The Subjective Experience of BPD Patients in the Punitive Parent Mode after a Failure Induction: A Qualitative Analysis
Daniëlle Bouma
Abstract
Borderline personality disorder (BPD) is characterized by a pattern of instability on multiple terrains. Young et al. (2005) developed the concept of schema-modes, hypothesized as dramatic shifts in emotions, cognitions and behaviors. Quantitative studies showed that BPD is associated with the Punitive Parent Mode (PPM) in which the patient internalizes a critical and punishing parental voice after, for instance, personal failure. Young’s PPM includes many assumptions about subjective experience. However, quantitative methods seem limited in measuring subjective experience. Therefore, this study uses semi-structured interviews to assess the subjective experience of BPD patients during the PPM. The experience of failure, that is expected to activate the PPM in PD patients, is induced with a manipulated,
computerized task in competing teams. The responses were subjected to a qualitative bottom-up analysis and the emergent themes were compared to Young’s PPM. The subjective experience of BPD-PPM patients is discussed and compared to the experience of AVPD-PMM patients and non-patients. Indications of other schema modes are also shortly discussed.
Borderline personality disorder (BPD) is characterized by a pattern of instability in affects, interpersonal relationships, self-image and marked impulsivity (DSM-IV; American Psychiatric Association, 1994). This instability on multiple terrains appears mostly in late adolescence or early adulthood. Since cognitive behavioral interventions are generally not sufficiently effective for treating BPD, Young, Klosko and Weishaar (2005) developed a treatment called Schema Therapy (ST). ST is based on the assumption that
individuals with personality disorders (PDs) developed early maladaptive schema’s (EMS). These EMS are conceptualized as broad, pathological patterns or themes, that consist of memories, emotions, physical sensations and cognitions, with regard to oneself and others. However, whereas EMS are conceptualized as stable constructs, BPD is characterized by rapid, dramatic emotional shifts. Therefore, Young et al. (2005) expanded ST theory with the concept of schema modes, to capture these emotional shifts during
treatment. Schema-modes are states that were experienced as a child and progressed into adulthood, leading to dramatic shifts in emotional states later on in life. Schema-modes are defined as organized patterns of emotions, cognitions and behaviors based on specific schemas, in combination with a specific form of coping, that are activated by events that relate to these aversive childhood experiences.
Research into the existence of schema-modes within PDs is still scarce. Young et al. (2005) proposed that BPD is related to four schema-modes, respectively the Detached Protector, Punitive Parent, Vulnerable Child and Angry Child mode. The Detached Protector mode is conceptualized as a state of
flat affect, emotional detachment and avoidance of intimacy to protect oneself from painful feelings. The
Punitive Parent mode is characterized by the internalization of the critical and punishing parental voice. The Vulnerable Child mode refers to a vulnerable, overwhelmed state that is accompanied by painful feelings,
such as anxiety, sadness and hopelessness. The Angry Child mode is known by feelings and expressions of
The emotions, cognitions and behaviors that characterize Young’s schema-modes have been translated into questionnaires to assess these schema modes (Young et al., 2007; Klokman, Arntz &
Sieswerda, 2001; Lobbestael et al., 2010). Studies that used quantitative measures to assess the relationship between BPD and schema-modes, indicated that BPD is indeed associated with these four schema-modes (Arntz, Klokman & Sieswerda, 2005; Lobbestael, Arntz & Sieswerda, 2005; Lobbestael et al., 2010). Moreover, Arntz et al. (2005) showed that individuals with BPD score higher on the Detached Protector Mode after watching BPD specific emotional filmclips, than before, indicating that the stress induction activated it.
All four schema-modes indicate distress to some extent, however, the Punitive Parent mode (PPM) in particular is associated with self-injury (Saldias, Power, Gillanders, Campbell & Blake, 2013) and suicide (Van Genderen & Arntz, 2010) and therefore the clinical relevance of research into this schema-mode is high. Young et al. (2005) proposed that, with respect to the PPM, individuals with BPD internalized the punishing stance of their parents or other caretakers. This internalized parental voice is harsh, critical, punitive and blaming for failures or the mere expression of needs, emotions and opinions. During this mode, the patients seem to perceive themselves as bad, to blame, evil, defective and deserving punishment. Moreover, emotions related to this schema mode are shame, guilt and self-directed anger. The latter is one expression of self-punishment, but punishment also becomes apparent in more concrete behaviors, such as subjecting the self to starvation or self-injury. During the PPM, individuals seem incapable of forgiving themselves for their failure to meet certain expectations, even if this lacks realism. Therefore, based on Young’s theory, the experience of failure is expected to activate this punishing and critical PPM (Young et al., 2005). Prior research in a non-patient population showed that the experience of failure is also related to emotional distress (Brown, Cai, Oakes & Dang, 2009), but non-patients seem to attribute their failure to external or modifiable causes, instead of their own ability (Hawi, 2010). Therefore, the PPM seems to differ from a normal self-critical stance.
Even though prior studies showed that quantitative measures of the PPM discriminate well between patients with BPD and healthy controls (Klokman et al., 2001; Lobbestael et al., 2005; Lobbestael et al., 2010; Arntz et al., 2005), Young’s schema-mode conceptualization includes many assumptions about the subjective experience of the patient (Young et al., 2005). However, quantitative methods are limited in measuring subjective experience, since the researcher translates his own theoretical assumptions into a questionnaire (Silverman, 2011). Moreover, the format in which questionnaires are presented ‘forces’ people to give a response, within a limited range of response-options. Therefore, the ‘a priori’ construction of these questionnaires, might also miss out on important information.
To investigate the subjective experience in the PPM, a qualitative method seems more appropriate to look into the full range of emotions, thoughts and behaviors related to this schema-mode. Qualitative research is generally inductive in nature (Bitektine, 2008). With a deductive approach one might overlook the themes that are brought forward by the participants themselves. Therefore, this study adopts a qualitative inductive approach, where emergent themes (Pietkiewicz & Smith, 2014), the themes that result
from the data, will be compared to Young’s conceptualization of the PPM.
As previously discussed, the experience of failure is expected to activate the PPM (Young et al., 2005). Moreover, Arntz and Jacobs (2011) propose that a subtle way to induce this schema mode, is to imply guilt, while omitting explicit punishing reactions. Therefore, this study involves a manipulated, computerized task in a team, that induces the experience of failure and more specifically, the experience of failing others. Following the task, the subjective experience of this team failure, to which they can be held accountable, is questioned in a semi-structured in-depth interview. The aim of this study is to uncover the subjective experience of BPD patients in the PPM and compare these experiences to the experience of non-patients and patients with an Avoidant Personality Disorder (AVPD). The latter group is also
associated with the PPM (Bamelis, Renner, Heidkamp & Arntz, 2011) and comparisons between the PPM in BPD and AVPD could therefore provide insight into the specificity of the PPM in BPD.
Method Participants
Both BPD patients and AVPD patients were recruited through several mental health care
institutes, such as “De Viersprong”, “Pro Persona” and “Arkin NPI”. Exclusion criteria for the BPD and AVPD-condition were: comorbid bipolar disorder, dissociative identity disorders, psychotic disorders, clinical detoxification, attention-deficit/hyperactivity disorder and immediate suicide risk. Other comorbid Axis-I disorders, Axis-II disorders and medication (if the doses were kept stable during treatment) were allowed.
The non-patients were recruited from the general population. Inclusion took place according to a matching principle, based on age (between 18-55 year), gender (female) and educational level
(MBO/HBO/WO). Non-patients who met more than 2 criteria for an Axis-I or Axis-II DSM-IV diagnosis or any BPD or AVPD criterion, were excluded from this study.
Table 1 presents the demographics for each group. A Kruskal-Wallis test shows that the groups do not differ on educational level (χ2 [2] = 2.04, p = .36), but they do differ on age (χ2 [2] = 7.56, p = .02). Pairwise comparisons show that AVPD patients have higher mean ages than BPD patients (Mann-Whitney: z[29] = - 2.25, p = .03) and non-patients (Mann-Whitney: z[27] = - 2.48, p = .01).
The PD groups do not differ in number of comorbid Axis-II diagnoses (Mann-Whitney: z[29] = - .94, p =
.35), eating disorders (Mann-Whitney: z[29] = - .46, p = .65), anxiety disorders (Mann-Whitney: z[29] = -
1.02, p = .31) and mood disorders (Mann-Whitney: z[29] = - 1.12, p = .26). However, the BPD patients
had more comorbid substance use disorders than the AVPD patients (Mann-Whitney: z[29] = - 2.04, p =
.04). Substance use disorders are, however, common for BPD (Taylor, 2005). Table 1
Mean Age and Educational Level Frequencies by Group
BPD patients (N=18) AVPD patients (N=13) Non-patients (N=16)
Mean age (SD) 26.9 (7.1) 34.9 (10.6) 26.3 (7.9) Educational level % MBO (N) 16.6 (3) 23.1 (3) 18.8 (3) % HBO (N) 44.4 (8) 46.2 (6) 18.8 (3) % WO (N) 38.8 (7) 30.8 (4) 62.5 (10) Materials
Diagnostics. DSM-IV Axes I and Axes II disorders were screened and assessed with the Dutch
version of the SCID-I and SCID-II (Weertman et al., 2000; Groenestijn et al., 1996). The SCID-I shows moderate to excellent inter-rater reliability, whereas the SCID-II shows excellent inter-rater reliability (Lobbestael, Leurgans, Arntz & Wiley, 2011).
Failure induction. The failure induction is a manipulated, computerized task, that was
subsequently introduced to their team members and the competing team members by pictures. The participants are told these other players are located in the same building somewhere. After all pictures are shown, the task instruction appears on the screen. This task, called ‘The Tower of London’, repeatedly shows two pictures of pegs with colored balls simultaneously, known as a start and end situation. Participants are asked to calculate the minimum number of steps they need from the start situation in order to make an arrangement of balls that is identical to the end situation. This instruction phase is succeeded by a practice phase of three items. In the first part of the task, the participants receive feedback about their own achievements. Once this practice phase is completed, the game phase is initiated. During this phase, they no longer receive feedback on their performance, however, two bar-plots on top of the screen reflect the interim scores of both teams separately. Another bar-plot in the middle of the screen represents the remaining time, with a maximum of one minute for each assignment. Once the game phase is completed, a text appears on the screen, saying: ‘We are sorry, your team lost.’ Subsequently, a bar-plot appears on the screen that represents the achievements of all players separately. The bar that represents their own achievement is significantly lower than that of other team members, implying they failed on the task and were accountable for the team loss.
Dependent variables. The Positive and Negative Affect Schedule (PANAS ; Krijns, Gaillard, Van Heck
& Brunia, 1994) serves as a manipulation check for negative affectivity. This scale contains two subscales; Negative Affectivity (NA) and Positive Affectivity (PA). Each subscale consists of 10 items, with a 5-point Likert scale ranging from 1, not at all, to 5, very much. These subscales show excellent convergent and
discriminant validity (Gijsbers van Wijk & Kolk, 1996), as well as high internal consistencies, with Cronbach’s α ranging from 0.86-0.90 for the PA scale and 0.84-0.87 for the NA scale.
A state version of the Short Schema Mode Inventory (SMI; Lobbestael et al., 2010), serves as a
manipulation check for the activation of the Punitive Parent Mode (10 items). The three other BPD related schema-modes, the Vulnerable Child (10 items), Angry Child (10 items) and Detached Protector Mode (9 items), are also assessed. The participants respond to the items by rating the extent to which they agree with the statement at that moment. All responses are made on 100mm VASs. High internal
consistencies were found for the SMI subscales of the Punitive Parent mode (α = .91), the Vulnerable
Child mode (α = .96), the Angry Child mode (α = .89) and the Detached protector mode (α = .91;
Lobbestael et al., 2010).
Semi-structured interview. To measure subjective experience a semi-structured in-depth interview
takes place (appendix 1). The interview contains a number of standardized questions, but also involves probing when additional information is required (Bryman, 2004). During the construction of the semi-structured interview, the aim was to cover all aspects of subjective experience, thoughts, feelings, behavioural tendencies and physiological experiences. Furthermore, three types of situations were covered. The present situation, in which the participant was only introduced by the other team through pictures on a screen. A second, hypothetical situation, were the participant played in the physical presence of the others. And a third situation, regarding a similar failure experience, within a real-life setting. The motivation to include these themes was to cover the effect of the presence of others and to determine if the experiences after the interview were indicative of real-life settings.
These interviews were carried out by two researchers (DB, REJV), whereas 83% of the interviews was carried out by the principle researcher (DB). The aim was to conduct the interviews without any knowledge about the research group, in order to minimize the possibility that biases about the research group influenced the interview-techniques. This was successful for both the control and BPD group, whereas the AVPD group was tested on location, because the majority took part in inpatient treatment, and therefore their primary diagnoses were inevitably known to the researchers. To further minimize bias
from the interviewers, the interview was restricted to open-ended questions and the responses of the participants were paraphrased and summarized with minimal interpretation from the researcher. Both researchers (DB, REJV) received a training on the relevant clinical interview techniques and gained sufficient experience with carrying out semi-structured interviews in clinical practice.
Procedure
This study received ethical approval from the Ethical Committee of the University of Amsterdam. All BPD and AVPD patients in this study were recruited through several mental health institutes, such as “De Viersprong”, “Pro Persona” and “Arkin NPI”. These institutions were contacted by email by professor Arnoud Arntz, in which they were asked to provide female participants between 18-55 years, with a BPD or a AVPD diagnosis. Once they agreed to participate in the study, contact information about the patients was provided to the second researcher (RV). The non-patients were recruited from the general population through advertisements, both on- and offline.
The majority of the BPD patients and all AVPD patients included in this study were already diagnosed during intake at the institution they were recruited from. However, participants that were recruited from “Arkin NPI” still required a full SCID-I (Groenestijn et al., 1996) and SCID-II (Weertman et al., 2000) assessment. BPD patients from Arkin NPI and all non-patients in this study received an email, including a link to the Dutch version of the SCID-I and SCID-II screener for DSM-IV Axes I and Axes II disorders (Weertman et al., 2000; Groenestijn et al., 1996). Non-patients that met ≥ 2 criteria for a DSM-IV Axes I or Axes II disorder, or any criterium of BPD or AVPD, were excluded from the study. The BPD patients that completed the screening procedure were subjected to further assessment by telephone, with the SCID-I and SCID-II interview for DSM-IV Axes I and Axis II disorders. This procedure was carried out by the second researcher to ensure that the principle researcher was blind for group during the interview.
All eligible participants were scheduled for the experiment. Preceding the experiment, participants were informed about participating in a study that looks into the difference in subjective experience of patients and non-patients after competitive play, in which financial incentives are involved. Following signed informed consent, they filled out the PANAS and SMI. To increase the credibility of the experiment, one of the researchers took a picture of the participant, that was told to be uploaded in the game. Afterwards, the participants received verbal instructions about the task as well as on a computer screen. Once this instruction phase was completed, they performed the manipulated, computerized task. Following this task, participants filled out another PANAS and SMI. Subsequently, their subjective experience as reaction to their failure was questioned in a semi-structured interview. This interview was followed by a debriefing, in which the true objective of the study was explained. PD patients were also asked about their current mental state.
Statistical analysis
The PANAS and SMI, state version, both serve as a manipulation check. If the experience of failure is well induced, we expect an increase in negative affect and a decrease in positive affect on the PANAS in all participants. Furthermore, prior research, into a BPD population, showed that the Detached Protector Mode was activated following a stress induction (Arntz et al. 2005). Therefore, it is not unlikely that schema modes, other than the PPM, become activated in the PD patients. Moreover, the SMI is expected to differentiate well between PD patients and non-patients (Arntz et al., 2005; Lobbestael et al., 2010; Bamelis et al., 2011). Thus, to determine if the PPM is specifically activated in the PD patients after the failure induction (Young et al., 2005; Arntz & Jacobs, 2012), and to ensure that this effect is specific to PD patients only, the scores on the SMI are also subjected to analysis.
These manipulation checks are performed with a 2 (Time) x 3 (Group) mixed ANOVA, to examine group and time differences on the SMI and the PANAS. However, if the assumption of normality and/or homogeneity of variances is violated, non-parametric analyses are performed on the difference scores between pre- and post-measures. Group differences between these scores are assessed with Kruskal-Wallis tests and additional Mann-Whitney U tests, if necessary.
Qualitative analysis
The interviews are transcribed verbatim by DB. Two randomly selected interviews are reviewed by the REJV to assess the accuracy of the verbal transcripts. Before the coding process is initiated, all interviews are assigned a random number, without any reference to their research group, in order to minimize the influence of group biases during the coding process.
Since the primary goal of this study is to unravel the subjective experience in the Punitive Parent mode, an inductive analytic approach is adopted, that remains open to whatever themes discern from the data (Charmaz, 2006). Therefore, the coding process starts with a phase of initial coding, where the
researcher (DB) remains open to every analytic possibility and creates codes that best fit the data. During this phase, two randomly selected interviews from every research group, six in total, are coded word-by-word, line-by-line, with no interferences from a theoretical framework. This initial coding process is carried out by DB. To ensure these initial codes are grounded in the data, they are limited to close summaries or even literal quotes (in-vivo codes) of the responses. This enables the researchers to determine
what themes are brought forward by the participants themselves, within the broader themes of the semi-structured interview.
Within these initial codes, certain transcript themes emerge (Pietkiewicz & Smith, 2014). However
the substraction of these emergent transcript themes, also involves interpretations of the researcher. Therefore, to maximize objectivity, both DB and REJV construct a list of emerging transcript themes independently, based on the initial codes (Charmaz, 2006). The most significant and frequent themes are merged into broader categories, creating a list of focused codes. Once all documents are coded with these focused codes by DB, five interviews within every research group, 15 in total, are randomly selected and independently coded by REJV. Once the process of focused coding is completed, these codes are subjected to theoretical integration based on the reflections and interpretations of DB and REJV. The resulting superordinate themes are therefore reflections of the emergent transcript themes and the
interpretation of the researchers.
Results Statistical results
Shapiro-Wilks tests showed that the scores on the SMI and PANAS were skewed (not normally distributed) for the non-patients. Levene’s tests also showed that the homogeneity of variances between the groups could not be assumed for all schema modes and the PANAS. Therefore, the non-parametric Kruskal-Wallis test seems a more appropriate statistical test to look into Group differences. Significant results on the KW-tests are further examined by pairwise comparisons, with the Mann-Whitney test. Time differences are examined with the Wilcoxon Signed Ranks Test. To examine Group x Time interactions, KW-tests are performed on the difference scores between pre- and post-measures, followed by pairwise comparisons, with the Mann-Whitney test. Table 2 includes means and standard deviations of all groups separately on all different outcome variables.
Table 2
Group Means and Standard Deviations of Pre- and Post-test Scores for Schema Modes and Affects
BPD patients (N = 18) AVPD patients (N = 13) Non-patients (N= 16)
M (SD) M (SD) M (SD) PPM pre-test 234.1 (178.1) 402.8 (246.5) 53.9 (82.7) PPM post-test 227.4 (218) 413 (308) 64.1 (109.1) ACM pre-test 397.2 (206.7) 443.1 (193.5) 71.9 (100.5) ACM post-test 312.8 (205.3) 371.5 (233.8) 53.9 (66.6) VCM pre-test 395.7 (253.1) 518.8 (235) 44.6 (82.7) VCM post-test 391.5 (281.7) 505.2 (303.5) 53.1 (68.4) DPM pre-test 251.6 (197) 349.5 (161.6) 29.5 (61.8) DPM post-test 185.3 (162.3) 292.5 (219.6) 29.3 (50.1) NA pre-test 18.8 (6.3) 23.8 (6.8) 11.8 (2.2) NA post-test 20.1 (9.4) 28.1 (5.5) 13.9 (4.5) PA pre-test 26.4 (5.6) 23.2 (6.3) 31.9 (5.3) PA post-test 20.6 (6.4) 20.2 (8) 28.4 (7.6)
PPM = Punitive Parent Mode; ACM = Angry Child Mode; VCM = Vulnerable Child Mode; DPM = Detached Protector Mode; NA = Negative Affect; PA = Positive Affect
On the pre-tests, Kruskal-Wallis tests showed that the groups differ on all schema modes (PPM:
χ2 [2] = 22.88, p < .001 ; VCM: χ2 [2] = 26.75, p < .001 ; ACM: χ2 [2] = 25.76, p < .001 ; χ2 [2] = 23.87, p < .001) and both negative (χ2 [2] = 13.9, p < .01) and positive emotions (χ2 [2] = 23.7, p < .001).
Mann-Whitney U tests were performed to make pairwise comparisons between groups. At pre-test, BPD patients scored significantly higher than non-patients on the Punitive Parent Mode (z[32] = -
3.54, p < .001), the Vulnerable Child Mode (z[32] = - 4.32, p < .001), the Angry Child Mode (z[32] = -
4.39, p < .001) and the Detached Protector Mode (z[32] = - 3.7, p < .001). This result was in line with the
hypothesis that BPD would score higher on the PPM than non-patients. BPD patients also reported significantly more negative emotions (z[32] = - 2.73, p < .01) and less positive emotions than non-patients
(z[32] = - 3.53, p < .001) on the pre-test.
AVPD patients also displayed significantly higher scores than non-patients on the Punitive Parent Mode (z[27] = - 4.15, p < .001), the Vulnerable Child Mode (z[27] = - 4.35, p < .001), the Angry Child
Mode (z[27] = - 4.26, p < .001) and the Detached Protector Mode (z[27] = - 4.39, p < .001) on the
pre-test. This result was in line with the hypothesis that AVPD patients would score higher on the PPM than non-patients. Furthermore, AVPD patients reported more negative emotions (z[27] = - 3.32, p < .001)
and less positive emotions (z[27] = - 4.42, p < .001) than non-patients.
With regard to the differences between BPD patients and AVPD patients on the pre-test, pairwise comparisons showed that AVPD patients scored significantly higher than BPD patients on the Punitive Parent Mode (z[29] = - 2.26, p = .02). However, the groups did not differ on the Vulnerable
Child Mode (z[29] = - 1.6, p = .11), the Angry Child Mode (z[29] = - .92, p = .36) and the Detached
Protector Mode (z[29] = - 1.72, p = .09). Nor was there any statistical difference between self-reported
Wilcoxon Signed Ranks Tests are performed to look into differences in the total sample between pre- and post-test scores on the SMI-state version and the PANAS. There was no statistical difference between pre- and post-tests on the Punitive Parent Mode (z = - .6, p = .55) and the Vulnerable Child
Mode (z = - 1.06, p = .29). However, the participants scored lower on the Angry Child Mode during the
post-test than at pre-test (z = - 3.2, p < .01) and the same accounted for the Detached Protector Mode (z
= - 2.86, p < .01). Furthermore, the total sample scored higher on negative emotions on the post-test,
compared to the pre-test (z = - 4.46, p < .001), but lower on positive emotions (z = - 2.85, p < .01).
To examine Time x Group interaction effects, new variables were computed, using difference scores between pre- and post-tests and Kruskal-Wallis tests were performed to look into group
differences on these difference scores. There was no significant Group effect on the difference scores of the Punitive Parent Mode (χ2 [2] = 3, p = .23), Angry Child Mode (χ2 [2] = 2.01, p = .36), Vulnerable Child Mode (χ2 [2] = .39, p = .82), negative emotions (χ2 [2] = 2.42, p = .3) and positive emotions (χ2 [2] = 1.32, p = .52), indicating the absence of group differences between the difference scores of these variables. The
absence of a Group effect on the differences scores of the PPM, negative emotions and positive emotions, contradicted the hypotheses, since it was expected that the PD groups would show a greater increase in PPM than the non-patients. However, there was a significant Group effect on the Detached Protector Mode (χ2 [2] = 7.81, p = .02). These differences were further examined with pairwise
comparisons between the groups. The decrease between pre-and post-test on the Detached Protector Mode is larger for the BPD patients, than for the non-patients (z[32] = - 2.95, p < .01). However, the
AVPD patients did not differ from the non-patients in difference score (z[27] = - 1.72, p = .09). The BPD
patients also did not differ from the AVPD patients in their difference score (z[29] = - .32, p = .75). Qualitative results
The verbatim transcriptions of the interviews by the principle researcher (DB) were evaluated as high quality by the second researcher (RV), with minimal deviations from the audio recordings. The initial coding process, carried out by DB, resulted in a total number of 246 codes (appendix 2). Based on these
initial codes, DB and REJV independently constructed a list of emergent themes, which were combined in one list of 134 focused codes, divided into parent- and sub-codes. These focused codes were used by DB to code all 47 documents. During this stage of focused coding, 43 additional codes were added, to adequately fit all responses, resulting in a final number of 177 focused codes (appendix 3).
To calculate intercoder agreement, 15 interviews, five within every group (BPD/AVPD/non-patients), were randomly selected and also independently coded by RV. Intercoder agreement was calculated as the percentage agreement of the selected text within similar codes. This percentage agreement ranged from 64% to 81.2% between the documents. To calculate Cohen’s kappa, the codes that were used to calculate the percentage agreement, were counted once for each document. Cohen’s kappa’s ranged from .6 to .83, indicating good to very good intercoder reliabilities (appendix 4).
Finally, the focused codes were merged into more theoretical, superordinate themes, by REJV and DB together, to facilitate an understanding of the meaning and relations between the responses (appendix 5).
Because this study looked into subjective experience, it was of relevance to give the participants a voice, by including quotations. However, this study was performed in the Netherlands, and therefore these quotations required translations to English. During these translations, it was attempted to translate the quotes as literally as possible, while preserving the meaning (Van Nes, Abma, Jonsson & Deeg, 2010). However, the preservation of the meaning was of higher priority than literal translations. In order to compare the translations to the original quotes, the original Dutch quotations are presented in appendix 7.
Characteristics of the Punitive Parent Mode. Within the focused codes, one theme emerged
with significant relevance to Young’s PPM; a negative general self-evaluation. Responses within this theme were
characterized by strong critical self-evaluations that could be task-specific, but mostly exceeded the task performance. A total number of 25 documents were coded with this code (53.2%). However, these responses were only indications of the presence of a PPM and therefore these documents represented the
PPM candidate list. The first analytic step was to look into these self-devaluating responses thoroughly, by
answering the following question:
What do these negative general self-evaluations have in common?
Within the responses, almost all participants, that appeared to have a general negative self-view, spoke about those self-devaluations with a high degree of determination (88%).
“I am just bad or worthless, I didn’t contribute anything, I drag other people down with me.”
(pt. 40, 18)
“That I am worthless. What am I still doing here? I am useless to others.” (pt. 37, 54)
Another common characteristic of these responses was the sense that failure was a conformation of
their general self-image (76%). In other words, they expected to fail on the task, because they expect to fail on everything, as is apparent in the following quote:
“I can’t do it. I told you so, I can’t do it, that’s what this shows.” (pt. 16, 28)
“During the game I still hope that someone else is dumber than I am. And afterwards it is like, surprise on the one hand, like, oh it is true, you always hope to know it, not that I do this every week, but then it is, yes it is another confirmation like, yes, I told you so.” (pt. 32,
18)
A third common theme that emerged from the responses was the tendency to generalize their
performance to other situations, their intelligence or their general self-image (92%).
(About other situations) “This confirms my other weaknesses. Participating in society.”
(pt. 40, 22)
(About her intelligence) “…normally that’s one of my strengths, and you accept that and when it fails to be true, then you think, I told you so, it’s not true, it’s just what people say.”
(pt. 20, 22)
(About her general self-image)“That’s were my insecurity presents itself, that I am not good enough. In a bigger picture, in all domains of my life.” (pt. 41, 30)
According to Young’s model, the PPM is known for its harsh, critical, punitive and blaming self-evaluation (Young et al., 2005). All responses that were categorised under the code negative self-evaluation,
showed self-criticism and/or self-blame. The harsh tone that is conceptualised by Young seems to overlap with the determinations that became apparent in the responses. However, not all responses were
characterized by this harsh, determinant tone. Moreover, Young proposed that, during the PPM, patients seem to perceive themselves as bad, to blame, evil, defective and deserving punishment. This implies that
patients in a PPM make general self-devaluating statements about themselves, which seems to overlap with the generalizations found in the responses. However, not all responses were characterized by general self-devaluations. Therefore, only responses that indicated determinations and generalizations were seen as indicative of the activation of a PPM.
During the second phase of the analysis, the interviews were assigned to their original research group (BPD/AVPD/Control), to enable group comparisons between the responses. Within every research group, the documents were further subdivided according to the presence or absence of the PPM, based on the presence or absence of determinant and generalizing negative self-evaluations. With regard to these exclusion criteria, seven documents were removed from the PPM candidate list (see appendix 6). This resulted in 8 PPMs within the BPD group (44.4%) and 10 PPMs within the AVPD group (76.9%). Frequency distributions of the PPM in the three research groups are presented in table 3.
Table 3
Frequencies of Included and Excluded PPM Documents by Group
BPD AVPD Control Total
PPM (candidate list) 8 (11) 10 (12) 0 (2) 18
No PPM (candidate list) 10 (7) 3 (1) 16 (14) 29
Total 18 13 16 47
The second phase of analysis looks deeper into the subjective experience in the PPM. Since this study adopted a more inductive method, it was aimed to subtract more theoretical constructs from the data. These were summarized in six broader superordinate themes, that resulted from the data and the interpretation of DB and REJV. The responses of the participants are described according to these superordinate themes:
1. Expectations about performance 2. Opinions about the task itself 3. Reasons for failure on the task 4. Effect from failure on the task 5. Ideas about self
6. Ideas about others
With respect to these superordinate themes, this phase of data-analysis follows several steps. To determine what characterizes the PPM in BPD, it is meaningful to look into the specificity of this subjective experience by determining how it deviates from the PPM in AVPD patients and the responses of healthy controls. Moreover, a little more than half of the BPD patients showed no activations of the PPM and, therefore, it seemed meaningful to examine what type of subjective experiences characterized this subgroup. The subjective experience of these four different subgroups is discussed separately, according to the previously described subordinate themes.
First, the subjective experience of patients with BPD in the PPM will be discussed, referred to as the BPD-PPM patients. This is followed by a description of the subjective experience of the PPM by AVPD
patients, referred to as the AVPD-PPM patients. Subsequently, the subjective experience of the healthy
controls is discussed and this group is referred to as the non-patients. Last, the experience of the BPD
patients, that did not show signs of a PPM activation, is discussed. The latter group is referred to as the
Punitive Parent Mode in BPD. The following descriptions are limited to the BPD patients that
showed signs of a PPM activation (N=8). Six out of eight BPD-PPM participants declared that they
disliked the task (75%). The main reasons were ‘it gets on my nerves’ (pt. 24, 14) or ‘it makes me mad that it’s so hard’ (pt. 3, 4). Two participants said they liked the game (25%).
Not many BPD-PPM participants spoke about their expectations about their performance. The ones that did varied in their responses. One participant was surprised about her score and declared that she ‘…did not expect to be the worst’ (pt. 3, 20) while another said that it was confirming, because ‘…I already doubted myself during the game’ (pt. 28, 14).
Six out of eight BPD-PPM participants brought forward task specific reasons for their failure (75%), such as choking under time pressure or flaws in the task, whereas only 1 participant spoke about her own ability, but she also gave state-related reasons for her performance and reported being very tired.
Half of the participants reported that this performance had a negative effect on their overall mental state. They declared that a failure like this could affect their mental state for hours, days or weeks.
“It’s bizarre how this affects me, it surprises me too, that a small test like this, I just, I just lose my mind.” (pt. 33, 14)
When they were presented with their scores, the majority of the participants had immediate thoughts about being the worst player (62.5%) and half thought about being responsible for the team failure. One participant declared that she felt betrayed, because she did not believe the other team-members to be real, whereas another started to criticize herself for her crying over something little like this.
With respect to emotional experiences, the majority reported feelings of guilt (75%). More than half of the participants also reported feelings of anger (62.5%). Four of them stated that they directed this anger towards themselves, appearing as self-blame for their responsibility in the team failure (50%). The other one felt betrayed and expressed her anger indirectly towards the researcher.
“No, but I was mad, more about the veracity of the team, so to speak, than about the game itself. I thought, is this real or isn’t it? That’s what made me mad.” (pt.3, 24)
Two participants felt ashamed about their overall intelligence and two reported feeling sad about their performance. The latter two also cried during the interview. In the BPD-PPM group, all emotions were experienced with intensity.
Four participants reported unpleasant physiological experiences (50%), such as a cold feeling, increased heartbeats, breath takes, tingling in her hands and tension in her stomach.
Five out of eight declared that they felt a tendency to leave the building when they were presented with the scores (62.5%). They reported feelings of shame towards their team-members and were anxious about running into them in the hallway. One participant said that she wanted to apologize to the others, whereas another felt the tendency to evaluate the game with the others. One participant said that if this would happen in real life she would practice until her performance would be perfect.
Within the responses a pattern of self-evaluations became apparent. All participants generalized their
performance to their general self-view (100%), with determination in their statements (100%). Seven
participants generalized their task performance to their overall intelligence and called themselves ‘stupid’ (87.5%), whereas one concluded that she was never good enough. Six out of eight considered this performance as confirmative for their general self-view (75%). Moreover, six of the BPD-PPM participants
compared themselves to others and declared they were not good enough and that they were subordinate to others (75%). Four stated that they always ruin it for others (40%).
“When I am compared to others, and since I subordinate myself to others, there are always others who do better. And I am always the one that ruins it for the rest.” (pt. 28, 32)
Another interesting aspect that four of the BPD-PPM participants brought forward is a discrepancy between their feelings and thoughts (50%):
“What I say, if I see their scores and my own, than I think, they can do it and I can’t. I doubt myself at that point. It doesn’t matter if that’s rational, somehow I realize that it’s not reasonable what I think, but that doesn’t impact me. It doesn’t win from the fact that I subordinate myself to others.” (pt. 28, 74)
When they spoke about others, three of them reported overall positive views, as if they looked up to them, stating they were smarter, better and more self-confident (37.5%), whereas one made a statement about others being better at this task. Half of the BPD-PPM participants also expected that the others would blame them for the team loss. Moreover, all participants viewed the others as rejecting or criticizing, expecting that others would evaluate them as ‘stupid’ (87.5%) and ‘very bad’ (12.5%). All expected evaluations by others matched the self-evaluations exactly (100%). One participant experienced people in her personal life as less rejecting:
“They know who I am, so they could think that it’s not my intelligence per se, that I have other qualities. But people I don’t know have no idea how I am in my personal life.”(pt. 28,
80)
If they played the game in the physical presence of others, they expected to experience more shame (64.2%) and stronger feelings of guilt (50%) regarding their failure.
Summary.
The majority of the BPD-PPM patients stated they disliked the game. Half reportedthat their mental state was negatively affected by their failure and that similar feelings could last for even weeks. The majority reported immediate self-devaluating and self-blaming thoughts and a minority reported unpleasant physiological experiences, but there was no consensus between them. Shame seems to be more activated in the presence of others, whereas guilt and self-directed anger seem more
internalised. Almost all self-views were characterized by determination, confirmation and generalisation. A majority reported being subordinate to others. BPD-PPM patients tend to generalise their failure
specifically to their overall intelligence. Self-views and expected evaluations by others were an exact match. A large majority blamed their performance on task specific difficulties. However, even though they reported task specific difficulties and did not see themselves as unskilled, their bad performance felt like a confirmation of their general self-image. This also becomes apparent in the discrepancy between their feelings and thoughts, they reported themselves. It appears as if they literally internalized the critical and punishing parental voice, without finding rationalisations for this critical tone.
Punitive Parent Mode in AVPD.The following descriptions are limited to the AVPD patients that showed signs of a PPM activation (N=10).
Two out of ten AVPD-PPM patients reported they disliked the task (20%), because they failed to improve themselves on similar tasks and because it was played on a computer. Two stated they liked the
game (20%). None reported feeling confident about their performance. Six declared that they were unskilled for the task (60%). Seven participants declared that they doubted their own performance during the task (70%). Four of the participants that stated they doubted their experience called themselves unskilled and also reported difficulties with the time pressure (40%). Three of them stated they doubted their performance during the task because of time-pressure only (30%).
“It was hard, I dislike time pressure anyway and the pressure of seeing the point-scale move and the fact that you’re in a team, so you want to put extra effort into it.” (pt. 9, 4)
One participant blamed the low scores on her efforts and one blamed it on her superficial reading of the instructions.
Five participants reported that their mental state was strongly affected by their failure on the task (50%). However, the duration of this effect ranged between two minutes, weeks or even months. They also stated that they felt it very intensely, expressed as ‘…it hit me hard’ (pt. 9, 110). Three participants questioned the veracity of the experiment, but said that this did not affect their experiences.
When they were presented with the scores, four participants reported they had immediate thoughts about being the worst player (40%), whereas six of the participants thought about being responsible for the team loss (60%). One participant said to herself ‘…you have the lowest score, so you abandoned the other players’ (pt. 41, 16). Two participants declared that they were disappointed by the team loss, another could not believe she was ‘that bad’ (pt. 40, 16).
With regard to their emotional experiences, four participants reported anger, directed at themselves (40%), one stated that it was no intense feeling, whereas another explained her intense self-directed anger as a way of ‘…punishing myself’ (pt. 37, 52). Three also reported fear about the anger of others or their possible punishment (30%). Feelings of shame were reported by four participants (40%), whereas one participant continued saying this feeling was not intense. Feelings of guilt were reported by six (60%) participants. Other reported feelings were despair (20%) and disappointment (10%). All, but two, of the reported emotions were experienced with intensity in this group.
Eight out of ten reported unpleasant physiological experiences (80%). Four participants reported muscle tension (40%) and another four reported a specific tension in or squeezing of their stomach (40%). Other psychological experiences were breath-takings, shaking, feeling cold, pressure on the chest and dissociation, the latter explained as ‘...no longer being able to feel my head’ (pt. 37, 68).
When presented with the scores, two participants reported the desire to leave the building immediately. One wanted to ‘…make myself invisible’ (pt. 9, 96). During the interview one participant excused herself to the researcher for her effort to rationalize her experience, suggesting that she might this might influence the experiment negatively. She also felt the tendency to punish herself with starvation. Another participant spoke about her tendency to overcompensate to others in real life, she declared:
“In social situations, where I get mad, I don’t express that, I act very cautiously to others, like I did something wrong. I overcompensate, or something like that. As if I want to make it up with them extra hard.” (pt. 37, 36)
The response patterns regarding the self-image showed a clear pattern of determination (100%). As
discussed before, five participants had ideas about being unskilled for the task (50%), whereas nine spoke about the fact that this failure felt like a confirmation for their expectation about themselves (90%). Seven
out of ten participants stated ‘I told you’ (70%), followed by ‘I can’t do it’ (50%), ‘...it’s me again’ (pt. 41, 28) and ‘…I am worthless’ (pt. 42, 16). Ten out of ten generalized their experience (100%), to their general
stated that she was good in other things, but experienced this performance as a failure of herself as a person nonetheless. Five participants spoke about being subordinate to others (50%), whereas three described themselves as worthless in general (30%).
Five spoke about a discrepancy between their feelings and their thoughts (50%). One participant spoke about being subordinate to others specifically and when she was asked what made her think that, she stated:
“It’s more like an idea I keep in my head, that’s maintained by me, then that it’s confirmed by others. It had been confirmed in the past and now I maintain it myself.” (pt. 41, 50)
When they spoke about others, half of them seemed to hold superior views of others, stating that others were more skilled and smarter than they were (50%). Four of them generalized this statement to being subordinate to others in general (40%), whereas one spoke about others only being better at the task. Moreover, five participants viewed others as blaming, and expected that others would indeed blame them for the team loss (50%). Four of them agreed, saying they fell short on the others, even though they all stated were unskilled for the task (40%). One stated that others good not write her off completely because of one failure. Seven experienced others as criticizing or rejecting (70%). They expected that others would evaluate them as ‘stupid’ (20%), ‘can’t do anything’ (20%), ‘the worst of all’ (20%) and one participant expected to be evaluated as ‘worthless’. All seven agreed saying they were subordinate to others or worthless in general (70%).
Six participants expected to feel guiltier in the physical presence of others (60%). One reported that she would feel more shame and another would feel worthless and subordinate to others.
Self-evaluations and expected evaluations by others showed a high degree of overlap (90%). That these are indeed connected becomes apparent in the following quote:
“It might sound weird, but I have a scale in my head, where everyone has its own rank based on their value. If other people do a good thing, they are squared upwards and when they do a bad thing, they drop only a little bit. I dangle somewhere at the bottom, so when I do
something bad, it’s squared downwards and when I do a good thing, I can only rise a little bit. In my head, everyone knows that unconsciously, that I’m worth nothing.” (pt. 42, 20)
Summary. Most AVPD-PPM patients tend to doubt themselves in general, as became apparent in
their expectations during the task. The majority experienced themselves as unskilled for the task, but they also reported task specific difficulties. Half reported that their mental state was negatively affected by their performance on the task and that similar feelings could last for even months. All participants hadimmediate self-blaming or self-devaluating thoughts. Unpleasant physiological experiences are highly prevalent among this group and there appears to be consensus between them. Mostly, they reported muscle tension and tension in their stomach. Feelings of guilt, shame and self-directed anger were most prevalent in this group, but they also tend to be anxious about the reactions from others. In the presence of others, they expected to feel more guilt. A minority reported avoidance tendencies and none felt the desire to get in contact with others. Self-punishing and overcompensating tendencies were also reported.
Self-descriptions were highly characterized by determinations, confirmations and generalisations. A minority generalizes her performance to her overall intelligence. Mostly, they generalized to other domains in life, other tasks that require skill and their general self-view. Self-evaluations and expected evaluations by others showed a high degree of overlap. Half of them tend to hold superior views of others, but also perceive them as criticizing and rejecting, thinking they were unskilled in general or worthless. They
expected that others would blame them and blamed themselves too, even though they felt unskilled for the task. This discrepancy between their feelings and thoughts is also explicitly described in their self-reports. AVPD-PPM patients speak to themselves as if they literally internalised a critical and punishing parental voice, without finding any rational ground for these claims.
Failure experience in non-patients.
The following descriptions are limited to the non-patients in
this study (N=16). Ten out of sixteen non-patients said they liked the task (62.5%), whereas one declaredshe ‘…hated it, because it fries her brains’ (pt. 47, 4).
Eleven participants doubted their performance during the task (68.8%). Seven of them said their scores confirmed their expectations (43.8%), whereas two said their scores were worse than they expected (12.5%) and another two stated that their score was higher than expected (12.5%). Eleven non-patients reported that they were unskilled for the task (68.8%) and five of them also reported task-specific reasons for their failure (31.3%), such as time-pressure and unclear instructions. Three non-patients blamed their failure on task-specific reasons only (18.8%).
With regard to the effect of the scores on their general mental state, thirteen declared that they were not affected by their scores and did not blame themselves too much (81.3%). None declared that the scores impacted their mental state in a negative way. Two questioned the veracity of the experiment and reported that these ideas decreased the intensity of their experience.
When they were presented with their scores, eight participants had immediate thoughts about being bothered by the scores (50%), however six of them continued saying ‘they didn’t feel bad about it’, it was more like ‘...a bummer’.
“Bummer, but I didn’t feel guilty or bad about it.” (pt. 17, 24)
Four had immediate self-blaming thoughts about being responsible for the scores (25%) and three reported self-devaluating thoughts about being the worst player (18.8%). Another three stated that they did not feel bad, because they expected this outcome (18.8%). One of the latter declared that she thought her score was funny.
Five stated they felt guilty (38.5%), whereas two of them declared they felt it only a little bit (12.5%). Five declared they felt ashamed (38.5%), with three of them stating it was no intense feeling (18.8%). Other participants said they were disappointed in the scores (18.8%), they could laugh about it (12.5%) and one participant stated that she was a little mad at herself.
“I was a little mad at myself, not really mad, but just a little bothered.” (pt.11, 32)
Two participants reported unpleasant physiological experiences (12.5%). One participant reported increased heartbeats and temperature, whereas another spoke about tension in her stomach.
When they were presented with the scores, eight of them felt the tendency to apologize to the other team-members (50%), four just wanted to have a chat with the others to evaluate the game (25%) and one felt the urge to leave the building. When a similar situation would occur in real life, nine participants would try to improve themselves in the future (56.3%).
“I try to imagine a situation like that. I would think, shit, I didn’t do well and then I would quickly analyse what went wrong and how I can improve that.” (pt. 35, 68)
Eight participants gave responses that included views of themselves (50%). Five participants declared that they were good in other things (31.3%) and three of them specifically stated that this
performance did not affect their general view. Two participants used the word ‘stupid’ in their self-descriptions. However, one statement lacked determination, because she stated that she ‘felt’ stupid,
whereas the other did not generalize this description to her overall intelligence:
“And sometimes I thought like, neh, actually I am a little stupid at tasks like these. And then I think, what are the averages?” (pt. 14, 20)
Another participant stated that she is never responsible for problems at work:
“The weird thing is that if situations like these happen at work, that I’m never the reason it fails. Because these are things that require precision and being one time and I’m mostly the one that guides the process. So this is hard, those things never happen to me, that I’m the one that fucks up.” (pt.10, 38).
None of the participants reported a specific discrepancy between their feelings and thoughts. With regard to their views of others, three participants reported that others were better or more skilled at the game (18.8%). Four participants expected others to criticize them for their efforts (25%). From the thirteen participants that stated they did not blame themselves for their performance, six perceived the others as more blaming than themselves (37.5%). Out of these thirteen, another six perceived the others to be as tolerant as they were themselves, expecting that they would not blame them either (37.5%). Two participants blamed themselves and expected the same from others (12.5%). One participant perceived others to be more tolerant than she is, saying she ruined it for the rest, whereas she expected others to be only bothered by the team loss in general.
In the physical presence of others, five expected they would have felt ashamed (43.8%), but two did not expect to feel this intensely (12.5%), whereas another two declared that the presence of others would not matter, because they would feel just as ashamed (12.5%). Five expected to feel stronger feelings of guilt (31.3%), but two declared that this feeling probably will not be very intense (12.5%).
Summary
. The majority of the non-patients enjoyed the game, but also doubted their performanceduring the game. A little less than half evaluated their score as confirming for their expectations, but these confirmations were restricted to their task performance. A large majority declared that she was not affected by her failure and none declared being negatively affected. However, they reported immediate thoughts about being bothered by their scores and more than half reported devaluating or self-blaming thoughts. Close to half of the non-patients reported feelings of guilt and shame, but half of them stated that these feelings were not very intense. Self-directed anger was very low prevalent and not very intense, and disappointment was also reported by a small minority. In the presence of others a large minority expected to feel more ashamed, whereas a smaller proportion would also experience more guilt. A small minority experienced unpleasant physiological experiences, described as tension in the stomach and increased heartrates and temperature. The majority felt the desire to contact the others, in order to apologize or evaluate the game. When a similar situation would occur in real life, more than half of the participants would try to improve themselves for the future.
A small minority reported self-devaluating descriptions, but they lacked determination or
generalisation. The non-patients also reported that they were skilled in other things or even flawless in real life. A small minority perceived the others as better than they were, but this was only restricted to the task. Furthermore, they reported a great variety in their expected evaluations by others. A large minority perceived them as more critical then themselves, the same amount perceived them just as tolerant, whereas a small minority perceived others to be just as critical as them.
Failure experience in non-PPM BPD patients.
The following descriptions are limited to the
BPD patients in this study that showed no signs of a PPM (N=10). Since a majority of the BPD patientsdid not show indications of a PPM activation, it was meaningful to look into the reports of these BPD patients, to determine what characterizes their responses. The following descriptions are limited to the BPD patients that did not show signs of a PPM activation (N=10).
Five out of ten non-PPM BPD patients declared they liked the task (50%), whereas four reported disliking it (40%). Two of the patients that disliked the task stated that it was ‘pointless’, one said it was a weird game and another said that she disliked it because of the time pressure.
Seven participants doubted their experience during the task (70%). Four of them blamed their failure on their task ability, another two attributed their failure to the time pressure and one said that she did not put any effort into it. The three participants that did not doubt their performance during the task (30%) blamed their failure on the time pressure and one also spoke about her low task ability.
One patient reported being a little negatively affected by her failure, but continued saying that she rationalised this by telling herself ‘she did the best she could’ (pt. 27, 30). Five stated that they were not negatively affected or bothered by their failure (50%), because they ‘didn’t care’ (30%), did not blame themselves and did the best they could. However, three participants had immediate self-devaluating thoughts (30%) and three also reported self-blaming thoughts when they were presented with their scores (30%). Furthermore, three participants reported feeling guilty (30%), one said she felt ashamed and one BPD patient reported self-directed anger. These feelings were all experienced with intensity. Moreover, five participants declared that they did not feel anything (50%). One of them questioned the veracity of the experiment and three of them specifically spoke about feeling indifferent (30%), saying they did not care about their scores at all. Two participants explained this indifference as a self-protecting mechanism, whereas one of them said:
“Maybe it’s because, so I don’t have to feel anything, because if I did feel something, it wouldn’t be a positive feeling, given the fact that I lost.” (pt. 26, 52)
From the three participants that reported feeling indifferent, two participants felt a tendency to avoid the situation by leaving the building (20%). One of them also reported that she would use verbal or physical aggression if other people would blame or criticise her for her failure.
Furthermore, five out of ten BPD patients reported a tendency to excuse themselves to others or evaluate the game with them (50%). None reported a tendency to improve themselves if a similar situation would occur in real life, nor did any BPD patient report unpleasant physiological experiences.
Three BPD patients reported critical self-descriptions (30%). One reported feeling incompetent in general, however, her statement lacked determination, because she continued that this was more a feeling than a fact. The latter was explained by her as a discrepancy between her feelings and thoughts. Another participant spoke critical about her physical appearance, stating that she looked dishevelled, but she did not generalize her self-criticism to other domains. The third BPD patient with a critical self-description showed an overall positive view of herself, however, her desire to outperform others in real life serves as a ground for self-criticism:
“For example, when I get a lower grade than someone else, I evaluate myself as bad. Not bad in the sense of morals and values, but I should have done better. I always have to be better than others.” (pt. 5, 22)
None of the BPD patients experienced themselves as subordinate to others. Four participants perceived others as tolerant, expecting that they would not blame them for the team failure (40%).
However, six participants expected others to be critical or rejecting (60%), saying they were stupid, ‘very bad’ or ‘a loser’ and one of them also expected that others would blame her for the team loss.
Four participants in this group reported self-evaluations that matched their expected evaluations by others (50%). Two of them were self-critical or -blaming and expected the same from others (20%), whereas three did not blame themselves and expected others to be just as tolerant (30%). Four
participants expected others to be more blaming and criticising than they were towards themselves (40%). The latter group included all participants that spoke about feeling indifferent (30%). The participant that reported a desire to outperform others expected others to be less critical towards her, than she is towards herself. In the presence of others, three expected to feel more ashamed, whereas another expected to feel more guilt.
Two out of ten BPD patients reported that they questioned the veracity of the experiment (20%). One declared that these doubts did not affect her experience, because - regardless of the veracity - she would not devaluate herself because of one failure on a specific task. The other declared that, if this would occur in real life, she would be really mad at herself, because she subjects herself to extremely high
standards.
Summary
. The BPD patients varied in their opinions about the task, whereas half of theparticipants liked it and almost half disliked the task. The majority of the BPD patients doubted their performance during the game and perceived their failure as conforming for their task-expectations. Half of the BPD patients said that their failure had no consequences for their overall mental state, whereas a small minority reported negative short-term effects. Moreover, a minority reported blaming and self-devaluating thoughts, feelings of guilt and feelings of indifference. The latter feeling seemed to mask an underlying vulnerability. Self-directed anger was very low prevalent in this group, as well as shame. However, a larger proportion would expect to feel more shame in the physical presence of others.
Avoidance and acting-out behaviours were only reported among the group that felt indifferent, whereas approach behaviours were reported among the rest. A minority reported critical self-evaluations, but they lacked determinations or generalisations. One self-report was specifically characterized by a desire to outperform others and critical thoughts if she fails to do so. Others were perceived both as tolerant and critical in this group, but self-evaluations only matched with the expected evaluations by others in half of the participants. A minority expected others to be more critical or more tolerant than they were towards themselves. A minority also questioned the veracity of the experiment, but this only affected the intensity of the experience, not the content.
Comparisons between codes
In order to make comparisons between the four groups and structure this great variety in qualitative data, the frequency of the most significant codes is presented by group, in table 4. Table 4
Frequencies of Most Significant Codes by Group
BPD-PPM (N=8) AVPD-PPM
(N=10) CON (N=16) Non-PPM BPD (N=10) Expectations performance
Low (%) 12.5 70 68.8 70
High (%) 12.5 0 12.5 0
Reasons for failure
Task specific (%) 75 70 50 50
Ability (%) 12.5 60 68.8 50
Overall negative (%) 50 50 0 10 Self-blaming thoughts (%) 50 60 25 30 Self-devaluating thoughts (%) 62.5 40 18.8 30 Indifferent (%) 0 0 0 30 Guilt (%) 75 60 38.5 30 Shame (%) 25 40 38.5 10 Fear of punishment (%) 0 30 0 0 Self-directed anger (%) 50 40 6.3 10 Self-punishment tendency (%) 0 10 0 0 Avoidance tendency (%) 62.5 20 6.3 20 Approach tendency (%) 25 20 75 50 Acting-out tendency (%) 10 0 0 10
Tendency to improve self (%) 12.5 0 56.3 0
Unpleasant sensations (%) 50 80 12.5 0 Self-image Self-critical (%) 100 100 12.5 30 Confirmations (%) 75 90 0 0 Determinations (%) 100 100 6.3 10 Generalizations (%) 100 100 6.3 20 Other image Blaming (%) 50 50 50 10 Criticizing/Rejecting (%) 100 70 25 60 Superior in general (%) 37.5 40 0 0 Tolerant (%) 0 10 43.8 40
Match self-by others
evaluations (%) 100 90 50 50
Discrepancy
feelings-thoughts (%) 50 50 0 10
Discussion
This study looked into the subjective experience of BPD patients in the Punitive Parent Mode. This schema mode is conceptualized by Young et al. (2005) as the internalization of the harsh, critical, blaming and punishing parental voice from their past. To activate this mode, a manipulated team task was performed that induced a sense of personal failure, as well as the experience of failing others. The task was performed by BPD patients, AVPD patients and non-patients, followed by semi-structured interviews to look into their subjective experience. Qualitative inductive analysis was performed on these interview responses and emergent themes were compared to Young’s conceptualization of the PPM. Based on the activation of the PPM, participants were divided into subgroups. The responses of BPD-PPM, AVPD-PPM, non-patients and non-PPM BPD patients were separately discussed, according to the superordinate themes that resulted from the data. Moreover, close-ended questionnaires were conducted to look into group and time differences between emotions and BPD schema modes. These statistical results will be first discussed, followed by the results of the qualitative analysis.
As expected, the PD groups differed on all BPD related schema modes from the non-patients (Lobbestael et al., 2010; Bamelis et al., 2011; Klokman et al., 2001; Arntz et al., 2005). Moreover, the AVPD group scored higher on the PPM than the BPD group. There was no increase in schema modes after the failure induction for all groups. However, negative emotions increased in all participants and there was an overall decrease in the Angry Child mode, Detached Protector mode and positive emotions. Furthermore, the BPD group showed a greater decrease in the DPM than the non-patients. These effects are remarkable, since it was expected that the failure induction would increase the PPM activation in the PD groups and therefore an interaction between time and group was expected for this schema mode (Young et al., 2005; Arntz & Jacobs, 2012; Arntz et al., 2005). However, it should be noted that the
