In the present study, we conducted three independent experiments to examine false memory susceptibility in BPD. First, we conducted two preliminary experiments to construct BPD-related DRM word lists that were consistent with the clinical description of BPD (see online supplementary materials for more details). Having compiled BPD-related wordlists, we then conducted the main experiment in which we examined false memory in BPD for these word lists as well the more standard word lists that are commonly used in the DRM paradigm. Below, we describe the methodology for the main experiment; the details of the two preliminary experiments that we conducted to develop the word lists are available as online supplementary materials.
Participants
A total of 300 undergraduate students were recruited for this experiment via an online recruitment database. All participants were required to be native speakers of English: two participants were excluded from the experiment because they did not meet this criterion. The final sample (N = 298) consisted of 211 females (M age = 21.22 years; SD = 3.12) and 87 males (M age = 21.55 years; SD = 3.80); no participants identified as gender-diverse. Participants identified as New Zealand European (66.4%), Māori (4.7%), Asian (13.8%), or “Other” (19.8%), which included those of other European descent. Using embedded validity indicators, no participants were excluded from the experiment due to suspected random or noncontent-based responding. All participants provided written, informed consent; and were tested individually using online Qualtrics (www.qualtrics.com) and E-Prime (Psychology Software Tools, Inc., 2012) software. They were reimbursed $40 NZD for their time. The research was reviewed and approved by the University’s Human Ethics Committee (Health).
Self-Report Measures
Descriptive statistics for each scale score appear in Tables 1 and 2.
McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD). The MSI-BPD [37] is a 10-item yes/no questionnaire designed to assess the symptoms of borderline personality disorder. Responses are summed to create an overall score ranging from 0–10 with higher scores indicating a greater experience of symptoms of borderline personality disorder. Although the MSI-BPD is as a continuous measure, a cut-off score of 7 has been used to identify individuals experiencing significant subthreshold symptoms of BPD [37]. The MSI-BPD is frequently used in research and has good psychometric properties. Reliability analysis (Cronbach’s α) indicated adequate internal consistency (α = .74).
SCID-II Personality Questionnaire (SCID-II). The SCID-II [38] is a 119-item true/false questionnaire designed to assess 10 personality disorders according to the DSM-IV diagnostic criteria. In the current study, only the BPD scale from the SCID-II was used; the BPD scale consists of 15 items and was initially designed as a screening device [39] for BPD based on the DSM-IV [40]. Responses are summed to create an overall score ranging from 0–15 with higher scores indicating a greater experience of symptoms of BPD. A cut-off score of 6 or greater is frequently used to identify individuals experiencing significant subthreshold symptoms of BPD [41]. Reliability analysis (Cronbach’s α) indicated adequate internal consistency for the BPD sub-scale (α = .79).
Borderline Personality Disorder Impairment Scale(BPD-IS). The BPD-IS [42] is an 11-item self-report rating scale developed to assess BPD impairment according to DSM–5 Section III Criterion A. Based on the Section III model, there are two broad domains (self, interpersonal) and four scales (identity, self-direction, empathy, intimacy) of functional impairment. Accordingly, in each of the four scales, individuals rate their level of functional impairment on several items which represent the four traits of functional impairment included in the DSM–5 Section III model. Participants indicated how much they agreed with 5 items. Each item asks individuals to choose one of five statements of ascending severity ranging from 0 (no impairment) to 4 (extreme impairment). Responses are averaged, with higher scores indicating greater levels of self- and interpersonal impairment [42]. Reliability analysis (Cronbach’s α) indicated adequate internal consistency for total score (α = .79). Due to brevity of the items in each subscale, we did not calculate internal consistency for the subscales.
Personality Inventory for DSM-5 Short Form (PID-5-SF). The PID-5-SF [43] is a 100-item self-report inventory, that was extracted from the original 220-item PID-5 [44], developed to assess the five DSM–5 Section III personality domains and their respective traits found in Criterion B. PID-5-SF covers the 25 personality traits facets as well as the five DSM-5 Section III personality domains (Negative Affectivity, Detachment, Antagonism, Disinhibition, and Psychoticism). Participants are asked to rate the extent to which they agree with each item using a Likert scale ranging from 0 (very false or often false) to 3 (very true or often true). The PID-5 is frequently used in research and has good psychometric properties (e.g. [45]). Reliability analysis (Cronbach’s α) indicated good internal consistency for each trait scale (α’s = .80 for hostility and separation insecurity; .84 for impulsivity and risk-taking; and .86 for anxiousness, emotional liability, and depressivity).
Beck Depression Inventory (BDI). The BDI [46] is a 21-item questionnaire developed to assess multiple aspects of depression, including cognitive, emotional, behavioral, and physical domains. For each item on the BDI, participants selected from one of four statements. Items were summed to create an overall score ranging from 0-63 with a higher score indicating greater experiences of symptoms of depression. A score of 10 or more is considered indicative of clinically significant levels of depressive symptoms [46]. Reliability analysis (Cronbach’s α) indicated good internal consistency (α = .91).
Dissociative Experiences Scale (DES). The DES [47] is a 28-item self-report questionnaire developed to measure dissociative experiences. Each item describes a kind of experience that the subjects may have had. Participants rate their agreement with each item using a 10-point scale (Never = 0% to Always = 100%). Items are summed to create a total score. The total is multiplied by 10, then divided by 28 to calculate the average score. Studies have shown that a score higher than 15 needs more investigation to diagnose dissociation, a score higher than 30 indicates the high probability of dissociative disorders and PTSD, and scores over 40 express high probability of dissociative identity disorder [47]. Reliability analysis (Cronbach’s α) indicated good internal consistency for the total score (α = .93).
Traumatic Life Events Questionnaire (TLEQ). The TLEQ [48] is a 23-item yes/no questionnaire assigned to assess the history of 22 potentially traumatic events (ranging from physical abuse, being stalked, or robbery involving a weapon, to exposure to warfare) and a 23rd category of “other events” with examples. For every event that is endorsed, participants are asked to answer questions about the event’s frequency (ranging from “never” to “more than 5 times”), the severity of the traumatic experience based on DSM-IV diagnostic criteria (i.e. whether they felt intense fear, helplessness, or horror during the event), and characteristics of perpetrator (e.g. stranger, relative). Participants are also asked to nominate the event that caused them the most distress and their age when the event first and last occurred.
Validity indicator. Given that these data were collected on computer and that none of the measures used in this study had built-in validity indicators, it was important to determine if individuals were responding appropriately to the item content (i.e. not randomly responding). For this reason, we distributed a series of questions throughout the protocol, to which most individuals should respond in a non-affirmative way, such as “Do you live on Mars?” or “I am allergic to water.” None of the participants were excluded based on their responses to the validity indicator questions.
False memory (DRM) task: material and procedure
We used 12 word lists in the DRM task (3 each for BPD-related, trauma-related, negative, and positive lists). The BPD-related word lists (e.g. Loneliness, Self-harm) were chosen based on the findings of our preliminary experiment to develop DRM word lists for BPD (see online supplementary materials). Two each of the negative (e.g. Lie, Sick) and positive (e.g. Happy, Nice) lists were taken from Zhang et al. [49] and one each from Palmer and Dodson [50]; the trauma-related lists (e.g. Rape, Sex) were taken from Goodman et al. [51]. Each word list was 10-words long and the words on each list were shown in the same serial order to all participants. All of the words in each word list were shown in the descending order of their BAS values so that the first word on the list was the one most associated with the critical lure, the second word was the one next most associated, and so on.
We used E-Prime to present participants with the word lists (120 words in total). Prior to conducting the experiment, the experimenter explained the procedure and then participants completed a neutral-word practice list. The presentation order of the word lists was partially counterbalanced across participants; that is, half of the participants learned the three BPD-related word lists, followed by the three positive lists, the three trauma-related lists, and the three negative word lists, and the other half of the participants learned the word lists in the reverse order (see also [49]). Each word was presented for 2 seconds on the computer screen and after all of the words in the word list were presented, a message came up warning the participant that the next word list was about to appear. Participants were asked to remember all 12 word lists. Between the learning phase and the test phase, participants completed a 5-minute subtraction exercise.
Following Zhang et al. [49], the recognition test contained 36 studied words from serial positions 1, 4, and 8 of each word list; and 36 non-studied words. The non-studied words consisted of 12 critical lures and 24 non-related lures that were unrelated to any list words; the positive and negative non-related lures were taken from Brainerd et al. [52] and the neutral ones were taken from Roediger et al. [53]. The presentation order of the 72 test words was completely random. Participants were asked to make an “Old” decision if they thought a word had been presented in the study phase and to make a “New” decision if they thought a word had not been presented in the study phase.
Analytic strategy
We analyzed all of the data using IBM SPSS Statistics 25 (IBM, USA) and Hayes’ (2018) PROCESS macro v3.4 for SPSS. For all of the analyses, p < .05 was set as the level of significance. The findings of the current study are presented in two major sections. In the first section, we conducted a series of bivariate correlations to examine false memory in 1) the traditional model of BPD, as the index of full disorder of BPD (i.e. SCID-II and MSI-BPD), and 2) BPD impairment in personality functioning (i.e. Criterion A), as well as 3) BPD personality traits (i.e. Criterion B), as the indices of AMPD-BPD. In the second section, we tested four separate mediation models to examine the indirect effect of BPD on false memory. To do this, we used Hayes’ PROCESS to run a series of simple mediation models (Model 4, 10,000 bootstrap resamples); statistical significance for an indirect effect is assumed if zero is not included in the bias-corrected 95% confidence interval for the indirect effect.