Personality and mental health changes throughout the course of university police training in Sweden

Introduction

Police officer trainees are internationally recognized as being highly selected, relatively mature and healthy individuals (e.g., du Preez, Cassimjee, Lauritz, Ghazinour, & Richter, 2011; Ghazinour, Lauritz, du Preez, Cassimjee, & Richter, 2010). However, a relatively high percentage of them develop psychological symptoms of various kinds after some years in policing (e.g., Lee, Kim, Won, & Roh, 2016; Schwarzer, Cone, & Bowler, 2016; Stanley, Hom, & Joiner, 2016; Violanti, Andrew, Mnatsakanova, Hartley, Fekedulegn, & Burchfiel, 2016). Due to the link between exposure to traumatic situations and pathogenic consequences (Gabriel et al., 2007), police officers have been considered at high risk of developing posttraumatic stress, psychological disturbances, and poor physical health (de Terte, Stephens, & Huddleston, 2014). This heightened risk might be one of the reasons for the considerable interest in research on the police profession in recent years. Personality characteristics seem to have an impact in this respect (du Preez et al., 2011; Ghazinour et al., 2010).

The police work is complex because officers are required to play very different roles; sometimes they are fighting for their survival on the street (e.g., when confronting life-threatening situations), and, at other times, they have to emotionally support victims of crime (e.g., battered children). This dual role exists because police are usually the first caregiving professionals to arrive at a crime scene (Manzella & Papazoglou, 2014). Both before and after police training, officers are likely to experience physiological arousal before, during, and even after exposure to stressful and potentially critical or traumatic events (Andersen, Papazoglou, Arnetz, & Collins, 2015; Wills & Schuldberg, 2016). The effect of traumatic life events within the police work environment has an impact on officers work performance. A negative consequence of police work on officers’ personality created research fields such as “police personality”.

A body of research investigating police personality have been concerned with negative aspects of personality, including cynicism, authoritarianism, and psychopathology (Gerber & Ward, 2008), with some emphasizing that authoritarianism increases over time as officers enter a work culture (Genz & Lester, 1976; Niederhoffer, 1967), adopting an occupational personality (Wills & Schuldberg, 2016). Additional characteristics have been associated with police personalities as well: suspicious, solidaristic, conservative, alienated and thoroughly bigoted (Balch, 1977; Skolnick, 1977).

Watson and Sterling (1969) described police officers as possessing these attributes: pragmatism, action oriented, valuing common sense rather than theory and success more than ideas. They also noted that the tendency towards pragmatism is closely associated to cynicism. Police cynicism was defined as an attitude of “contemptuous distrust of human nature and motives (Behrend, 1980). According to Graves (1996) police cynicism is a distinct characteristic of police personality. He observed that cynicism, particularly prevalent in larger urban departments, develops as a result of burnout and stress, the emotional conditions caused largely by the excessive demands of police work (Twersky-Glasner, 2005).

Forero and colleagues (2009) conducted a structural equation model analysis based on longitudinal data to examine the predictive validity of the training process for actual performance after graduation. Their results suggest that actual job performance is indeed influenced by personality, but that this influence is mediated by training. Williams, Ciarocchi, and Dean (2010) followed 60 police trainees from recruitment to integration at workplace to determine if mindfulness, emotional awareness, and resilience predict improvement in mental health and wellbeing in the police recruits after 1 year of service. Their study indicated that mindfulness predicted depression at follow up, while emotion identification skills predicted general mental health. The authors concluded that police officers and police organisations may benefit from interventions aiming at developing and promoting mindfulness and emotion identification.

To help police academies to enhance police trainees’ coping resources related to the adversities that they will encounter in their future work, several research projects have recently been conducted in Scandinavia: in Finland, leadership in police organizations, a changing security environment and evolving security discussions (Huotari, 2011; Laitinen, 2011); in Norway, personalities and preferences for conflict resolution tactics (Abrahamsen, 2006); in Denmark, the consequences of losing local affiliations (Holmberg, 2011); and, in Sweden, police trainees’ socialization (Rantatalo, Karp, Ghazinour, & Lauritz, 2015), police trainees’ mental health and personalities (Ghazinour et al., 2010), and anger related to psychopathology, temperament and character in police trainees (Ghazinour & Richter, 2009);

Personality characteristics of police officers and their development represent important research topics because they are, for example, known to be major determinants of mental illness/health (García-Montes, Martínez, Pérez-Álvarez, & Díaz, 2005; Hansenne et al., 1999; Richter et al., 2003), coping behaviors (Carver & Connor-Smith, 2010; Richter, Lauritz, Preez, Cassimjee, & Ghazinour, 2013), and decision-making processes (Dewberry, Juanchich, & Narendran, 2013; Pohling, Bzdok, Eigenstetter, Stumpf, & Strobel, 2016; Rahaman, 2014; Schwarz, 2000).

Cloninger’s (1984) psychobiological model of personality contributes to enhance our understanding of human behavior not only from character or performance angles what several investigations did. It is assumed that applying both constructs, temperament and character in combination, can lead to comprehensive solutions to protect police staff.

Methods

Measurement

Results

Fifty two students participated in this study (67% males). Their baseline age ranged between 20 and 38 years (mean age = 25, SD = 4). Nearly half of the participants were single, , more than half (55%) had already gained a university degree; and nearly one third (31%) completed secondary school before entering police academy.

At the end of their university training, the students scored, on average, significantly lower in harm avoidance (t = 2.91; p = 0.002) and in self transcendence (t = 3.02; p = 0.002) and significantly higher in reward dependence (t = 2.25; p = 0.013), self-directedness (t = 3.77; p < 0.001), and cooperativeness (t = 3.11; p < 0.001) than Swedish individuals of the same age from the general population (Tables 1 and 2). Furthermore, they scored significantly lower on most of the psychopathological scores of the SCL-90-R (except for obsessive-compulsive and depression) than individuals from the Swedish general population, with t-test values ranging between 2.30 (p = 0.013) for somatization and 9.43 (p < 0.001) for phobic anxiety (p < 0.001 also for anxiety, paranoid ideation, and hostility) (data from the general population not presented; taken from Brändström et al., 2008; and Fridell et al., 2002, respectively).

In a multivariate analysis of variance (MANOVA), repeated measure, related to the TCI domains, a significant difference was found between the two assessments (Tables 1 and 2), which was mainly based on the significant reduction in self transcendence and a borderline significant reduction in novelty seeking More specifically, the reduction in self transcendence was based on the reductions in subscales ST3 (spiritual acceptance versus rational materialism – F = 10.57; p = 0.002) and in NS1 (exploratory excitability versus stoic rigidity – F = 3.39; p = 0.072) and NS2 (impulsiveness versus thoughtfulness – F = 6.28; p = 0.015) (mean scores for TCI subscales not presented).

Furthermore, a reduction in psychopathological symptoms was indicated by significant MANOVA results related to the SCL-90-R scores based on the significant reduction in phobic anxiety and in psychoticism (borderline significant) from assessment one to assessment two (Tables 1 and 3). Mean score for phobic anxiety upon entry was 0.08 which was reduced to 0.02 at the end of the follow-up and psychoticism reduced from 0.22 to 0.14, suggesting that students experience less phobic anxiety and less psychoticism after the training.

The variance of seven of the nine SCL-90-R scores from second assessment were significantly explained by the variation of a variety of combinations of personality scores and psychopathological scores from the first assessment (exclusively the same variables like those predicted of the second assessment were applied – e.g., depression scores from first and second assessment or in other words, for example, depression from first assessment needs to be included in the model to assess its predictability for depression at second assessment, since it is known that a score from a previous assessment usually is its best predictor of subsequent scores). This means that personality of students could predict their psychopathological symptoms. Independent from the number and type of significantly contribution variables the amount of explained variance ranged from 8% in paranoid ideation to 39% in obsessive compulsion. In multiple regression analyses, with SCL-90-R scores from the completion of university training as dependent variables and the respective SCL-90-R, temperament and character scores from entry into training as independent variables, no significant model based on the stepwise method could be found for phobic anxiety. The proportion of the variation of the response variable (SCL-90-R in 2^nd assessment) that is explained by the explanatory variables is provided as adjusted r-square. The larger the value of adjusted r-square, the better the set of explanatory variables predicts response variable. For example variation in somatic symptoms at first assessment can explain 18% of variance in somatic symptoms of the 2^nd assessment. The unexplained percentage of 82% could be explained by variation in not measured variables and measurement errors. However, the variance of the independent variables from the first assessment could explain substantial variance in the other SCL-90-R scores from the completion of training—between 13% (psychoticism from the second assessment by psychoticism from the first assessment) and 42% (somatic symptoms from the second assessment by anxiety, hostility, and somatic symptoms from the first assessment), with harm avoidance as the personality characteristic that often significantly contributed to the findings (for obsessive-compulsive, interpersonal sensitivity, depression, anxiety, and hostility, see Table 4).

The variance in corresponding SCL-90-R scores from the first assessment alone could explain between 6% (interpersonal sensitivity) and 26% (obsessive-compulsive) of the variance in the scores from the second assessment. The variance in harm avoidance , the only significant temperament score from the first assessment remaining in the regression equation could explain substantial amount of variance in five of the psychopathological scores (SCL-90-R subscales) from the second assessment : obsessive compulsive, anxiety, interpersonal sensitivity, depression and hostility. This means students with higher harm avoidance, experience higher level of obsessive compulsive, anxiety, interpersonal sensitivity, depression and hostility. Whereas variance in character scores only explained an additional amount beyond the SCL-90-R from first assessment and the harm avoidance score for depression (-self-directedness explained 5%) and obsessive-compulsive (-cooperativeness explained 7%). This means on top of baseline harm avoidance and SCL-90-R, self-directedness explained depression at 2^nd assessment additional 5% and cooperativeness explained obsessive compulsive symptoms additional 7%.

The changes in personality characteristics between the two assessments have been significantly correlated with changes in psychopathological symptoms and with changes in character scores self-directedness (negative – highest with psychoticism, r=–0.38) and self -transcendence (positive – highest with anxiety, r=+0.52).Changes in self-directedness were significantly negative associated with changes in five of the nine SCL-90-R scores (obsessive compulsive (r=–0.33), depression (r=–0.28), anxiety (r=–0.32), hostility (r=–0.28) and psychoticism (r=–0.38). This means students with improved self-directedness at second assessment experience less mental health symptoms, compare to their baseline status. Substantial positive associations were observed for changes in self transcendence with regard to changes in four of nine SCL-90-R scores (depression: r=+32, anxiety: r=+0.52, hostility r=+0.35, phobic anxiety r=+0.38). Changes in self-directedness and self transcendence were the most important, as indicated by their substantial associations with changes in five and four, respectively, of the nine SCL-90-R scores. Changes in harm avoidance were significantly positively associated with four subscales – highest with somatic symptoms and psychoticism), novelty seeking (negatively associated with obsessive-compulsive and phobic anxiety) and cooperativeness (negatively associated with hostility and phobic anxiety when theoretically predicting a substantial association. However, changes in reward dependence and persistence did not significantly relate to changes in any of the SCL-90-R scales (see Table 5).

Discussion

In our present study, changes in personality characteristics and general health among Swedish police trainees were investigated in the course of their two-year university training. Furthermore, we analyzed time-lagged relationships between personality characteristics and mental health and the associations between changes in personality characteristics and changes in mental health during this period.

Our findings indicate that there is a tendency towards a small reduction in one temperament dimension, novelty seeking (mainly based on a reduction in subscale 2: impulsiveness versus reflection), and a more substantial reduction in self transcendence (mainly due to a reduction in subscale 1: self-forgetfulness). Seemingly, Swedish police trainees learned to focus on long-term goals and to act with somewhat more appropriate levels of deliberation, which often may be useful in their daily practices when they are trying to keep calm, even in difficult situations, to help them make thoughtful decisions. These findings might indicate a slightly increased reflective capacity of the police trainees, which represents one of the core issues in police training.

Therefore, increased self-insight and social learning processes might trigger and inform such changes (Caspi & Roberts, 2001). One other explanation to such changes might depend on task oriented training as problem based learning pedagogical modules. Police trainee’s take part of number of complex situations where they have to act, organize and cope with stress. One example between many of such a complex training module is a scenario where police trainee has to carry on a report on a car accident and in the same time get calm angry people and also protect people meanwhile keep in contact with the communication central.

Our results confirm associations between different scores for the personality characteristics and the psychopathological syndrome scores of the SCL-90-R with regard to the impact of several temperament and character traits on a wide spectrum of psychopathological syndromes, thereby confirming previous findings (Conrad, Geiser, Kleiman, Zur, & Karpawitz-Godt, 2014; du Preez et al., 2011). Harm avoidance and self-directedness have been found to be the most closely related (positively and negatively, respectively) to a wide spectrum of psychopathological symptoms, confirming many findings from previous research (Rettew & McKee, 2005). As these associations could even be established in a sample of very healthy and highly mature individuals (substantially lower SCL-90-R scores and higher self-directedness and cooperativeness scores as indicators of maturity compared with individuals from the Swedish general population; see Ghazinour et al., 2010), we suggest that these relationships can be described as universal given their independence from an individual’s level of psychopathology or maturity.

Furthermore, self-transcendence decreased over the course of police training, mainly because of a decrease in the subscale ST1 (self-forgetfulness versus self-consciousness). Through police training and its intensive focus on facts and analysis, police trainees’ creative minds moved along this dimension, and they tended to be more aware of their individuality or concentrating on their work (Cloninger et al., 1994). Possibly, the combination of being lectured by skilled police officers; the training in self-awareness as well as improving self-confidence, and mindfulness techniques might have contributed to the found differences in personality characteristics, psychopathological symptoms and associations between both constructs over time. However, most personality characteristics and psychopathological symptoms are seemingly stable and rather unchanged. This can be due to the highly selected participants in terms of highly mature and healthy individuals who decided for a police officer career. Alternatively, the theoretical assumption of the trait-like nature of personality characteristic in Cloninger’s personality theory might have been confirmed by the results. Additionally, the relative short follow-up period of two years might represent another reason for the found relative stability of most scores.

Several limitations of our present investigation have a negative impact on the generalizability and interpretability of the findings. The very small sample size and its high retention rate between the two assessments indicate this study’s explorative character, which was the main reason that corrections were not applied for multiple comparisons in the data analysis. Another limitation is the study’s exclusive use of self-report questionnaires. However, the use of internationally established assessment methods provides the opportunity for a wide range of comparisons between our results and other findings in the literature. The sample can be criticized for not being adequately representative to analyze the relationships between personality and psychopathological phenomena. Interestingly, representativeness does not usually occur when investigating such associations in clinical populations. We claim that the use of a sample from the other ‘positive’ end of psychopathological symptoms, combined with mature personality characteristics, provided us the opportunity to make conclusions regarding the generalizability of these associations across humankind.

The recruitment of police trainees based on a rigorous selection criteria and also our findings might provide evidence for the fact that police officers burn out and stress are much due to the line of the duty. This observation addresses previous assumption about police personality.

Despite a number of methodological limitations, we conclude that, over the course of two years, police trainees’ personality characteristics have shown some minor changes that seemingly fit the demands of policing and are potentially valuable with regard to the trainees’ future careers. Personality characteristics are predictors of mental health even in a highly selected sample of mature and healthy individuals like Swedish police trainees.