Affective personality type, post-traumatic stress disorder symptom severity and post-traumatic growth in victims of violence
The current study explored the differential association between affective personality type, post-traumatic stress disorder (PTSD) symptom severity, and post-traumatic growth (PTG) in victims of violence (N = 113). Relying on previous research, median cut off-scores on the Positive and Negative Affect Schedule Short Form were used to classify participants as high affective [i.e. high positive affectivity (PA) and high negative affectivity (NA)], self-actualizing (i.e. high PA and low NA), self-destructive (i.e. low PA and high NA) and low affective (i.e. low PA and low NA). Results indicated that the self-destructive and high affective personality styles were strongly associated with increased PTSD symptoms severity. High affective personality type was found to be the only significant predictor of PTG. Results, study limitations and directions for future research were discussed. Copyright © 2010 John Wiley & Sons, Ltd.
An extensive body of research exists regarding the psychosocial aftermath of violent victimization. Many studies have shown that victims of interpersonal violence suffer from psychological distress, such as post-traumatic stress disorder (PTSD) (e.g. Kunst, Winkel, & Bogaerts, in press; Orth, Cahill, Foa, & Maercker, 2008). Previous research suggests that the prevalence of violence-related PTSD approximates 10–15 per cent (see Winkel, 2007). However, individual studies have recorded much higher prevalence rates, with percentages as high as 25 per cent amongst victims recruited through police stations (Wohlfarth, Winkel, & Van den Brink, 2002) and 32 per cent in emergency department samples (Birmes et al., 2003). Other outcomes observed in victims of violence include: symptoms of depression (e.g. Kilpatrick et al., 2003), somatization, hostility, anxiety and phobic anxiety (e.g. Norris & Kaniasty, 1994; Winkel, 2009), decreases in general well being (Denkers, 1996; Denkers & Winkel, 1997) and eating disorders (Brady, 2008).
Along with these negative outcomes, several studies have reported on the positive side of coping with violent victimization (e.g. Borja, Callahan, & Long, 2006; Cobb, Tedeschi, Calhoun, & Cann, 2006), particularly experiences of post-traumatic growth (PTG)—a state of functioning superior to that which existed prior to trauma exposure (e.g. Linley & Joseph, 2004; Tedeschi & Calhoun, 2004; Zoellner & Maercker, 2006). PTG comprises varying domains of personal change, including changes in perception of the self, changes in interpersonal relationships, and changes in philosophy of life (Calhoun & Tedeschi, 1999). The current state-of-the-art seems to suggest that PTG and manifestations of distress following traumatic events are only marginally correlated and thus represent rather separate, independent dimensions of psychological functioning (Zoellner & Maercker, 2006). In other words, PTG is not merely an indicator of well-being.
Individual differences in trait affectivity have been suggested to underlie both adverse and beneficial outcomes of traumatization (e.g. Affleck & Tennen, 1996; Erbes et al., 2005; Frederickson, Tugade, Waugh, & Larkin, 2003; Tedeschi & Calhoun, 1996). NA seems to play an important role in the onset and maintenance of emotional problems following violent victimization (e.g. Kunst, Bogaerts, & Winkel, 2009; Mikkelsen & Einarsen, 2002; Zoellner, Goodwin, & Foa, 2000). NA involves the stable tendency to experience negative emotions (e.g. Watson & Clark, 1984), negative self-evaluations, and over-sensitivity to adverse stimuli (e.g. Watson & Pennebaker, 1989). When exposed to potentially stressful events, people with high levels of NA may be assumed to respond more intensely than others (e.g. Zeidner, 2006). Furthermore, if levels of heightened distress do not automatically resolve within a reasonable amount of time, they are at an increased risk of developing psychiatric disorder. Previous research, although preliminary in nature, indicates that NA is not associated with PTG (e.g. Tedeschi & Calhoun, 1996, 2004; Sheikh, 2004). By contrast, PA reflects one's ability to maintain a positive out-come both over time and across various situations (Cropanzano et al., 1993; Watson, Pennebaker, & Folger, 1987) and has been suggested to act as a determinant of PTG (Park, Cohen, & Murch, 1996; Tedeschi & Calhoun, 2004). PA and NA broadly correspond with the Big Five traits extraversion and anxiety/neuroticism, respectively (Watson, Clark, & Tellegen, 1988). Generally stated, NA and PA seem to be rather weakly correlated with each other (Thompson, 2007; Watson et al., 1988).
Although earlier studies have shown that both NA and PA play an independent role in the process of trauma resolution, research on the explanatory power of configurations of high and low NA and PA values is scarce. Such combinations enable the identification of personality subtypes or cognitive-emotional presymptom profiles (Archer, Adolfsson, & Karlsson, 2008) that are differently related to both positive and negative psychological outcomes (Denollet, 2000; Van Yperen, 2003). Norlander, Bood, and Archer (2002) developed a procedure to create four affective personality styles based on the Positive and Negative Affect Schedule (PANAS; Watson et al., 1988): high affective (i.e. high PA and high NA), self-actualizing (i.e. high PA and low NA), self-destructive (i.e. low PA and high NA), and low affective (i.e. low PA and low NA) personality. In a study of 46 grocery store employees and 44 flying squad policemen (Norlander, Von Schedvin, & Archer, 2005) who had experienced a wide array of negative life events, they investigated, amongst other things, differences in experienced stress, as measured by the Stress and Energy Scale (Kjellberg & Iwanowski, 1989) and PTG between the four personality types. Based on two earlier studies (Bood, Archer, & Norlander, 2004; Norlander, Bood, & Archer, 2002), they hypothesized that self-actualizing individuals would report the lowest levels of distress and the highest levels of PTG. Self-destructive persons were assumed to experience they highest levels of distress and the lowest levels of PTG. In partial support of their expectations, they found that the highest levels of distress were experienced by self-destructive and high affective persons, whereas self-actualizing and low affective persons reported the lowest stress scores. Contrary to expectations, the high affective group scored highest on PTG. An intermediate response was found for the self-actualizing group. The lowest scores were observed for the self-destructive and low affective groups. Presumably, and in line with the view held by several PTG scholars that a certain degree of distress is a prerequisite for PTG to occur (e.g. Tedeschi & Calhoun, 1995, 2004), PA is most likely to induce growth if one is vulnerable to experience distress in the first place. More precisely, the stress high affective individuals are likely to experience in response to trauma exposure is necessary to trigger the process of growth.
Given the aforementioned, the primary purpose of the current study was to explore associations amongst affective personality type, PTSD symptom severity, and PTG in victims of interpersonal violence. To our knowledge, this has never been done before. Building on the observations made by Norlander, Von Schedvin, et al., (2005), it was expected that the self-destructive and high affective personality styles would be associated with increased PTSD symptom severity, while the latter was also expected to be positively associated with PTG. For the self-actualizing group, an association with PTG was expected as well.
The current study was part of a larger study into the psychosocial aftermath of violent victimization (Kunst, Bogaerts, Wilthagen, & Winkel, 2010; Kunst et al., in press). Participants were recruited through the Dutch Victim Compensation Fund (DVCF). Inclusion criteria were: age ≥18; filing a claim during the first quarter of 2006; and no missing file data on age, gender, and date of crime (cf. Kunst et al., in press). All victims eligible for participation were invited to fill out a set of internet questionnaires on affective personality type, PTSD symptom severity, and PTG in October 2007. Those who did not have access to the World Wide Web or preferred to complete the measures by pencil could request for a paper version. Background variables (age, gender, time relapse since victimization, compensation level for pain and suffering, and type of violence: sexual violence, severe physical assault, moderate physical assault, theft with violence and other1) were retrieved from victims' electronic files. Compensation level for pain and suffering was mainly used as an indicator of the objective severity of the violence experienced and ranged from 0 to 8. Approval for the study was obtained from the DVCF Committee. Unfortunately, reminders were not allowed to be sent to non-respondents.
In total, 641 victims were approached for participation. One hundred and thirty-four (20.1%) of them responded. Twenty-one of them were dropped from statistical analyses due to missing data. Those participating in the study did not differ significantly from non-participants on any of the background variables included in the study. The study sample contained 52 (46%) men and 61 women (54%). Their average age was 40.9 (SD = 14.4) years. Mean time since victimization was 4.9 (SD = 3.6) years. Mean compensation levels for pain and suffering was 2.2 (SD = 1.8). Thirteen (11.5%) participants were victims of sexual violence, 12 (10.6%) of severe physical assault, 32 (28.3%) of moderate physical assault, and 30 (26.5%) of theft with violence. The remainder of the sample had experienced another type of violence.
NA and PA
NA and PA were measured by the NA subscale of the 10-item PANAS Short Form (PANAS-SF; Thompson, 2007). The PANAS-SF is derived from the original 20-item PANAS version (Watson et al., 1988). The PANAS has been validated in The Netherlands by Peeters, Ponds, & Vermeeren (1996). The Dutch version of the 20-item PANAS was used to select the items that correspond to the English version of the PANAS-SF. Both NA and PA are assessed by five items and need to be rated on a five-point likert scale (1 = never, 5 = always). Each item consists of one word describing a particular emotion or feeling. ‘Nervous’ and ‘determined’ are examples of the NA and PA subscales, respectively. In accordance with Thompson (2007), participants were instructed to indicate to what extent they generally feel a particular emotion to measure trait affectivity. The psychometric properties of the PANAS-SF have been found to be acceptable in many different language environments (Thompson, 2007). Internal consistency reliability of the NA subscale of the PANAS-SF was Cronbach's α = 0.82 and Cronbach's α = 0.63 for the PA subscale. Deletion of one item (‘alert’) would have increased its reliability to 0.70. It was decided to retain the PA subscale in its original format though, for internal reliability levels of <0.7 are acceptable for scales containing only a few items. Furthermore, the value for the mean inter-item correlation of the NA items in this study (0.24) lied within the range of 0.2–0.4 proposed by Briggs and Cheek (1986). Following on the procedure developed by Norlander et al. (2002), a variable representing the four affective personality styles was created by median split dichotomization. In addition, four categorical variables were created for each personality type: high affective (PA ≥ 18 and NA ≥ 14), self-actualizing (PA ≥ 18 and NA < 14), self-destructive (PA < 18 and NA ≥ 14) and low affective (PA < 18 and NA < 14). Those who fulfilled the criteria for a particular personality profile were coded as 1 and those who did not as 0.
PTSD symptom severity
The Dutch version of the PTSD Symptom Scale, Self-Report version (PSS-SR; Arntz, 1993; Foa, Riggs, Dancu, & Rothbaum, 1993) was used to measure PTSD symptom severity. The PSS-SR has often been used as a screening instrument for PTSD symptomatology amongst victims of crime (e.g. Andrews, Brewin, Rose, & Kirk, 2000; Dunmore, Clark, & Ehlers, 1999; Rose, Brewin, & Kirk, 1999). For each of the 17 items, respondents had to indicate to what extent they had experienced the corresponding symptom during the past week on a four-point Likert scale (0 = never, 1 = once, 2 = 2–4 times, 3 = 5 times or more). An example of the PSS-SR is ‘Upsetting images or thoughts about the event’. The psychometric properties of the PSS-SR have been found to be satisfactory in crime victim samples (Foa et al., 1993; Wohlfarth, Van den Brink, Winkel, & Ter Smitten, 2003). In the current study, internal consistency reliability of the PSS-SR was Cronbach's α = 0.95.
Perceptions of PTG were measured by the Dutch version of the PTG Inventory (PTGI; Jaarsma, Pool, Sanderman, & Ranchor, 2006). The PTGI was originally developed by Tedeschi and Calhoun (1996) and assesses five domains of PTG: relating to others, new possibilities, personal strength, spiritual change, and appreciation of life. Each item needs to be answered on a six-point Likert scale (0 = not at all, 5 = extremely). An example of a PTGI item is ‘I changed my priorities about what is important in life’. Both subscale and total scale scores can be calculated for statistical analysis. For the current study, only the PTG total score was used. Participants were instructed to report PTG levels related to the act of violence leading to the request for compensation. The psychometric properties of the PTGI have been well established in victims of violence (e.g. Cobb et al., 2006; Grubaugh & Resick, 2007; Updegraff & Marshall, 2005). Internal consistency reliability for the PTG in the present study was α = 0.96.
To explore the association between PTSD symptom severity and PTG in the current study, the correlation between the PSS-SR and PTG total scores was computed. Two one-way ANOVAs were performed to determine whether the four affective personality groups differed in their mean PSS-SR and PTGI total scores. Tukey LSD post hoc tests were performed to locate observed differences. Next, two multivariate regression analyses were performed to examine the independent contributions of each personality type to the variance in PTSD symptom severity and PTG when adjusting for background variables. In each model, background variables were entered on the first step. On the second step, three of the four categorical personality variables were added to the model. When estimating their associations with PTSD symptom severity (model 1), low affective personality style served as reference category. Self-destructive personality type was used as reference category in the second model. Selection of reference categories followed on the hypotheses presented in the introduction. Prior to regression analyses, data were checked for underlying assumptions (i.e. multicollinearity, outliers, normality, linearity, homoscedasticity and independence of residuals). The alpha level was set at 0.05 in all statistical tests. Analyses did not adjust for multiple testing, as this is not required for exploratory studies (Bender & Lange, 2001). All statistical analyses were performed using the software package SPSS 16.0 for Windows (SPSS Inc., Chicago, IL, USA).
PSS-SR and PTG total scores were not significantly correlated with each other (r = 0.12, p = ns). Thirty participants could be classified as high affective individuals, 34 as self-actualizing, 20 as low affective, and 29 as self-destructive. ANOVAs revealed overall effects for affective personality type on PTSD symptom severity, F(3, 109) = 27.97, p < 0.001, and PTG, F(3, 109) = 2.92, p < 0.05. Post hoc analyses indicated that self-destructive individuals (M = 29.1, SD = 12.1) reported higher symptom levels than high affective (M = 23.0, SD = 9.6, p < 0.05), self-actualizing (M = 8.6, SD = 10.9, p < 0.001), and low affective (M = 8.2, SD = 8.3, p < 0.001) participants. Mean PTG scores for the high affective group (M = 50.7, SD = 27.4) were higher than for the low affective (M = 33.5, SD = 27.9, p < 0.05) and self-destructive (M = 33.4, SD = 22.9, p < 0.025) groups. High affective individuals also reported higher PTG scores than those in the self-actualizing group (M = 46.5, SD = 30.2), although they did not differ significantly from each other. Regression analyses yielded significant positive associations between the self-destructive and high affective personality types and PTSD symptom severity. None of the other variables included in the model was significantly related to PTSD symptom severity (Table I). High affective personality type was the only factor independently related to higher PTG levels (Table II), although a marginal positive effect for self-actualizing personality was found as well (p = 0.08).
Table I. PredictingPTSDsymptomseverity(N = 113)
|Step 1|| ||0.138*|
| Age||0.20*|| |
| Gender||−0.07|| |
| Time since victimization||0.19*|| |
| Sexual violence||−0.04|| |
| Physical assault (severe)||0.12|| |
| Physical assault (minor)||−0.17|| |
| Robbery||0.08|| |
| Compensation level||−0.10|| |
|Step 2|| ||0.341**|
| Age||0.11|| |
| Gender||0.10|| |
| Time since victimization||0.03|| |
| Sexual violence||0.00|| |
| Physical assault (severe)||0.06|| |
| Physical assault (minor)||−0.12|| |
| Robbery||0.00|| |
| Compensation level||−0.07|| |
| Self-destructive||0.63**|| |
| High affective||0.46**|| |
| Self-actualizing||0.00|| |
Table II. Predictingpost-traumaticgrowth(N = 113)
|Step 1|| ||0.065|
| Age||0.06|| |
| Gender||−0.02|| |
| Time since victimization||−0.11|| |
| Sexual violence||0.16|| |
| Physical assault (severe)||0.10|| |
| Physical assault (minor)||−0.03|| |
| Robbery||−0.13|| |
| Compensation level||0.02|| |
|Step 2|| ||0.074*|
| Age||0.05|| |
| Gender||−0.03|| |
| Time since victimization||−0.07|| |
| Sexual violence||0.16|| |
| Physical assault (severe)||0.14|| |
| Physical assault (minor)||−0.01|| |
| Robbery||−0.09|| |
| Compensation level||0.03|| |
| High affective||0.27*|| |
| Self-actualizing||0.22|| |
| Low affective||−0.01|| |
The present study investigated self-reports of PTSD symptom severity and PTG with regard to affective personality type in a sample of victims of rather heterogeneous categories of interpersonal violence. It intended to build on the work provided by Norlander and colleagues (Bood et al., 2004; Norlander, Johansson, et al., 2002; Norlander, Von Schedvin, et al., 2005), who used PANAS median scores to create four configurations of affective personality: high affective, self-actualizing, low affective, and self-destructive.
In line with many previous studies, PTSD symptom severity and PTG were not correlated with each other. In support of our expectations, and in line with Norlander, Von Schedvin, et al. (2005), results showed that the self-destructive and high affective personality styles were strongly associated with increased PTSD symptom severity in multivariate regression analyses. Mean symptom score values for individuals in these groups even suggested that individuals in these groups are likely to develop symptom levels that lie well beyond the cutoffs proposed to qualify for probable diagnosis of PTSD (Wohlfarth et al., 2003). Also in accordance with Norlander, Von Schedvin, et al. (2005), high affective individuals reported the highest levels of PTG. Moreover, high affective personality type was found to be the only significant predictor of PTG when adjusting for background variables, while self-actualizing personality type was, contrary to expectations, only marginally associated with PTG.
The results further underline the importance of considering configurations of personality traits when studying the psychological aftermath of trauma rather than solely focussing on single personality dimensions. Admittedly, on the one hand, they seem to indicate that NA is a correlate of PTSD symptom severity irrespective of concurrent PA. However, on the other hand, they suggest that PTG will particularly occur if PA is accompanied by NA. The latter is in line with the notion that a certain level of distress is required to enable the experience of PTG (e.g. Tedeschi & Calhoun, 1995, 2004) and may prove to be an important starting-point for future research into the relationship between PTG and PTSD.
As mentioned in the introduction, currently no consensus exists regarding the relationship between PTG and PTSD. Several explanations have been proposed for this lack of consistency (Helgeson, Reynolds, & Tomich, 2006). Most of them refer to the different designs employed by previous studies, including the operationalization of PTG, the types of trauma under investigation, and the failure to test for moderation by third variables (Helgeson et al., 2006; Hobfoll et al., 2007; Zoellner & Maercker, 2006). Based on the current study's results, one might argue that the dominant affective personality style within a particular sample may also determine whether both outcomes are correlated or not. For example, a significant positive correlation would be expected when most study participants qualify as high affective individuals, because in that case most participants combine high PTSD with high PTG scores.
With the exception of a study by Norlander, Von Schedvin, et al. (2005), the finding that self-actualizing individuals reported equally low levels of PTSD symptomatology with participants in the low affective group seems to contrast with several previous studies conducted in non-traumatized samples. Archer, Adrianson, Plancak, and Karlsson (2007), for example, found that the self-actualizing group reported significantly lower levels of anxiety and depression on the Hospital and Anxiety Scale (Zigmund & Snaith, 1983) than each of the other three groups. Similar results were reported by Karlsson and Archer (2007, in press). Slightly different results were reported by Garcia and Siddiqui (2009), who found that self-actualizing individuals reported the highest levels of psychological well-being on a short version of Ryff's Measurement of Psychological Well-Being (Clarke, Marschall, Ryff, & Wheaton, 2001), although they failed to observe a significant difference between self-actualizing and high affective participants. Presumably, under non-stress conditions low affective individuals report higher levels of distress/lower levels of general well-being than those characterized by self-actualizing personality, while equally low symptom levels are reported in response to trauma.
Although highly speculative, one explanation for the incongruence of the current study's results with those found in other studies is that both self-actualizing and low affective subjects are characterized by high internal locus of control and low external control, whereas high affective and self-destructive individuals score low on internal and high on external control (Archer et al., 2008). Internal locus of control involves the belief that forces shaping one's life are largely within one's control, whereas external locus of control comprises the degree to which one conceives the outcome of an event as uncontrollable and due to luck or chance (Rotter, 1966, 1975). Particularly in the aftermath of trauma, when the adverse impact of the event needs to be overcome, high internal and low external locus of control may be assumed to protect against unsuccessful coping (e.g. Bisson, 2007; Hoge, Austin, & Pollack, 2007), possibly irrespective of affective personality type. By contrast, in the absence of trauma exposure or when external stimuli are not salient enough to trigger the autonomic stress response, perceptions of controllability are perhaps not always capable of predicting psychological distress above and beyond dispositional affectivity. Unfortunately, these contentions seem to have gone largely untested in the existing literature. To address this topic in a methodologically proper manner, future studies will need to employ a design that allows exploring the main effects of trauma exposure, affective personality type, and locus of control on PTSD and the interactions between (severity of) trauma exposure and affective personality type/locus of control. Preferably, they also account for religious believes about the controllability of life events (cf. Göral, Kesimci, & Gençöz, 2006; Karanci & Erkam, 2007).
When interpreting the study's results several limitations must be considered though. Firstly, the cross-sectional nature of the collected data does not allow interpretation of the results in terms of cause and effect. Secondly, due to the rather low response rate, findings may not be generalizable to the population of victims applying for compensation. However, at first sight this possibility does not seem to be very likely, since no differences were observed between participants and non-participants on a wide array of background variables. Thirdly, the specific focus on victims applying for state compensation prevents generalization to the general population of victims of violence and other populations. Fourthly, results may have been biased by participants' interest in compensation (cf. Frueh et al., 2003). This was not deemed very likely though, as all of them had received a final decision on their claim.
Despite these limitations, the study was the first to assess the associations between affective personality type, PTSD symptom severity, and PTG. Its results raised several issues which deserve further clarification. In addition to the propositions made above, a topic that may be addressed in future research is how the four affective personality types relate to other concepts. Given their theoretical similarities, it would be worthwhile to determine whether high affective personality overlaps with borderline personality traits. A recent study by McCormick and colleagues (2007), for example, suggests that average PANAS scores for patients diagnosed with borderline personality disorder are high for both NA and PA. Another issue that may be addressed is the similarities between low affectivity and the concepts of apathy and emotional indifference (Peterson & Janssen, 2007) and type C characteristics. Type C individuals deliberately suppress negative emotions and respond passively in the face of distress (Temoshok, 1987). Consequently, one might argue that Type Cs are likely to report low levels of PA and NA on the PANAS. Finally, in line with previous research (e.g. Klein & Shih, 1998), the association between self-destructive personality type and depressive or dysthymic personality traits deserves to be scrutinized.
The author thanks the Dutch Victim Compensation Fund for their financial and organizational support with the data collection.
The DVCF categorizes type of violence according to their legal classification used in the Dutch Penal Code (DPC). To enable statistical testing, the number of different categories was reduced from 30 to 5 (cf. Kunst et al., in press). Severe and moderate physical assault and theft with violence corresponded to the original file categorization. Sexual violence included all individuals that had experienced an offence falling under Book 2, Title XIV of the DPC. The remainder of the sample is a mixture of offences that were too low in number to form a category of their own.