healthcare professions. The professional groups of psychotherapists and dentists may be overrepresented, while the large group of doctors from different specialties and paramedical staff may be underrepresented. Furthermore, our sample consists exclusively of people from German-speaking DACH countries. All this limits the generalizability of the results.

5.6 Implications and outlook

Despite these limitations, the results offer important implications for research and practice. The consistent differences in the personality profiles between health care professionals and the general population underline the importance of personal characteristics for the successful practice of health care professions. This could be relevant for career counseling, personnel selection and training.

The observation that young HCPs already differ greatly in their personality profile from students of the same age indicates early selection and/or adaptation processes. This could be used for the design of educational programs by specifically promoting competencies associated with functional personality styles. Discussion and research on this have only begun hesitantly, occasionally in medical subjects (Costa et al, 2014; Apedzi & Apedzi, 2024), somewhat more clearly in the psychotherapeutic field (Nodop & Strauß, 2013; Evers et al, 2019; Taubner & Evers, 2022).

The identified clusters also offer a differentiated view of different personality types in the healthcare sector. The functional cluster, in which HCPs are overrepresented, is characterized by a profile that reflects emotional stability, empathy and social skills. These characteristics could be seen as resources that promote both the quality of patient care and the professional satisfaction and resilience of professionals.

Future research should examine the personality profiles identified here in longitudinal studies in order to better understand the developmental dynamics. It would also be interesting to use the PSDI to analyze the relationship between personality styles and concrete professional outcomes such as patient satisfaction, treatment success or burnout risk, i.e. in addition to the personality studies mentioned above, which have been conducted using well-known measurement instruments such as the Big Five. Finally, the conceptual basis of the PSDI model could be further developed in order to make the interpretation of extreme scale values more precise and to provide a better empirical foundation for the bipolar constructs.

5.7 Conclusion

In summary, the results show a consistent, differentiated picture of the personality structure of health care professionals. The stable four-factor structure of the PSDI scales depicts superordinate dimensions of clinically relevant personality styles but only allows limited conclusions to be drawn about group-specific differences. Only a more detailed analysis at the level of individual styles and their configurations – for example in the context of cluster analysis – reveals clear differences between functional and dysregulated personality profiles.

The comparison between young HCPs, experienced professionals and students is particularly revealing. While students differ significantly from health care professionals in several conflictual styles and show a personality profile that is close to the norm or less mature, young HCPs already have a style profile that is almost identical to that of their experienced colleagues. This finding clearly suggests a job-related socialization process through which certain adaptive personality traits apparently develop and consolidate in the early course of the profession – possibly in response to the specific demands of the healthcare sector.

The results thus not only suggest the structural validity of the PSDI but also provide indications of the occupational plasticity of clinically relevant personality styles. Future longitudinal studies could clarify whether this development can be confirmed prospectively – and whether certain personality profiles may be predictive of professional satisfaction, stress tolerance or patient interaction.

Informed consent:

Participation in this study was voluntary, with no rewards or disadvantages for taking part. All participants were of legal age and did not receive any form of compensation. By completing the questionnaire, they gave their written consent for their data to be used in research. All data was fully anonymized, so approval from ethics committees was not needed. The study followed the ethical guidelines of the 1064 Declaration of Helsinki.

Patient and public involvement:

Not applicable.

Credit authorship contribution statement:

Wolfgang H.R. Miltner: Conceptualization, Data analysis, Formal analysis, Writing – original draft, Writing – review & editing, Visualization. Burkhard Peter: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Declaration of interests:

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements:

We are very grateful to Miguel Kazén for providing the data for the normalization sample of the PSDI. We also thank different bots of ChatGPT for the recommendation of several robust statistical methods, testing and correction of scripts for the R-analysis and Python programs, and for helpful recommendations where and how to shorten the text. The English translation was assisted by DeepL.

Funding: There was no funding

Clinical trial number: not applicable

Ammi, M., Fooken, J., Klein, J., & Scott, A. (2023). Does doctors’ personality differ from those of patients, the highly educated and other caring professions? An observational study using two nationally representative Australian surveys. BMJ Open, 13(4). https://doi.org/10.1136/ bmjopen-2022-069850
Andersen, W., Berning, A., Sell, S., Strauss, B., & Taubner, S. (2025). What do psychotherapists in training and patients think about psychotherapeutic competencies? Psychotherapie Psychosomatik Medizinische Psychologie, 75(05), 173-180. https://doi.org/10.1055/a-2553-1326
Anni, K., Vainik, U., & Möttus, R. (2025). Personality profiles of 263 occupations. Journal of Applied Psychology, 110(4), 481–511. https://doi.org/10.1037/apl0001249
Apedzi, A. K., & Apedzi, C. (2024). Personality traits and productivity of healthcare workers: Case study of st. Elizabeth, Holy Family, and st. Patrick Hospitals, Ghana. Biomedical. Journal of Scientific & Technical Research, 58(5). https://doi.org/BJSTR.MS.ID.009210
Asokan, S., Geethapriya, P. R., Dhanabalan, O., & Kumar, T. D. (2023). Assessment of personality traits among pediatric dentists in India: A cross-sectional study. International Journal of Clinical Pediatric Dentistr, 16(3), 489–493
Baldwin, S. A., & Imel, Z. E. (2013). Therapists effects: Findings and methods. In L. M. Lambert & p.-W. nd Garfield´s handbook of psychotherapy and behavior change (6 ed. (Eds.), Bergin and Garfield´s handbook of psychotherapy and behavior change (Vol. 6, pp. 258-297). Wiley.
Betts, C., Stoneley, A., & Picker, T. (2024). Exploring paramedic personality profiles and the relationship with burnout and employment retention: A scoping review. Australasian Emergency Care, 27(4), 227-236. https://doi.org/10.1016/j.auec.2024.04.003
Blanca, M. J., Alarcón, R., Arnau, J., Bono, R., & Bendayan, R. (2017). Non-normal data: Is ANOVA still a valid option? Psicothema, 29(4), 552-557.
Bochter, B., Hagl, M., Piesbergen, C., & Peter, B. (2014). Persönlichkeitsstile von Psychologiestudierenden im Vergleich zu Studierenden sogenannter MINT-Fächer [Personality styles of students of psychology in contrast to STEM students]. Report Psychologie, 39(4), 154–165.
Bühler, J. L., Orth, U., Bleidorn, W., Weber, E., Kretzschmar, A., Scheling, L., & Hopwood, C. J. (2024). Life Events and Personality Change: A Systematic Review and Meta-Analysis. European Journal of Personality, 38(3), 544–568. https://doi.org/10.1177/08902070231190219
Caspi, A., Roberts, B., & Shiner, R. (2005). Personality development: Stability and change. Annual review of psychology, 56, 453–484. https://doi.org/10.1146/annurev.psych.55.090902.141913
Cattell, R. B., Cattell, A. K., & Cattell, H. E. P. (1993). 16PF Fifth Edition Questionnaire. Champaign, IL: Institute for Personality and Ability Testing.
Chamberlain, T. C., Catano, V. M., & Cunningham, D. P. (2005). Personality as a predictor of professional behavior in dental school: comparisons with dental practitioners. Journal of Dental Education, 69(11), 1183–1292. https://doi.org/10.1002/j.0022-0337.2005.69.11.tb04021.x
Costa, P. T., & McCrae, R. R. (1992). NEO Personality Inventory-Revised (NEO-PI-R) and NEO Five-Factory Inventory (NEO-FFI) Professional Manual. Psychological Assessment Resources. .
Delacre, M., Lakens, D., & Leys, C. (2017). Why Psychologists Should by Default Use Welch’s t -test Instead of Student’s t -test. International Review of Social Psychology, 30(1), 92–101. doi:10.5334/irsp.82
Delgadillo, J., Branson, A., Kellett, S., Myles-Hooton, P., Hardy, G., & Shafran, R. (2020). Therapist personality traits as predictors of psychological treatment outcomes. Psychotherapy Research, 30(7), 857–870. doi:10.1080/10503307.2020.1731927o
Demisch, A. M., & Kuchinke, L. (2022). Do the relationships between age and the personality of psychotherapists differ from expected trajectories? A cross-sectional study. Counselling and Psychotherapy Research, 22, 970–981. https://doi.org/10.1002/capr.12529
Efron, B., & Tibshirani, R. J. (1993). An Introduction to the Bootstrap. Chapman & Hall.
Elliott, R., Watson, J. C., Bohart, A. C., & Murphy, D. (2018). 4Therapist empathy and client outcome: An updated meta-analysis. Elliott, R., Watson, J. C., Bohart, A. C., & Murphy, D. (2018). Therapist empathy and client outcome: An updated meta-analysis. Psychotherapy75, 55(4), 399–410. https://doi.org/10.1037/pst0000175
Evers, O., Schröder-Pfeifer, P., Möller, H., & Taubner, S. (2019). How do personal and professional characteristics influence the development of psychotherapists in training? Results from a longitudinal study. Research in Psychotherapy: Psychopathology, Process and Outcome, 22(3), 389–401. 10.4081/ ripppo.2019.424
Field, A., Miles, J., & Filed, Z. (2012). Descovering Statistics Using R. Sage.
Furnes, M. E., Lillejord, S., Lillejord, V., & Johnsen, J. A. K. (2025). The relationship between the perceived personality traits of dentists, dental anxiety, negative stories, and negative experiences with dental treatment: A cross-sectional study. Dentistry Journal, 13(162). https://doi.org/10.3390/dj13040162
Games, P. A., & Howell, J. F. (1976). Pairwise multiple comparison procedures with unequal n’s and/or variances. Journal of Educational and Behavioral Statistics, 1(1), 113–125.
Good, P. I. (2005). Permutation, Parametric, and Bootstrap Tests of Hypotheses (3rd ed.). Springer.
Good, P. I. (2013). Permutation Tests: A Practical Guide to Resampling Methods for Testing Hypotheses (3rd ed.). Springer.
Gumz, A., Longley, M., Franken, F., Janning, B., Hosoya, G., Derwahl, L., & Kästner, D. (2024). Who are the skilled therapists? Associations between personal characteristics and interpersonal skills of future psychotherapists. Psychotherapy Research, 34(6), 817–827. https://doi.org/10.1080/10503307.2023.2259072
Hampel, F. R., Ronchetti, E. M., Rousseeuw, P. J., & Stahel, W. A. (1986). Robust Statistics: The Approach Based on Influence Functions. Wiley.
Hojat, M., Louis, D. Z., Markham, F. W., Wender, R., Rabinowitz, C., & Gonnella, J. S. (2011). Physicians' Empathy and Clinical Outcomes for Diabetic Patients. Academic Medicine, 86(3), 359-364. https://doi.org/10.1097/ACM.0b013e3182086fe1
Holland, J. L. (1997). Making vocational choices: A theory of vocational personalities and work environments (3 ed.). Psychological Assessment Resources.
Jones, L. M., & Huggins, T. J. (2014). Empathy in the dentist-patient relationship: Review and application. New Zealand Dental Journal, 110(3), 98 – 104.
Kruskal, W. H., & Wallis, W. A. (1952). Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association, 47, 583–621.
Kuhl, J. (2000). A theory of self-development: Affective fixation and the STAR model of personality disorders and related styles. In J. Heckhausen (Ed.), Motivational Psychology of Human Development. Elsevier Science. Elsevier Science.
Kuhl, J. (2001). Motivation und Persönlichkeit. Interaktionen psychischer Systeme [Motivation and personality. Interactions of psychological systems]. Hogrefe.
Kuhl, J., & Kazén, M. (2009). Persönlichkeits-Stil- und Störungs-Inventar (PSSI). Manual [Personality Style and Disorder inventory, PSDI] (2 ed.). Hogrefe.
Kuhl, J., & Kazén, M. (2024). Persönlichkeits-Stil- und Störungs-Inventar (PSSI). Manual [Personality Style and Disorder inventory, PSDI] (3 ed.). Hogrefe.
Leiner, D. J. (2024). SoSci Survey (Version 3.5.02) [Computer software]. Available at: https://www.soscisurvey.de.
Levene, H. (1960). Robust Tests for Equality of Variances. Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling. Stanford University Press.
Louwen, C., Reidlinger, D., & Milne, N. (2023). Profiling health professionals’ personality traits, behaviour styles and emotional intelligence: a systematic review. BMC Medical Education, 23:120. https://doi.org/10.1186/s12909-023-04003-y
Nodop, S., & Strauß, B. (2013). Mangelnde Eignung bei angehenden Psychotherapeuten. Kriterien und Umgangsmöglichkeiten aus Sicht der Institutsleiter [Lack of aptitude in prospective psychotherapists. Criteria and ways of dealing with it from the perspective of the institute directors]. Psychotherapeut, 58(5), 446–454. https://doi.org/10.1007/s00278-013-1001-9
Peter, B., & Böbel, E. (2020). Significant Differences in Personality Styles of Securely and Insecurely Attached Psychotherapists: Data, Reflections and Implications. Frontiers in Psychology, 11, Article 611. https://doi.org/10.3389/fpsyg.2020.00611
Peter, B., Böbel, E., Hagl, M., Richter, M., & Kazén, M. (2017). Personality Styles of German-Speaking Psychotherapists Differ from a Norm, and Male Psychotherapists Differ from Their Female Colleagues. Frontiers in Psychology, 8, Article 840. https://doi.org/10.3389/fpsyg.2017.00840
Peter, B., Bose, C., Piesbergen, C., Hagl, M., & Revenstorf, D. (2012). Persönlichkeitsprofile deutschsprachiger Anwender von Hypnose und Hypnotherapie [Personality styles of German-speaking practitioners of hypnosis and hypnotherapy]. Hypnose-ZHH, 7(1 + 2), 31–59. www.MEG-Stiftung.de,
Peter, B., & Wolf, T. G. (2021). Replication Studies on Significant Differences in Personality Profiles of Securely and Insecurely Attached Psychotherapists and Dentists. Frontiers in Psychology, 12, Article 662828. https://doi.org/10.3389/fpsyg.2021.662828
Peter, B., & Wolf, T. G. (2022). Personality Styles of Hypnosis-Practicing Dentists: A Brief Report. International Journal of Clinical and Experimental Hypnosis, 70(3), 314-324. https://doi.org/10.1080/00207144.2022.2097082
Richardson, J. D., Lounsbury, J. W., Bhaskar, T., Gibson, L. W., & Drost, A. (2009). Personality traits and career satisfaction of health care professionals2009. Health Care Manager, 28(3), 218–226. https://doi.org/10.1097/HCM.0b013e3181b3e9c7
Rogers, C. R. (1957). The Necessary and Sufficient Conditions of Therapeutic Personality-Change. Journal of Consulting Psychology, 21(2), 95-103. https://doi.org/10.1037/0022-006x.60.6.827
Rossetti, C., Biemann, T., & Dlouhy, K. (2025). The emergence of similar personalities in similar occupations. Journal of Organizational Behavior, n/a(n/a). https://doi.org/10.1002/job.2873
Saxon, D., Barkham, M., Foster, A., & Parry, G. (2017). The Contribution of Therapist Effects to Patient Dropout and Deterioration in the Psychological Therapies. Clinical Psychology & Psychotherapy, 24(3), 575-588. doi:10.1002/cpp.2028
Schneider, B. (1987). The people make the place. Personnel Psychology, 40(3), 437–453. https://doi.org/10.1111/j.1744-6570.1987.tb00609.x
Schwartz, B., Hehlmann, M. I., Deisenhofer, A. K., Rubel, J. A., Fischer, L., Lutz, W., & Schöttke, H. (2025). Elucidating therapist differences: Therapists' interpersonal skills and their effect on treatment outcome. Behaviour Research and Therapy, 186, Article 104689. https://doi.org/10.1016/j.brat.2025.104689
Tabachnick, B. G., & Fidell, L. S. (2019). Using Multivariate Statistics (7th ed.). Pearson.
Taubner, S., & Evers, O. (2022). Kann man Super-Shrinks ausbilden? Kompetenzentwicklung in der Psychotherapie [Can super-shrinks be trainedelliot? Competence development in psychotherapy]. Psychotherapie. doi: https://doi.org/10.1007/s00278-022-00609-7
The jamovi project. (2024). Jamovi. Version 2.6. Retrieved from https://www.jamovi.org.
University of Minnesota. (2023). College of Continuing & Professional Studies. Retrieved April 18 2025, from https://ccaps.umn.edu/story/10-must-have-characteristics-health-care-professionals; access April 18 2025
Vermeulen, M. A. A. P., Hill, J. M., van Vilsteren, B., Brandt-Hagemans, S. C. F., & van Loon, F. H. J. (2024). Personality characteristics of Dutch nurse anesthetists and surgical nurses when compared to the normative Dutch population, a quantitative survey study. Applied Nursing Research, 76. https://doi.org/10.1016/j.apnr.2024.151781
Welch, B. L. (1947). The generalization of "Student's" problem when several different population variances are involved. Biometrika, 34(1–2), 28–35. https://doi.org/10.1093/biomet/34.1-2.28
WHO. (2013). Transforming and scaling up health professionals‘ education and training: World Health Organization guidelines. . World Health Organization
Wilcox, R. R. (2012). Introduction to Robust Estimation and Hypothesis Testing (3rd ed.). Academic Press.
Wolf, T. G., Baumgärtner, E., & Peter, B. (2022). Personality styles of dentists practicing hypnosis confirm the existence of the homo hypnoticus. Frontiers in Psychology, 13. https://doi.org/10.3389/fpsyg.2022.835200

Tables 1 to 11 are available in the Supplementary Files section

No competing interests reported.

Tables.docx

Download PDF

Reviews received at journal
22 Dec, 2025
Reviews received at journal
22 Dec, 2025
Reviews received at journal
12 Dec, 2025
Reviews received at journal
09 Dec, 2025
Reviewers agreed at journal
01 Dec, 2025
Reviewers agreed at journal
01 Dec, 2025
Reviewers agreed at journal
01 Dec, 2025
Reviewers agreed at journal
29 Nov, 2025
Reviewers invited by journal
04 Sep, 2025
Editor assigned by journal
23 Jul, 2025
Submission checks completed at journal
23 Jul, 2025
First submitted to journal
19 Jul, 2025

You are reading this latest preprint version

Nature or Nurture? Have Health Care Professionals Always Been Good People or Do They Only Become So During Their Professional Career? An Empirical Study of Personality Styles in Health Care Professionals and the General Population

Status:

Version 1

Abstract

Figures

1. Introduction

2. Questions and objectives of the current study

3. Materials and methods

3.1 Survey instrument

3.2 Sample

3.3 Data analysis

4. Results

4.1 Results of the original DACH study by Peter et al. (2017)

4.2 Results of the PSDI of the HCP group vs. the normal population

4.3 Analysis of extreme T-values of the PSDI styles

4.4 Comparison with students

4.5 Exploratory and confirmatory factor analysis

4.6 Cluster analysis

5. Discussion

5.1 Personality profiles of healthcare professionals compared to the normal population

5.2 Role of extreme manifestations of the PSDI personality styles

5.3 Development of personality styles during a career

5.4 Complementary findings from factor analysis and cluster analysis

5.5 Limitations and methodological restrictions

5.6 Implications and outlook

5.7 Conclusion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1