Development of a Brief Version of the Dissociative Symptoms Scale and the Reliability and Validity of DSS-B Scores in Diverse Clinical and Community Samples

Abstract

The Dissociative Symptoms Scale (DSS) was developed to assess moderately severe types of dissociation (depersonalization, derealization, gaps in awareness and memory, and dissociative reexperiencing) that would be relevant to a range of clinical populations, including those experiencing trauma-related dissociation. The current study used data from 10 ethnically and racially diverse clinical and community samples (N = 3,879) to develop a brief version of the DSS (DSS-B). Item information curves were examined to identify items with the most precision in measuring above average levels of the latent trait within each subscale. Analyses revealed that the DSS-B preserved the factor structure and content domains of the full scale, and its scores had strong reliability and validity that were comparable to those of scores on the full measure. DSS-B scores showed high levels of measurement invariance across ethnoracial groups. Results indicate that DSS-B scores are reliable and valid in the populations studied.

Keywords

dissociation dissociative experiences traumatic stress PTSD measurement short form Dissociative Symptoms Scale

Dissociation is multifaceted and encompasses a range of phenomena that vary in severity and impact on functioning. It has been proposed that clinically relevant dissociative experiences include distortions in perception of self (depersonalization) or the environment (derealization), gaps in awareness, memory, or identity, and disruptions in subjective experience accompanied by involuntary and unwanted intrusions (Cardeña & Carlson, 2011). Moreover, the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association [APA], 2013) defined dissociation as the “disruption of and/or discontinuity in the normal integration of consciousness, memory, identity, emotion, perception, body representation, motor control, and behavior.” These symptoms are associated with significant distress, impairment, and poor treatment response (Kleindienst et al., 2016; Lanius et al., 2012) and may be experienced by individuals with a variety of psychiatric disorders (Lyssenko et al., 2018). Dissociative symptoms are most prevalent and severe in patients with dissociative disorders, such as dissociative identity disorder, and are also very common and elevated in those with posttraumatic stress disorder (PTSD; Carlson et al., 2012; Waelde et al., 2009) and acute stress disorder (Cardeña & Carlson, 2011). High levels of dissociation have also been associated with disordered attachment experiences (Dutra et al., 2009; Maniglio, 2009), exposure to traumatic stress (Carlson et al., 2012), complex PTSD (Hyland et al., 2020), and borderline personality disorder (Scalabrini et al., 2017). The relevance of dissociation to traumatic stress is further underscored by the recent addition of a dissociative subtype to the DSM diagnostic criteria for PTSD (APA, 2013), and by findings of dissociative symptoms being more strongly associated with the severity of PTSD than depression and general psychological distress (Eidhof et al., 2019). Research also indicates that dissociation is essentially unrelated to some psychological issues, such as alcohol use (e.g., Chana et al., 2021). This may be explained by dissociation and alcohol consumption serving similar emotion regulatory functions for responding to intense emotions and stressors (Cardeña & Carlson, 2011; Centanni et al., 2019). The relevance and complexity of dissociation’s relationships with psychological problems support the need for a brief measure of dissociation for use in research and clinical care.

The Dissociative Symptoms Scale (DSS) is a 20-item self-report measure that was developed to assess moderately severe types of dissociation (depersonalization, derealization, gaps in awareness and memory, and dissociative reexperiencing) that would be relevant to a range of clinical populations, including those experiencing trauma-related dissociation (Carlson et al., 2018). It does not include types of dissociation that are commonly observed in non-clinical populations or relatively rare and experienced mostly by people with dissociative disorders. The DSS assesses the frequency of dissociative symptoms in the past week. The phenomena assessed are neither “in the moment” measures of states nor expected to be stable over long periods of time (i.e., months or years). However, the total frequency of dissociative symptoms is expected to be fairly stable from week to week. Data from four clinical and five community samples were used to select items for the DSS, and initial psychometric analyses demonstrated the reliability and validity of DSS scores to measure dissociation in similar populations (Carlson et al., 2018). The DSS has been used to measure dissociative experiences in other clinical (Hyland et al., 2018) and non-clinical samples (Chiu et al., 2017), in adults with PTSD (Műllerová et al., 2016), and to evaluate the validity of scores on other measures of dissociation (Eidhof et al., 2019). In addition, scores on DSS items have shown excellent correspondence with dissociative symptoms measured in real time in daily life (Carlson et al., 2016).

The DSS was developed to address the limitations of other measures of dissociation. For example, the Dissociative Experiences Scale (DES; Bernstein & Putnam, 1986) contains many items that assess mild forms of dissociation that are frequently endorsed in non-clinical samples and some items that assess severe dissociative experiences that are rarely endorsed and primarily relevant to those with dissociative disorders (Carlson & Putnam, 1993). Results of item response theory (IRT) analyses comparing the information provided by DES and DSS items have demonstrated that DSS items assess symptoms with latent trait severity levels that fall between mild and severe DES items (Carlson et al., 2018). Another advantage of the DSS is that it assesses a range of clinically relevant dissociative experiences rather than focusing on a specific type (e.g., depersonalization), which is important given the diversity and complexity of dissociative symptoms. Armour and colleagues (2014) found that depersonalization and derealization were no more or less related to PTSD symptom clusters than other types of dissociative experiences studied. Although DSS scores do seem to provide a valid and reliable assessment of dissociation in populations studied, a briefer measure may be valuable for use in many research and clinical contexts when time for assessment is limited or the number of scale items needs to be minimized, such as studies assessing a large number of variables or measuring dissociation repeatedly over time (e.g., daily diary, longitudinal, or treatment studies).

The objectives of this study were to develop a brief version of the DSS (DSS-B) that (a) included items that were the most precise in measuring clinically relevant levels of the types of dissociation assessed, (b) preserved coverage of the content domains (subscales) and four-factor structure of the DSS, (c) produced scores that demonstrate reliability and validity comparable to the full measure, and (d) measured the same construct across racial/ethnic groups. This research involved secondary analyses on data drawn from nine clinically and ethnically diverse samples in which the full DSS was administered, including trauma-exposed adults, patients with trauma-related disorders, and student and community samples, as well as a tenth sample in which only DSS-B items were assessed.

To select items for the DSS-B, we conducted IRT analyses to identify items with the greatest precision in measuring above average levels of the latent trait. Structural analyses were used to evaluate whether the factor structure and content domains of the full measure were preserved. We also examined the internal consistency and temporal stability of DSS-B scores. The total frequency of dissociative symptoms was expected to be fairly stable from week to week with correlations in the range of .5 to .7, similar to the full DSS (Carlson et al., 2018). The validity of DSS-B scores was investigated by testing for elevations in clinical compared to non-clinical samples, among individuals with probable PTSD versus without probable PTSD, and in adults who were admitted to a trauma center as a result of injury from interpersonal or intentional violence as opposed to a vehicle collision, fall, or other accident. Convergent and discriminant validity were assessed using correlations of DSS-B scores with other measures. Scores on the DSS-B were expected to correlate most strongly with scores on measures of dissociation and PTSD symptoms, moderately with scores on measures of depression, perceived stress, and trauma exposure, and least strongly or essentially uncorrelated with scores on measures of alcohol use. To evaluate whether the validity of DSS-B scores was comparable to that of scores on the full measure, we tested for differences in the correlations of DSS-B and DSS scores with other measures. Measurement invariance analyses were performed to examine whether DSS-B scores measured the same construct across four ethnoracial groups.

Method

Data Sources

The current study utilized a combined dataset of 3,401 individuals from nine samples in which the full DSS was administered and a 10th sample of 478 individuals who completed the brief version (DSS-B). Demographic characteristics are shown in Tables 1 and 2 for the first nine and 10th samples, respectively. Informed consent was obtained in all samples. Seven of the samples were used to develop the full DSS, which included veterans in a Veterans Affairs (VA) residential PTSD program (Residential Veterans; n = 240), adults with PTSD in a medication study (PTSD Diagnosis; n = 33), adults assessed one to 14 days after a traumatic injury to self or a loved one that resulted in hospitalization (Trauma Center; n = 149), veterans in a VA residential homeless program (Homeless Veterans; n = 115), students at a Midwestern university (Midwest Students; n = 210), students at a Western state university (Western Students; n = 406), and adults in university, college, and shopping areas (Community Adults; n = 447). Additional information about these samples and relevant study methods is elsewhere (Carlson et al., 2018).

Table 1.

Demographic Characteristics and DSS-B Total Score Descriptive Statistics by Sample and Split of Data in the Nine Samples That Completed the Full DSS.

Sample	First split half (item selection)					Second split half (validation)
Sample	N	Female ^a	White/ European American ^a	Age ^b	DSS-B ^b	N	Female ^a	White/ European American ^a	Age ^b	DSS-B ^b
PTSD Diagnosis	17	10 (59%)	14 (82%)	43 (8.4)	7.0 (5.1)	16	11 (69%)	14 (88%)	42 (12.7)	6.5 (4.4)
Residential Veterans	120	12 (10%)	72 (63%)	52 (5.3)	6.5 (5.8)	120	14 (13%)	58 (53%)	51 (6.1)	7.1 (5.4)
Homeless Veterans	57	3 (5%)	23 (41%)	46 (6.8)	6.5 (5.9)	58	2 (3%)	30 (52%)	44 (5.7)	5.8 (5.6)
Trauma Center	74	35 (47%)	41 (57%)	41 (12.7)	3.8 (4.0)	75	45 (60%)	54 (75%)	44 (13.9)	4.6 (5.1)
Midwest Students	105	105 (100%)	64 (61%)	18 (1.0)	4.6 (3.4)	105	105 (100%)	63 (60%)	19 (1.1)	5.0 (3.5)
Western Students	203	118 (58%)	49 (24%)	20 (2.4)	3.6 (3.9)	203	105 (52%)	42 (21%)	20 (3.5)	3.2 (3.5)
Northeast Students	776	481 (62%)	53 (7%)	20 (2.3)	2.8 (3.5)	777	480 (62%)	52 (7%)	20 (2.3)	2.9 (3.2)
Community Adults	224	141 (63%)	130 (59%)	24 (8.6)	3.1 (4.3)	223	134 (60%)	135 (61%)	25 (10.4)	2.9 (4.3)
Caregivers	124	104 (92%)	20 (17%)	33 (7.6)	2.7 (3.6)	124	103 (94%)	19 (16%)	32 (9.3)	2.7 (4.0)
Combined Sample	1700	1009 (60%)	466 (28%)	25 (11.4)	3.5 (4.2)	1701	999 (60%)	467 (28%)	25 (11.6)	3.5 (4.1)

Note. DSS-B = Dissociative Symptoms Scale-Brief version; PTSD = posttraumatic stress disorder.

n (%). ^b M (SD).

Table 2.

Sample Characteristics and Descriptive Statistics for DSS-B Total Scores in the COTI DSS-B Sample (N = 478).

Gender	n or M	(%) or (SD)
Male	312	(65.3%)
Female	166	(34.7%)
Race/Ethnicity
White/European American	159	(33.3%)
Black/African American	186	(38.9%)
Hispanic/Latinx American	115	(24.1%)
Other Race/Ethnicity	18	(3.8%)
Age	40.23	(13.95)
Injury Mechanism
Violence (Intentional/Interpersonal)	116	(24.3%)
Vehicle Collision	201	(42.1%)
Falls/Other Accidents	171	(35.8%)
DSS-B total score	4.02	(5.86)

Note. DSS-B = Dissociative Symptoms Scale-Brief version; COTI = consequences of traumatic injury.

The eighth sample included primary caregivers of young children (ages 2–6) brought to a university-affiliated clinical research program at an urban public hospital between 2003 and 2011 (Caregivers; n = 248; Hagan et al., 2017). Caregivers were seeking outpatient mental health services for their children due to interpersonal trauma exposure (e.g., community violence, domestic violence, caregiver death). Most were the child’s biological mother (81%), and the rest were the biological father (9%) or another adult caregiver (10%). Exclusion criteria included severe intellectual disability, suicidal/homicidal ideation, severe psychosis, and active substance abuse. Baseline assessments were administered in participants’ native language (English or Spanish) by mental health providers using interview or paper format.

The ninth sample consisted of undergraduate students from a large urban public university system in the Northeastern U.S. (Northeast Students; n = 1,553; Anglin et al., 2014). Northeast Students were recruited through a participant recruitment website from 2011 to 2016 and were eligible for the study if they were between the ages of 18-29 years and self-identified as Black/African American/African descent or as a first- or second-generation immigrant. Participants completed self-report questionnaires on a computer in a research lab and were given course credit for participation.

The 10th sample included adults who were hospitalized due to a serious injury and completed the DSS-B 3 months later as part of a longitudinal study on consequences of traumatic injury (COTI DSS-B; n = 478). The sample was recruited from two large trauma centers in the Midwest and the Southern U.S. between 2017 and 2020. Adults aged 18–65 years without a cognitive impairment were eligible for the study if they suffered a physical injury that was not related to a preexisting condition or intentional self-harm, were admitted to the hospital for at least 24 hours, and were not in police custody. Assessments were conducted in English or Spanish and administered either in person or by phone. Measures were translated into Spanish by a bilingual translator, then reviewed by two other bilingual translators, and the three translators reviewed and reconciled any discrepancies. The present analyses focus on data from the 3-month follow-up; although information on the type of injury mechanism and demographic characteristics were collected at baseline.

Measures

The full DSS was administered to participants in all samples except for the COTI DSS-B sample, in which case the brief version of the DSS (DSS-B) was administered instead. All other measures were administered in a subset of the samples, as indicated below. Internal consistency of scores for measures in the current data are reported below in terms of coefficient alpha with 95% confidence intervals (CIs), which was estimated using the latent variable approach with 5000 bootstrap draws (Raykov & Marcoulides, 2015). Descriptive statistics for all measures are in the Supplemental Materials (Tables S1–S3).

The DSS is a 20-item self-report measure that assesses the frequency of dissociative symptoms in the past week (Carlson et al., 2018). The instructions at the top read: “For each statement below, check a box to show how much each thing has happened to you in the past week.” Response options for each item range from 0 = “not at all” to 4 = “more than once a day.” The DSS includes four subscales that cover the domains of depersonalization and derealization (e.g., “Things around me seemed strange or unreal”), sensory misperceptions (e.g., “I heard something that I know really wasn’t there”), gaps in awareness and memory (e.g., “I suddenly realized that I hadn’t been paying attention to what was going on around me”), and cognitive-behavioral reexperiencing (e.g., “I reacted to people or situations as if I were back in an upsetting time in my past”). Total scores are calculated as the sum of item responses. DSS scores have shown strong internal consistency, temporal stability, and validity in four clinical and five non-clinical samples (Carlson et al., 2018). In the nine samples that completed the full DSS, internal consistency of DSS scores was strong ( $\hat{α}$ = .922; 95% CI = [.915, .928]).

The DES (Bernstein & Putnam, 1986; Carlson & Putnam, 1993) was administered in the Residential Veterans and Western Student samples ( $\hat{α}$ = .946; 95% CI = [.936, .953]). The DES is a 28-item measure that assesses the frequency of dissociative symptoms, including amnesia, depersonalization, derealization, absorption, and imaginative involvement. Total scores reflect the mean of item responses and can range from 0 to 100. DES scores have been validated in clinical and community samples and in a wide range of diagnostic groups (Bernstein & Putnam, 1986; Carlson & Putnam, 1993). Previous research has shown good internal consistency of DES scores in college students and patients with dissociative disorders (α = .95; Frischholz et al., 1990).

The PTSD Checklist-Civilian (PCL-C) and Military (PCL-M) versions (Weathers et al., 1993) were used to assess PTSD symptoms in the Midwest Students, Residential Veterans, and Homeless Veterans samples ( $\hat{α}$ = .974; 95% CI = [.970, .977]). The PCL is a 17-item measure that assesses how much an individual has been bothered by each Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; APA, 1994) PTSD symptom in the past month on a 5-point scale with response options that range from 1 = “not at all” to 5 = “extremely.” The PCL-C assesses PTSD symptoms related to any traumatic stressor, whereas the PCL-M assesses PTSD symptoms related to traumatic stressors that occurred during military service. Total scores are calculated as the sum of item responses. PCL scores have shown good internal consistency (αs = .92 to .98) and convergent validity with other measures of PTSD in samples of veterans, college students, and individuals with a range of medical and psychiatric diagnoses (Wilkins et al., 2011). Individuals with probable PTSD were identified using a cut score of 44 on the PCL (McDonald & Calhoun, 2010).

The PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013) was used to assess PTSD symptoms following traumatic injury in the COTI DSS-B sample ( $\hat{α}$ = .938; 95% CI = [.930, .946]). The PCL-5 is a 20-item measure that assesses how much an individual has been bothered by each DSM-5 PTSD symptom on a 5-point scale that ranges from 0 = “not at all” to 4 = “extremely.” Total scores are calculated as the sum of item responses. PCL-5 scores have shown good internal consistency (αs = .94 to .95) and convergent validity with other measures of PTSD in samples of trauma-exposed college students and mental health service users (Blevins et al., 2015; Roberts et al., 2021). Individuals with probable PTSD were identified using a cut score of 33 on the PCL-5 (Bovin et al., 2016).

The Screen for Posttraumatic Stress Symptoms (SPTSS; Carlson, 2001) was used to assess PTSD symptoms in the Trauma Center and Community Adults samples ( $\hat{α}$ = .905; 95% CI = [.887, .920]). The SPTSS is a 17-item measure that assesses the frequency of DSM-IV PTSD symptoms in the past week using a 5-point scale with response options that range from 0 = “not at all” to 4 = “more than once every day.” Item responses are summed up to obtain the total score. SPTSS scores have shown good internal consistency (α = .91) and concurrent validity in predicting PTSD diagnosis in samples of psychiatric patients and veterans (Carlson, 2001; Caspi et al., 2007). A cut score of 20 on the SPTSS was used to identify individuals with probable PTSD based on results of the Clinician Administered PTSD Scale (see Carlson et al., 2017).

The Davidson Trauma Scale (DTS; Davidson et al., 1997) was used to assess PTSD symptoms in the Caregivers sample ( $\hat{α}$ = .943; 95% CI = [.930, .953]). The DTS assesses the frequency and severity of the 17 DSM-IV PTSD symptoms in the past week on a 5-point scale with response options that range from 0 = “not at all” to 4 = “every day” or “extremely distressing.” DTS scores have shown good reliability, internal consistency (α = .99), and convergent validity with scores on other PTSD measures in a sample of individuals exposed to different types of trauma (Davidson et al., 1997). For the current study, frequency and severity scores were summed up to create an overall score of PTSD symptoms. In addition, individuals with probable PTSD were identified using a cut score of 40 on the DTS (McDonald et al., 2014).

The Clinician Administered PTSD Scale (CAPS; Weathers et al., 2001) was used to assess PTSD symptoms in the PTSD Diagnosis and Residential Veterans samples ( $\hat{α}$ = .853; 95% CI = [.816, .883]). The CAPS is a structured interview for DSM-IV PTSD, which assesses the frequency and intensity of each symptom on a 5-point scale. Support for the internal consistency (αs = .87 to .95) and validity of its scores have been found in samples of individuals exposed to military and civilian trauma (Weathers et al., 1999, 2001). In the current study, frequency and intensity scores were summed up to obtain an overall severity score of PTSD symptoms. Individuals with probable PTSD were identified using a cut score of 45 on the CAPS (Weathers et al., 1999).

The Trauma History Screen (THS; Carlson et al., 2011) was used to measure trauma exposure in the Trauma Center, Homeless Veterans, Midwest Students, and Community Adults samples. The THS measures exposure to high magnitude stressor (HMS) events. Respondents indicate whether they have experienced each of 15 types of HMS events and how many times an event of that type has happened to them. Total HMS scores are the sum of frequencies for all HMS events. Total HMS scores from the THS have demonstrated good retest reliability and convergent validity in clinical and nonclinical samples (Carlson et al., 2011).

The Life Events Checklist (LEC; Gray et al., 2004) was used to measure trauma exposure in the Northeast Students sample. The LEC screens for exposure to 17 types of potentially traumatic events. Respondents indicate whether each type of event happened to them personally, whether they witnessed it happen to someone else, and whether they learned about it happening to someone close to them. Total scores were calculated as the number of items for which “it happened to you personally” was endorsed. LEC exposure reports have shown adequate temporal stability over one week and convergent validity with another trauma exposure measure in samples of college students and combat veterans (Gray et al., 2004).

The Beck Depression Inventory (BDI; Beck et al., 1961; Beck & Steer, 1993) is a 21-item measure that assesses attitudes and symptoms of depression on a 4-point scale. The BDI short-form (BDI-SF) comprises 13 BDI items that assess cognitive-affective symptoms of depression. The BDI was administered in the Caregivers sample ( $\hat{α}$ = .923; 95% CI = [.907, .936]), and the BDI-SF was administered in the Trauma Center sample ( $\hat{α}$ = .866; 95% CI = [.812, .901]). Total scores are obtained by summing item responses. BDI scores have shown moderate to high internal consistency (αs = .73 to .95) and strong concurrent validity with clinical ratings in clinical and nonclinical samples (Beck et al., 1988).

The Center for Epidemiologic Studies Depression Scale Revised (CESD-R-10; Kohout et al., 1993) was used to measure depression symptoms in the Northeast Students sample ( $\hat{α}$ = .794; 95% CI = [.778, .809]). The CESD-R-10 is a 10-item measure that assesses the frequency of depressive symptoms in the past week on a 4-point scale that ranges from 0 = “rarely or none of the time (less than one day)” to 3 = “all of the time (5-7 days).” Total scores are calculated by summing item responses. CESD-R-10 scores have demonstrated good psychometric properties, including adequate internal consistency (αs = .70 to .89) in clinical and community samples (Björgvinsson et al., 2013; Mohebbi et al., 2018).

The Patient Health Questionnaire (PHQ-8; Kroenke et al., 2009) was used to assess depression symptoms in the COTI DSS-B sample ( $\hat{α}$ = .893; 95% CI = [.875, .908]). For assessments conducted in Spanish, the Spanish translation of the PHQ-8 from Pfizer was used (www.phqscreeners.com/select-screener). The PHQ-8 assesses how often in the past 2 weeks an individual has been bothered by each symptom of major depression, except suicidal ideation. Respondents are asked to rate the frequency of each item using a 4-point scale with response options that range from 0 = “not at all” to 3 = “nearly every day.” The sum of item responses is calculated as a total score. PHQ-8 scores have shown good internal consistency (α = .88) and convergent validity with other measures of depression and health-related impairment in general population and psychiatric outpatient samples (Kroenke et al., 2009; Shin et al., 2019).

The Perceived Stress Scale (PSS; Cohen et al., 1983) is a measure that assesses perceived life stress in the past month on a 5-point scale that ranges from 0 = “never” to 4 = “very often.” The original PSS included 14 items (PSS-14), and a shortened, 10-item version (PSS-10) has also been created based on a factor analysis of PSS-14 scores in a large U.S. probability sample (Cohen & Williamson, 1988). The PSS-14 was administered in the Northeast Students sample ( $\hat{α}$ = .787; 95% CI = [.767, .804]), and the PSS-10 was administered in the Trauma Center sample ( $\hat{α}$ = .905; 95% CI = [.880, .923]). Item responses are summed to obtain a total score. Both versions have shown good reliability (αs = .84 to .91) and validity in college, psychiatric, and medical populations, although a review paper has concluded that psychometric properties of the PSS-10 are generally stronger (Lee, 2012).

The Alcohol Use Disorders Identification Test (AUDIT; Babor, 2000) is a 10-item measure that was used to assess problematic alcohol use in the Trauma Center sample ( $\hat{α}$ could not be calculated as item-level data were unavailable). Total scores are calculated by summing item responses and can range from 0 to 40. AUDIT scores have been shown to have adequate internal consistency (α = .75 to .94), to correlate strongly with other measures of alcohol use, such as quantity and frequency of consumption, and to have good predictive validity with alcohol use disorder and problematic drinking in samples of college students and medical and psychiatric patients (Allen et al., 1997).

Alcohol use was assessed in the COTI DSS-B sample using two items from the 2015 National Survey on Drug Use and Health (NSDUH; Center for Behavioral Health Statistics and Quality, 2016). The first item assessed the frequency of alcohol consumption in the past month with the question: “During the past 30 days, on how many days did you drink one or more drinks of an alcoholic beverage?” The typical quantity of alcohol consumed on drinking days in the past month was assessed with the question: “During the past 30 days, on the day/days that you drank, how many drinks did you usually have each day?”

Results

Item Selection

Each of the nine samples that completed the full DSS was randomly split in half, and the first half of these samples (Item Selection sample; n = 1,700) was used to select items for the DSS-B. The R package mirt (Chalmers, 2012) was used to conduct IRT analyses using the graded response model (GRM), which treats items as ordinal. Due to low endorsement of the highest response option (4 = “more than once a day”; 0.6% to 5.4% across DSS items), the top two options of each item were combined for the IRT analysis. To preserve coverage of the content domains and the four-factor structure of the DSS, two items were retained from each DSS subscale, and IRT analyses were performed for each subscale separately so that the selection of items within each subscale was not influenced by items from other subscales (Smith et al., 2000).

The local independence assumption was evaluated for each GRM model using the likelihood ratio (G²) statistic with Bonferroni correction (Chen & Thissen, 1997). The Bonferroni correction was chosen because it is recommended when the primary focus is on evaluating the universal null hypothesis (e.g., all tests of local dependence for item pairs are not significant; Perneger, 1998). Results of these tests revealed that each of the fitted GRMs contained at least one pair of items with a significant residual association. These residual associations were primarily between pairs of items with similar wording or structure, suggesting that they may be explained by method factors as opposed to conceptual similarities that were not adequately captured by the latent trait. Person fit was examined using Drasgow’s Z_h (Drasgow et al., 1985) to determine whether there were improbable participant response patterns that could impact results. The absolute value of Z_h exceeded 3.5 for fewer than 0.5% of individuals, suggesting that overall, the latent trait was an appropriate measure of severity given participant response patterns and item estimates. Due to the small proportion of improbable response patterns, all participants were retained in the analysis; however, results did not change if these participants were excluded.

GRM parameter estimates were used to derive item information curves (IICs) so that each item’s precision in measuring the latent trait (item information) could be examined across the latent trait continuum (Baker, 2001). Because the latent trait has a mean of 0 and SD of 1, values along the continuum describe the severity of the latent trait relative to the mean in units of standard deviation (i.e., z-score). Items with the greatest precision in measuring above average levels of the latent trait were selected for the DSS-B by estimating the area under each item’s information curve (AUIIC) between the latent trait (θ) values of zero (the mean) and four (4 SD above the mean) using adaptive quadrature. This range of the latent trait was assumed to reflect clinically relevant levels of the content domains assessed by the DSS given the mix of clinical, student, and community samples in the data. Item information results from the IRT analyses are summarized in Table 3. The first column of numerical values shows each item’s AUIIC between the latent trait mean and 4 SD above the mean. Plots depicting these AUIICs for each item are included in the Supplemental Materials (Figures S1 and S2). Peak information indicates an item’s highest information value, which reflects its maximum level of precision in measuring the latent trait. These values can be compared among items measuring the same latent trait, with higher values reflecting greater measurement precision. The next column reports the latent trait severity that corresponds to the peak information value (i.e., severity of the latent trait that the item is most precise in measuring). The last three columns describe the range of latent trait severities where item information is greater than one (e.g., maintains at least a moderate level of measurement precision). These latent trait values represent locations (i.e., severity levels) on the latent trait continuum in units of SD from the mean and can be interpreted as z-scores. The two items with the largest AUIIC values from each subscale (in boldface) were selected for the DSS-B. These items also had the highest peak information values within each subscale and demonstrated moderate to high levels of precision in measuring severities of the latent trait between -0.835 and 3.287 SD from the mean. The R code for this analysis is provided in the Supplemental Materials. A copy of the DSS-B measure is provided in the appendix.

Table 3.

Summary of Item Response Theory (IRT) Analysis Item Information Results From the First Half of the Nine Samples That Completed the Full DSS.

DSS subscale	AUIIC(0 < θ < 4)	Peak info	θ at peak info	Range of θ where Info > 1
DSS subscale	AUIIC(0 < θ < 4)	Peak info	θ at peak info	Min	Max	Max-Min
Depersonalization/Derealization
1. Body felt strange	4.689	1.719	1.968	0.113	2.853	2.740
2. Things strange	6.594	2.737	1.845	−0.013	2.825	2.838
5. Outside myself	4.584	1.812	2.095	0.652	3.024	2.372
8. In a movie, not real	5.162	1.963	2.134	0.425	3.069	2.644
13. Wasn’t myself	2.974	0.988	1.795	—	—	—
16. Body parts distorted	2.511	0.841	2.437	—	—	—
Sensory Misperceptions
6. Heard something not there	4.365	1.591	2.379	0.819	3.250	2.431
7. Upset, can’t remember	3.288	1.012	2.559	2.290	2.794	0.504
9. No pain when hurt	2.633	0.818	2.524	—	—	—
14. Saw something, seemed real	5.497	2.343	2.297	0.833	3.287	2.454
20. Smelled something not there	3.137	1.097	2.488	1.540	2.992	1.452
Gaps in Awareness/Memory
11. Staring into space	2.797	1.115	1.169	−0.397	1.806	2.203
12. Couldn’t remember	3.710	1.237	1.920	0.251	2.677	2.426
15. Not paying attention	4.365	1.926	1.298	−0.835	2.315	3.150
18. So focused, lost track	3.922	1.556	1.467	−0.484	2.379	2.863
19. Gaps in memory	2.977	0.969	1.816	—	—	—
Cognitive/Behavioral Reexperiencing
3. Reminded, spaced out	3.311	1.269	1.456	−0.491	2.247	2.738
4. Lost control, back in past	6.268	2.480	1.863	−0.038	2.870	2.908
10. Memory came back	4.187	1.499	1.811	−0.083	2.689	2.772
17. Reacted as if in past	5.305	1.919	2.003	0.202	3.024	2.822

Note. Bolded items were selected for the DSS-B (Dissociative Symptoms Scale-Brief version); θ = latent trait severity; AUIIC (0 < θ < 4) = area under the item information curve between the latent trait values of zero (the mean) and 4 SD above the mean; Peak information = highest point of the information curve.

Factor Structure

Psychometric properties of the 8-item DSS-B were then examined using the second half of the nine samples that completed the full DSS (Validation sample; n = 1,701) and the 10th sample that completed the DSS-B (COTI DSS-B sample; n = 478). Confirmatory factor analysis (CFA) was used to examine whether the DSS-B preserved the factor structure and content domains of the DSS. A correlated four-factor CFA model was fitted to DSS-B items using the WLSMV estimator for ordinal indicators in Mplus version 8.4 (Muthén & Muthén, 2019). The factor structure was assessed by examining model fit and standardized estimates for factor loadings and correlations. Model fit was evaluated using the chi-square statistic, which provided a test of the null hypothesis of an exact fit between the specified model and the data. Approximate fit indices and item residual correlations were also used to evaluate model fit. Higher comparative fit index (CFI) values, and lower root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) values were interpreted as better fit.

In the Validation sample, the chi-square test was significant, χ² (14) = 56.485, p < .001, suggesting that the correlated four-factor CFA model deviated from exact fit to the data. However, in the COTI DSS-B sample, the chi-square test was not significant, χ² (14) = 21.486, p = .090, suggesting the model adequately represented relationships among DSS-B items in the data. In both cases, approximate fit indices suggested excellent fit (Validation sample: RMSEA = 0.042, 90% CI = [0.031, 0.054]; CFI = 0.994; SRMR = 0.016; COTI DSS-B sample: RMSEA = 0.033, 90% CI = [0.000, 0.060]; CFI = 0.998; SRMR = 0.014), and the absolute values of item residual correlations were all below 0.06. Thus, overall, model fit results suggested that the four-factor model adequately fit the structure of DSS-B item-level data. In the Validation and COTI DSS-B samples, respectively, average standardized loading estimates were .839 (range: .789 to .896) and .892 (range: .834 to .941), and average inter-factor correlations were .707 (range: .646 to .816) and .816 (range: .685 to .921). Path diagrams of the CFA model with standardized parameter estimates are in the Supplemental Materials (Figure S3).

Reliability Analyses

To assess composite reliability, categorical omega was estimated using the R package MBESS (Kelley & Pornprasertmanit, 2016), and coefficient alpha was estimated using the latent variable approach in Mplus (Raykov & Marcoulides, 2015). CIs were obtained using 5000 bootstrap samples. According to categorical omega, internal consistency of DSS-B scores was strong (Validation sample: ${\hat{ω}}_{C}$ = .886, 95% CI = [.866, .904]; COTI DSS-B sample: ${\hat{ω}}_{C}$ = .943, 95% CI = [.915, .969]). Coefficient alpha estimates for internal consistency of DSS-B scores also indicated adequate reliability but were slightly lower than the estimates for categorical omega (Validation sample: $\hat{α}$ = .829, 95% CI = [.805, .849]; COTI DSS-B sample: $\hat{α}$ = .883, 95% CI = [.856, .906]). Temporal stability was examined with Spearman correlations, which indicated that DSS-B scores were moderately stable over a 1-week period in both the Midwest Student (r = .664, 95% CI = [.504, .780]) and Trauma Center (r = .574, 95% CI = [.340, .741]) samples.

Validity Analyses

Descriptive statistics for DSS-B total scores are provided in Table 1 for each of the nine samples that completed the full DSS and in Table 2 for the COTI DSS-B sample. In the Validation sample (second half of the nine samples that completed the full DSS), DSS-B scores for individuals in clinical (n = 269; 16%) and non-clinical (n = 1,432; 84%) samples and for individuals with (n = 262; 36%) and without (n = 460; 61%) probable PTSD were compared using Welch’s independent samples t-tests. As expected, DSS-B total scores were higher in clinical samples compared to non-clinical samples, t(314.41) = 8.910, p < .001, d = 0.432, 95% CI = (0.336, 0.528), and higher among individuals with probable PTSD relative to those without probable PTSD, t(349.67) = 11.577, p < .001, d = 0.862, 95% CI = (0.709, 1.014). DSS-B scores were also compared for individuals with (n = 156; 33%) and without (n = 322; 67%) probable PTSD in the COTI DSS-B sample. Similar to the Validation sample, DSS-B total scores in the COTI DSS-B sample were higher among individuals with probable PTSD relative to those without probable PTSD, t(181.08) = 12.213, p < .001, d = 1.117, 95% CI = (0.924, 1.310). Item-level analyses indicated that in both samples, individuals with probable PTSD endorsed significantly higher levels of dissociation on all DSS-B items compared to those without probable PTSD (all ps < .001). Finally, in the COTI DSS-B sample, DSS-B scores were higher among individuals who were injured as a result of interpersonal or intentional violence (n = 116; 24%) compared to those who were injured due to a vehicle collision, fall, or other accident (n = 362; 76%), t(154.84) = 4.140, p < .001, d = 0.379, 95% CI = (0.198, 0.559).

Convergent and discriminant validity of DSS-B scores were evaluated by calculating Spearman correlations for the relationships of DSS-B and DSS scores with scores for other measures of the same, similar, and dissimilar constructs. CIs and Holm-adjusted p-values were obtained for correlations (Holm, 1979). Holm’s method was used because it is less conservative than the Bonferroni correction and more appropriate when conducting multiple tests of pre-planned hypotheses (Wright, 1992). Zou’s (2007) CI method for comparing two overlapping dependent correlations was used to compare the correlations of DSS-B and DSS scores with scores on other measures using the R package cocor (Diedenhofen & Musch, 2015). Correlations of DSS-B and DSS total scores with scores for other measures administered in the Validation samples are shown in Table 4, and correlations of DSS-B scores with scores for other measures administered in the COTI DSS-B sample are in Table 5. In Table 4, a weighted (by sample size) average correlation was calculated for each construct that was assessed with multiple measures (e.g., PTSD). Overall, DSS-B scores tended to be more strongly correlated with scores on measures of dissociation, PTSD, and depression than with scores on measures of perceived stress and trauma exposure, and there was essentially no correlation between scores on the DSS-B and measures of alcohol use. Correlations with other measures tended to be slightly lower when DSS-B scores were used instead of scores for the full DSS (see Table 4). However, Zou’s (2007) CI method indicated that most differences between correlations of DSS-B and DSS scores with other measures were not statistically significant (only 4 of 13 differences were significant; 31%). The maximum difference between DSS-B and DSS correlations (r_diff) was .07, indicating a very small effect size for the difference between correlations with other measures.

Table 4.

Correlations of DSS-B and DSS Scores With Other Measures in the Second Half of the Samples That Completed the Full DSS.

Other measures	Sample(s)	n	DSS-B		DSS		Zou’s (2007) test ^b
Other measures	Sample(s)	n	r	(95% CI)	r	(95% CI)	r _diff	(95% CI)
DES	RV, WS	236	.621***	[.536, .694]	.643***	[.562, .713]	.022	[−.016, .062]
PTSD ^a	—	—	.563***	[.463, .648]	.615***	[.522, .693]
SPTSS	TC, CA	298	.681***	[.615, .737]	.729***	[.671, .778]	.048**	[.017, .084]
DTS	CG	105	.536***	[.384, .660]	.590***	[.449, .702]	.054	[−.019, .134]
PCL	RV, HV, MS, WS	304	.471***	[.379, .554]	.541***	[.456, .616]	.070***	[.034, .110]
CAPS	PD, RV	67	.498***	[.293, .659]	.479***	[.270, .645]	−.019	[−.101, .058]
Depression ^a	—	—	.510***	[.436, .576]	.557***	[.486, .618]
BDI-SF	TC	75	.602***	[.475, .704]	.604***	[.478, .706]	.002	[−.077, .081]
BDI	CG	122	.498***	[.305, .651]	.529***	[.344, .675]	.031	[−.040, .105]
CESD-R-10	NS	777	.497***	[.442, .548]	.552***	[.501, .599]	.055***	[.030, .081]
Perceived Stress ^a	—	—	.401***	[.327, .470]	.450***	[.378, .514]
PSS-14	NS	777	.406***	[.345, .463]	.456***	[.398, .510]	.050***	[.023, .078]
PSS-10	TC	75	.356**	[.141, .540]	.387**	[.175, .564]	.031	[−.053, .118]
Trauma Exposure ^a	—	—	.185***	[.109, .260]	.204***	[.128, .278]
LEC	NS	777	.210***	[.141, .276]	.222***	[.154, .288]	.012	[−.017, .041]
HMS	HV, TC, MS, CA	449	.144**	[.052, .233]	.173**	[.082, .262]	.029	[−.005, .063]
AUDIT	TC	74	−.048	[−.274, .182]	.019	[−.210, .247]	.067	[−.019, .152]

Note. DSS = Dissociative Symptoms Scale; DSS-B = DSS-Brief version; CI = confidence interval; DES = Dissociative Experiences Scale; RV = Residential Veterans; WS = Western Students; PTSD = posttraumatic stress disorder; SPTSS = Screen for Posttraumatic Stress Symptoms; TC = trauma center; CA = community adults; DTS = Davidson Trauma Scale; CG = caregivers; PCL = PTSD Checklist; HV = Homeless Veterans; MS = Midwest Students; CAPS = Clinician Administered PTSD Scale; PD = PTSD diagnosis; BDI-SF = BDI-Short Form; BDI = Beck Depression Inventory; CESD-R-10 = Center for Epidemiologic Studies Depression Scale Revised; NS = Northeastern Students; PSS = Perceived Stress Scale; LEC = number of event types endorsed as “happened to you personally” on Life Events Checklist; HMS = total number of high magnitude stressor events reported on the Trauma History Screen (THS); AUDIT = Alcohol Use Disorders Identification Test.

Weighted (by sample size) average correlation across measures of the same construct. ^b Zou’s (2007) test for comparing dependent overlapping correlations.

Holm-adjusted p-values: *p < .05. **p < .01. ***p < .001.

Table 5.

Correlations of DSS-B Scores With Other Measures in the COTI DSS-B Sample.

Other measures	n	r	(95% CI)
PCL-5 total score	478	.689***	[.639, .733]
PHQ-8 total score	478	.635***	[.578, .686]
Drinking days per month	470	.027	[−.064, .117]
Drinks per drinking day	470	.006	[−.085, .096]

Note. DSS-B = Dissociative Symptoms Scale-Brief version; COTI = consequences of traumatic injury; CI = confidence interval; PCL-5 = PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013); PTSD = posttraumatic stress disorder; PHQ-8 = Patient Health Questionnaire 8-item version.

Holm-adjusted p-values: *p < .05. **p < .01. ***p < .001.

Measurement Invariance

Measurement invariance was examined across four ethnoracial groups, including White/European American (n = 933), Black/African American (n = 625), Hispanic/Latinx American (n = 794), and Asian American (n = 786). These data were drawn from the total combined data set of the nine samples that completed the full DSS to maximize group sizes. Individuals in the combined data set were excluded from the analysis if they were missing race/ethnicity data, identified a group other than the groups listed above, or reported multiple identities (n = 263; 8%). Due to low endorsement of higher response options (0% to 17%), DSS-B items were dichotomized as 0 = “not at all” and 1 = all other responses. Measurement invariance analyses were conducted using the chi-square DIFF TEST procedure in Mplus.

The baseline model of configural invariance fit the data well, χ² (56) = 60.96, p = .302; RMSEA = 0.011, 95% CI = [0.000, 0.025]; CFI = 0.999; SRMR = 0.025. Constraining factor loadings to be equal (while allowing group-specific factor variances) did not significantly worsen model fit, χ²_diff (12) = 14.70, p = .258, which supported metric invariance across all four groups. However, when item thresholds were also constrained to be equal across groups (while allowing group-specific factor means), model fit significantly worsened, χ²_diff (12) = 43.04, p < .001. An investigation into the source of scalar non-invariance revealed two threshold equality constraints that were problematic. Freely estimating the threshold for DSS item 17 (DSS-B item 7) in the Asian American group and DSS item 18 (DSS-B item 8) in the White/European American group (while retaining all other threshold equality constraints) produced a model that did not fit significantly worse than the metric invariance model, χ²_diff (10) = 16.45, p = .088, supporting partial scalar invariance for White/European American and Asian American individuals.

Given that none of the equality constraints between the Black/African American and Hispanic/Latinx American groups were found to be problematic, we then tested strict factorial invariance between these groups specifically. A baseline model of scalar invariance for the two groups fit the data well, χ² (36) = 41.53, p = .242; RMSEA = 0.015, 95% CI = [0.000, 0.032]; CFI = 0.999; SRMR = 0.032, supporting the presence of full scalar invariance. Model fit did not worsen if item residual variances were constrained equal across groups rather than freely estimated, χ²_diff (8) = 9.17, p = .328, supporting the presence of strict factorial (measurement) invariance between the Black/African American and Hispanic/Latinx American groups.

We also conducted analyses to assess invariance of results (e.g., factor structure, internal consistency, convergent and discriminant validity) across administration methods (e.g., in person vs. phone). A description of these analyses and results are in the Supplemental Materials.

Discussion

The purpose of this study was to develop a brief measure of multiple domains of clinically relevant dissociation symptoms that provides reliable and valid scores in clinical and community samples. The DSS-B items were selected by identifying the two items from each DSS subscale that were most precise in measuring above average levels of the latent trait of dissociation among individuals from a variety of clinical and non-clinical samples. The finding that items selected for the DSS-B also had the highest peak information values within each subscale, and these peak information values were between 1.30 and 2.38 SD above the mean of the latent trait, provided further evidence that these items were most precise in measuring above average severities of dissociation within the samples studied. In addition, the selected items had at least an adequate level of precision in measuring a large range of clinically relevant levels of dissociation, as indicated by item information being greater than 1.0 for ranges of severity on the latent trait that spanned 2.43 SDs or more, which included minimums near the mean (-0.84 to 0.83) and maximums that were 2 to 3 SDs above the mean (2.32 to 3.29) on the latent trait.

Psychometric analyses conducted using separate samples of respondents, including an independent sample in which only DSS-B items were administered, revealed that the factor structure and content domains (subscales) of the DSS-B were adequately preserved and invariant across modes of administration (e.g., in person vs. phone), and that DSS-B scores reliably and validly assessed clinically significant dissociation in the clinical and community samples studied. The CFA fit results indicated that the four-factor structure of the DSS (Carlson et al., 2018) adequately represented the relationships among DSS-B items, especially when only DSS-B items were administered rather than the full DSS. The CFA model estimates for standardized loadings and inter-factor correlations suggest that DSS-B items measure four highly related content domains. It should be noted that although we aimed to create a brief version of the DSS that preserved coverage of all four content domains of the full scale, the DSS-B is only intended to provide a total score for dissociation. When the goal is to assess the individual content domains (subscales), the full DSS should be used instead.

DSS-B scores were shown to have strong internal consistency, with higher estimates observed in the sample in which only DSS-B items were administered and when internal consistency was examined using categorical omega rather than coefficient alpha. Although coefficient alpha is the more commonly reported estimate of scale reliability, it assumes that scale items are continuously distributed, which may not be appropriate when items are measured with a limited number of ordered response categories (Green & Yang, 2009). Thus, categorical omega may provide a more accurate estimate of composite reliability for DSS-B scores than alpha. However, we also report alpha to facilitate comparison with past research. DSS-B scores were also found to be fairly stable over a 1-week period, as indicated by test–retest correlations of .57 and .66, which were consistent with expectations. The lower stability of DSS-B scores in the Trauma Center sample may be explained by the recency of trauma exposure, as more rapid changes in symptoms are likely to be experienced in the early aftermath of trauma. Furthermore, internal consistency and temporal stability of DSS-B scores were only slightly lower than the estimates for the full DSS (Carlson et al., 2018), suggesting good preservation of scale reliability.

The validity of DSS-B scores was supported by the findings that DSS-B scores were higher for participants in clinical samples than in non-clinical samples, higher in participants with probable PTSD than those without probable PTSD, and higher in participants injured by interpersonal or intentional violence as opposed to a vehicle collision, fall, or other accident. The validity of DSS-B scores was further demonstrated by correlations with other measures that were generally consistent with expectations, comparable to the correlations for the full DSS, and invariant across mode of administration. These results are consistent with the results of other studies of the relationships between dissociation, PTSD, and trauma exposure (Carlson et al., 2012) and between dissociation and alcohol use (Chana et al., 2021). The correlations with measures of depression were somewhat higher than expected; however, this finding may have been related to the samples in which depression was measured. For example, the correlations of DSS-B scores with measures of PTSD and depression in the Caregivers sample are consistent with results of previous research on women exposed to interpersonal trauma (Taft et al., 2009). In the Trauma Center and COTI DSS-B samples, the higher than expected correlations between dissociation and depression symptoms may have been due to the recency of trauma exposure, as there is considerable variability in symptoms and mood in the early aftermath of trauma (Carlson et al., 2016), which can produce stronger associations among variables. The fact that correlations with scores on measures of the same and similar constructs were slightly lower for DSS-B scores than for the full measure was expected given the impact that removing items has on reducing reliability and content coverage of a scale (Smith et al., 2000). Despite variation in the strength of correlations across measures of the same construct, all weighted average correlations were weaker than the correlation with another measure of dissociation, and the strongest weighted average correlation was for the most similar construct (PTSD), consistent with expectations.

Measurement invariance analyses provided strong support for DSS-B items measuring the same construct across Black/African American and Hispanic/Latinx American individuals. This suggests that differences in DSS-B scores across these groups could be meaningfully interpreted as true differences in dissociation as opposed to differences in how items were interpreted. In addition, results of the measurement invariance analyses supported the presence of partial scalar invariance across all four ethnoracial groups. Only two items were identified as scalar non-invariant: one item from the Gaps in Awareness and Memory scale, which was likely explained by White/European American individuals endorsing the symptom more frequently than other groups, and one item from the Cognitive-Behavioral Reexperiencing scale that was non-invariant for Asian American individuals but was not explained by group differences in endorsement. It is relevant to note that there were very few Asian American participants in clinical samples (n = 34; 4%), which may have contributed to difficulties establishing full measurement invariance. Moreover, the present findings for measurement invariance are based only on data collected from administering the full DSS, not the DSS-B, which may have impacted the results. Future studies should further examine the measurement invariance of the DSS-B across racial/ethnic groups, particularly in samples in which the DSS-B was administered. Nonetheless, these initial results are promising in that they suggest that DSS-B scores may largely be interpreted as measuring the same construct across ethnoracial groups in the U.S.

There are many benefits to developing a brief measure that assesses multiple domains of clinically relevant dissociation symptoms. Brief measures are less time-consuming and can be administered in a variety of settings to efficiently assess and monitor dissociation with a reduced burden on patients, providers, and researchers. In addition, the development of the DSS-B is timely given the addition of the dissociative subtype to the PTSD diagnostic criteria (APA, 2013), which has increased interest in assessing dissociation among individuals with PTSD. One potential use of the DSS-B is to screen for dissociation, which may be important in clinical contexts given that individuals with PTSD who also experience dissociation may require specialized clinical treatment (Cloitre et al., 2012). In addition, use of the DSS-B may be desirable in medical settings where assessments need to be conducted quickly and efficiently. A brief measure is also beneficial for assessing symptoms over time. The DSS-B may facilitate collection of longitudinal or intensive data on dissociation symptoms, which could be important given that such symptoms have been found to vary across time (Carlson et al., 2016) and predict improvements in PTSD during treatment (Kleindienst et al., 2016). Finally, development of a brief measure of dissociation is important because dissociation is often measured in conjunction with other symptoms, and using brief measures can enable assessment of a greater number of constructs or help to improve study response rates, representativeness of samples, and generalizability of results while minimizing burden on participants. Thus, the availability of the DSS-B may facilitate the measurement of dissociation, which could increase knowledge about multiple domains of dissociation in the context of PTSD and other disorders.

Strengths of this study include the large number of individuals studied and the diversity of samples. For example, nearly three quarters of participants did not identify as White/European American. Several limitations of this research should be noted. Almost half of the data used in this study were used to develop the full DSS (Carlson et al., 2018). However, the present study used a different split of these data for the analyses and three new samples that had not been used to develop the full DSS (Caregivers, Northeast Students, and COTI DSS-B), including an independent sample in which only the DSS-B was administered. It should be noted that the DSS-B was only administered to participants in the COTI DSS-B sample. DSS-B scores for participants in all other samples were computed based on responses to the full DSS, and in these samples, it is possible that DSS items not included in the DSS-B affected responses to DSS-B items. Thus, results for the COTI DSS-B sample may be a more accurate representation of the structure, reliability, and validity of DSS-B scores than the results from the other samples. Moreover, because the measurement invariance analyses were not conducted with data from the COTI DSS-B sample, these findings should be interpreted with caution and verified by future studies in which the DSS-B is administered rather than the full measure. Another limitation was that some of the samples were convenience or clinical samples and information about diagnostic status other than PTSD was not available. These characteristics of the samples may have impacted the findings and generalizability of results, and future studies should examine the psychometric properties of the DSS-B in other diagnostic groups that experience dissociation.

In sum, this study provides substantial support that DSS-B scores measure clinically relevant levels of dissociation and have strong psychometric properties in the populations studied. Analyses revealed that the reduction in items had minimal impact on coverage of the content domains and factor structure. The reliability and validity of DSS-B scores were comparable to scores on the full measure, and DSS-B items had a high level of measurement invariance across ethnoracial groups. Thus, the DSS-B could be an important tool for quickly and efficiently assessing a range of clinically relevant dissociative symptoms in a variety of contexts.

Supplemental Material

sj-pdf-1-asm-10.1177_10731911221133317 – Supplemental material for Development of a Brief Version of the Dissociative Symptoms Scale and the Reliability and Validity of DSS-B Scores in Diverse Clinical and Community Samples

Supplemental material, sj-pdf-1-asm-10.1177_10731911221133317 for Development of a Brief Version of the Dissociative Symptoms Scale and the Reliability and Validity of DSS-B Scores in Diverse Clinical and Community Samples by Kathryn S. Macia, Eve B. Carlson, Patrick A. Palmieri, Steven R. Smith, Deidre M. Anglin, Chandra Ghosh Ippen, Alicia F. Lieberman, Eunice C. Wong, Terry L. Schell and Lynn C. Waelde in Assessment

Footnotes

Appendix

Authors’ Note

Preparation of this article was supported with resources of the Department of Veterans Affairs’ Office of Academic Affiliations Advanced Fellowship Program in Mental Illness Research and Treatment, the Department of Veterans Affairs’ National Center for PTSD, and the Veterans Affairs Palo Alto Health Care System. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or the United States Government. This article is not subject to US copyright law and is in the public domain.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: A portion of this research was supported by grant funding from the National Institute on Minority Health and Health Disparities (NIMHD) under Award Number: R01 MD010372.

ORCID iDs

Kathryn S. Macia

Deidre M. Anglin

Supplemental Material

Supplemental material for this article is available online.

References

Allen

J. P.

Litten

R. Z.

Fertig

J. B.

Babor

(1997). A review of research on the Alcohol Use Disorders Identification Test (AUDIT). Alcoholism: Clinical and Experimental Research, 21, 613–619. https://doi.org/10.1111/j.1530-0277.1997.tb03811.x

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.).

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing.

Anglin

D. M.

Lighty

Greenspoon

Ellman

L. M.

(2014). Racial discrimination is associated with distressing subthreshold positive psychotic symptoms among U.S. urban ethnic minority young adults. Social Psychiatry and Psychiatric Epidemiology, 49, 1545–1555. https://doi.org/10.1007/s00127-014-0870-8

Armour

Contractor

A. A.

Palmieri

P. A.

Elhai

J. D.

(2014). Assessing latent level associations between PTSD and dissociative factors: Is depersonalization and derealization related to PTSD factors more so than alternative dissociative factors? Psychological Injury and Law, 7, 131–142. https://doi.org/10.1007/s12207-014-9196-9

Babor

T. F.

(2000). Alcohol Use Identification Test (AUDIT). In John Rush

(Ed.) Handbook of psychiatric measures (pp. 460–462). American Psychiatric Association.

Baker

F. B.

(2001). The basics of item response theory (2nd ed.). http://files.eric.ed.gov/fulltext/ED458219.pdf

Beck

A. T.

Steer

R. A.

(1993). Beck Depression Inventory manual. Psychological Corporation.

Beck

A. T.

Steer

R. A.

Garbin

M. G.

(1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100. https://doi.org/10.1016/0272-7358(88)90050-5

10.

Beck

A. T.

Ward

C. H.

Mendelson

Mock

Erbaugh

(1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561–571. https://doi.org/10.1001/archpsyc.1961.01710120031004

11.

Bernstein

E. M.

Putnam

F. W.

(1986). Development, reliability, and validity of a dissociation scale. The Journal of Nervous and Mental Disease, 174, 727–735. https://doi.org/10.1097/00005053-198612000-00004

12.

Björgvinsson

Kertz

S. J.

Bigda-Peyton

J. S.

McCoy

K. L.

Aderka

I. M.

(2013). Psychometric properties of the CES-D-10 in a psychiatric sample. Assessment, 20, 429–436. https://doi.org/10.1177/1073191113481998

13.

Blevins

C. A.

Weathers

F. W.

Davis

M. T.

Witte

T. K.

Domino

J. L.

(2015). The Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5): Development and initial psychometric evaluation. Journal of Traumatic Stress, 28, 489–498. https://doi.org/10.1002/jts.22059

14.

Bovin

M. J.

Marx

B. P.

Weathers

F. W.

Gallagher

M. W.

Rodriguez

Schnurr

P. P.

Keane

T. M.

(2016). Psychometric properties of the PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition (PCL-5) in veterans. Psychological Assessment, 28, 1379–1391. https://doi.org/10.1037/pas0000254

15.

Cardeña

Carlson

E. B.

(2011). Acute stress disorder revisited. Annual Review of Clinical Psychology, 7, 245–267. https://doi.org/10.1146/annurev-clinpsy-032210-104502

16.

Carlson

E. B.

(2001). Psychometric study of a brief screen for PTSD: Assessing the impact of multiple traumatic events. Assessment, 8, 431–441. https://doi.org/10.1177/107319110100800408

17.

Carlson

E. B.

Dalenberg

C. J.

McDade-Montez

(2012). Dissociation in posttraumatic stress disorder part 1: Definitions and review of research. Psychological Trauma: Theory, Research, Practice, and Policy, 4, 479–489. https://doi.org/10.1037/a0027748

18.

Carlson

E. B.

Field

N. P.

Ruzek

J. I.

Bryant

R. A.

Dalenberg

C. J.

Keane

T. M.

Spain

D. A.

(2016). Advantages and psychometric validation of proximal intensive assessments of patient reported outcomes collected in daily life. Quality of Life Research, 25, 507–516. https://doi.org/10.1007/s11136-015-1170-9

19.

Carlson

E. B.

Palmieri

P. A.

Spain

D. A.

(2017). Development and preliminary performance of a risk factor screen to predict posttraumatic psychological disorder after trauma exposure. General Hospital Psychiatry, 46, 25–31. https://doi.org/10.1016/j.genhosppsych.2016.12.011

20.

Carlson

E. B.

Putnam

F. W.

(1993). An update on the Dissociative Experiences Scale. Dissociation: Progress in the Dissociative Disorders, 6, 16–27.

21.

Carlson

E. B.

Smith

S. R.

Palmieri

P. A.

Dalenberg

Ruzek

J. I.

Kimerling

Burling

T. A.

Spain

D. A.

(2011). Development and validation of a brief self-report measure of trauma exposure: The Trauma History Screen. Psychological Assessment, 23, 463–477. https://doi.org/10.1037/a0022294

22.

Carlson

E. B.

Waelde

L. C.

Palmieri

P. A.

Macia

K. S.

Smith

S. R.

McDade-Montez

(2018). Development and validation of the Dissociative Symptoms Scale. Assessment, 25, 84–99. https://doi.org/10.1177/1073191116645904

23.

Caspi

Carlson

E. B.

Klein

(2007). Validation of a screening instrument for posttraumatic stress disorder in a community sample of Bedouin men serving in the Israeli Defense Forces. Journal of Traumatic Stress, 20, 517–527. https://doi.org/10.1002/jts.20237

24.

Centanni

S. W.

Bedse

Patel

Winder

D. G.

(2019). Driving the downward spiral: Alcohol-induced dysregulation of extended amygdala circuits and negative affect. Alcoholism: Clinical and Experimental Research, 43, 2000–2013. https://doi.org/10.1111/acer.14178

25.

Center for Behavioral Health Statistics and Quality. (2016). Key substance use and mental health indicators in the United States: Results from the 2015 National Survey on Drug Use and Health (HHS Publication No. SMA 16-4984, NSDUH Series H-51). http://www.samhsa.gov/data/

26.

Chalmers

R. P.

(2012). mirt: A multidimensional item response theory package for the R environment. Journal of Statistical Software, 48, 1–29. https://doi.org/10.18637/jss.v048.i06

27.

Chana

S. M.

Wolford-Clevenger

Faust

Hemberg

Ramaswamy

Cropsey

(2021). Associations among betrayal trauma, dissociative posttraumatic stress symptoms, and substance use among women involved in the criminal legal system in three U.S. cities. Drug and Alcohol Dependence, 227, 108924. https://doi.org/10.1016/j.drugalcdep.2021.108924

28.

Chen

W.-H.

Thissen

(1997). Local dependence indexes for item pairs using item response theory. Journal of Educational and Behavioral Statistics, 22, 265–289. https://doi.org/10.3102/10769986022003265

29.

Chiu

C.-D.

Chang

J.-H.

Hui

C. M.

(2017). Self-concept integration and differentiation in subclinical individuals with dissociation proneness. Self and Identity, 16, 664–683. https://doi.org/10.1080/15298868.2017.1296491

30.

Cloitre

Petkova

Wang

(2012). An examination of the influence of a sequential treatment on the course and impact of dissociation among women with PTSD related childhood abuse. Depression and Anxiety, 29, 709–717. https://doi.org/10.1002/da.21920

31.

Cohen

Kamarck

Mermelstein

(1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. https://doi.org/10.2307/2136404

32.

Cohen

Williamson

G. M.

(1988). Perceived stress in a probability sample of the United States. In Spacapan

Oskamp

(Eds.), The social psychology of health (pp. 31–66). SAGE.

33.

Davidson

J. R. T.

Book

S. W.

Colket

J. T.

Tupler

L. A.

Roth

David

Hertzberg

Mellman

Beckham

J. C.

Smith

Davison

R. M.

Katz

Feldman

(1997). Assessment of a new self-rating scale for post-traumatic stress disorder. Psychological Medicine, 27, 153–160. https://doi.org/10.1017/S0033291796004229

34.

Diedenhofen

Musch

(2015). cocor: A comprehensive solution for the statistical comparison of correlations. PLOS ONE, 10, Article e0121945. https://doi.org/10.1371/journal.pone.0121945

35.

Drasgow

Levine

M. V.

Williams

E. A.

(1985). Appropriateness measurement with polychotomous item response models and standardized indices. British Journal of Mathematical and Statistical Psychology, 38, 67–86. https://doi.org/10.1111/j.2044-8317.1985.tb00817.x

36.

Dutra

Bureau

Holmes

Lyubchik

Lyons-Ruth

(2009). Quality of early care and childhood trauma: A prospective study of developmental pathways to dissociation. The Journal of Nervous and Mental Disease, 197, 383–390. https://doi.org/10.1097/NMD.0b013e3181a653b7

37.

Eidhof

M. B.

ter Heide

F. J. J.

van Der Aa

Schreckenbach

Schmidt

Brand

B. L.

Lanius

R. A.

Loewenstein

R. J.

Spiegel

Vermetten

(2019). The Dissociative Subtype of PTSD Interview (DSP-I): Development and psychometric properties. Journal of Trauma & Dissociation, 20, 564–581. https://doi.org/10.1080/15299732.2019.1597806

38.

Frischholz

E. J.

Braun

B. G.

Sachs

R. G.

Hopkins

Schaeffer

D. M.

Lewis

Leavitt

Pasquotto

M. A.

Schwartz

D. R.

(1990). The Dissociative Experiences Scale: Further replication and validation. Dissociation: Progress in the Dissociative Disorders, 3, 151–153.

39.

Gray

M. J.

Litz

B. T.

Hsu

J. L.

Lombardo

T. W.

(2004). Psychometric properties of the Life Events Checklist. Assessment, 11, 330–341. https://doi.org/10.1177/1073191104269954

40.

Green

S. B.

Yang

(2009). Reliability of summed item scores using structural equation modeling: An alternative to coefficient alpha. Psychometrika, 74, 155–167. https://doi.org/10.1007/s11336-008-9099-3

41.

Hagan

M. J.

Browne

D. T.

Sulik

Ghosh Ippen

Bush

Lieberman

A. F.

(2017). Parent and child trauma symptoms during child-parent psychotherapy: A prospective cohort study of dyadic change. Journal of Traumatic Stress, 30, 690–697. https://doi.org/10.1002/jts.22240

42.

Holm

(1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6, 65–70. https://www-jstor-org.web.bisu.edu.cn/stable/4615733

43.

Hyland

Shevlin

Fyvie

Cloitre

Karatzias

(2020). The relationship between ICD-11 PTSD, complex PTSD and dissociative experiences. Journal of Trauma & Dissociation, 21, 62–72. https://doi.org/10.1080/15299732.2019.1675113

44.

Hyland

Shevlin

Fyvie

Karatzias

(2018). Posttraumatic stress disorder and complex posttraumatic stress disorder in DSM-5 and ICD-11: Clinical and behavioral correlates. Journal of Traumatic Stress, 31, 174–180. https://doi.org/10.1002/jts.22272

45.

Kelley

Pornprasertmanit

(2016). Confidence intervals for population reliability coefficients: Evaluation of methods, recommendations, and software for composite measures. Psychological Methods, 21, 69–92. https://doi.org/10.1037/a0040086

46.

Kleindienst

Priebe

Görg

Dyer

Steil

Lyssenko

Winter

Schmahl

Bohus

(2016). State dissociation moderates response to dialectical behavior therapy for posttraumatic stress disorder in women with and without borderline personality disorder. European Journal of Psychotraumatology, 7, 30375. https://doi.org/10.3402/ejpt.v7.30375

47.

Kohout

F. J.

Berkman

L. F.

Evans

D. A.

Cornoni-Huntley

(1993). Two shorter forms of the CES-D depression symptoms index. Journal of Aging and Health, 5, 179–193. https://doi.org/10.1177/089826439300500202

48.

Kroenke

Strine

T. W.

Spitzer

R. L.

Williams

J. B. W.

Berry

J. T.

Mokdad

A. H.

(2009). The PHQ-8 as a measure of current depression in the general population. Journal of Affective Disorders, 114, 163–173. https://doi.org/10.1016/j.jad.2008.06.026

49.

Lanius

R. A.

Brand

Vermetten

Frewen

P. A.

Spiegel

(2012). The dissociative subtype of posttraumatic stress disorder: Rationale, clinical and neurobiological evidence, and implications. Depression and Anxiety, 29, 701–708. https://doi.org/10.1002/da.21889

50.

Lee

E.-H.

(2012). Review of the psychometric evidence of the Perceived Stress Scale. Asian Nursing Research, 6, 121–127. https://doi.org/10.1016/j.anr.2012.08.004

51.

Lyssenko

Schmahl

Bockhacker

Vonderlin

Bohus

Kleindienst

(2018). Dissociation in psychiatric disorders: A meta-analysis of studies using the Dissociative Experiences Scale. The American Journal of Psychiatry, 175, 37–46. https://doi.org/10.1176/appi.ajp.2017.17010025

52.

Maniglio

(2009). The impact of child sexual abuse on health: A systematic review of reviews. Clinical Psychology Review, 29, 647–657. https://doi.org/10.1016/j.cpr.2009.08.003

53.

McDonald

S. D.

Calhoun

P. S.

(2010). The diagnostic accuracy of the PTSD Checklist: A critical review. Clinical Psychology Review, 30, 976–987. https://doi.org/10.1016/j.cpr.2010.06.012

54.

McDonald

S. D.

Thompson

N. L.

Stratton

K. J.

Calhoun

P. S.

, & VA Mid-Atlantic MIRECC Workgroup. (2014). Diagnostic accuracy of three scoring methods for the Davidson Trauma Scale among U.S. military veterans. Journal of Anxiety Disorders, 28, 160–168. https://doi.org/10.1016/j.janxdis.2013.09.004

55.

Mohebbi

Nguyen

McNeil

J. J.

Woods

R. L.

Nelson

M. R.

Shah

R. C.

Storey

Murray

A. M.

Reid

C. M.

Kirpach

Wolfe

Lockery

J. E.

Berk

(2018). Psychometric properties of a short form of the Center for Epidemiologic Studies Depression (CES-D-10) scale for screening depressive symptoms in healthy community dwelling older adults. General Hospital Psychiatry, 51, 118–125. https://doi.org/10.1016/j.genhosppsych.2017.08.002

56.

Műllerová

Hansen

Contractor

A. A.

Elhai

J. D.

Armour

(2016). Dissociative features in posttraumatic stress disorder: A latent profile analysis. Psychological Trauma: Theory, Research, Practice, and Policy, 8, 601–608. https://doi.org/10.1037/tra0000148

57.

Muthén

L. K.

Muthén

B. O.

(2019). Mplus (Version 8.4) [Computer software].

58.

Perneger

T. V.

(1998). What’s wrong with Bonferroni adjustments. British Medical Journal, 316, 1236–1238. https://doi.org/10.1136/bmj.316.7139.1236

59.

Raykov

Marcoulides

G. A.

(2015). A direct latent variable modeling based method for point and interval estimation of coefficient alpha. Educational and Psychological Measurement, 75, 146–156. https://doi.org/10.1177/0013164414526039

60.

Roberts

N. P.

Kitchiner

N. J.

Lewis

C. E.

Downes

A. J.

Bisson

J. I.

(2021). Psychometric properties of the PTSD Checklist for DSM-5 in a sample of trauma-exposed mental health service users. European Journal of Psychotraumatology, 12, 1863578. https://doi.org/10.1080/20008198.2020.1863578

61.

Scalabrini

Cavicchioli

Fossati

Maffei

(2017). The extent of dissociation in borderline personality disorder: A meta-analytic review. Journal of Trauma & Dissociation, 18, 522–543. https://doi.org/10.1080/15299732.2016.1240738

62.

Shin

Lee

S.-H.

Han

K.-M.

Yoon

H.-K.

Han

(2019). Comparison of the usefulness of the PHQ-8 and PHQ-9 for screening for major depressive disorder: Analysis of psychiatric outpatient data. Psychiatry Investigation, 16, 300–305. https://doi.org/10.30773/pi.2019.02.01

63.

Smith

G. T.

McCarthy

D. M.

Anderson

K. G.

(2000). On the sins of short-form development. Psychological Assessment, 12, 102–111. https://doi.org/10.1037/1040-3590.12.1.102

64.

Taft

C. T.

Resick

P. A.

Watkins

L. E.

Panuzio

(2009). An investigation of posttraumatic stress disorder and depressive symptomatology among female victims of interpersonal trauma. Journal of Family Violence, 24, 407–415. https://doi.org/10.1007/s10896-009-9243-6

65.

Waelde

L. C.

Silvern

Carlson

E. B.

Fairbank

J. A.

Kletter

(2009). Dissociation in PTSD. In Dell

P. F.

O’Neil

Somer

(Eds.), Dissociation and the dissociative disorders: DSM-V and beyond (pp. 447–456). Routledge.

66.

Weathers

F. W.

Keane

T. M.

Davidson

J. R. T.

(2001). Clinician-Administered PTSD Scale: A review of the first ten years of research. Depression and Anxiety, 13, 132–156. https://doi.org/10.1002/da.1029

67.

Weathers

F. W.

Litz

Herman

Huska

Keane

(1993, November). The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility [Paper presentation]. Annual Convention of the International Society of Traumatic Stress Studies, San Antonio, TX, United States.

68.

Weathers

F. W.

Litz

B. T.

Keane

T. M.

Palmieri

P. A.

Marx

B. P.

Schnurr

P. P.

(2013). The PTSD Checklist for DSM-5 (PCL-5). https://www.ptsd.va.gov/

69.

Weathers

F. W.

Ruscio

A. M.

Keane

T. M.

(1999). Psychometric properties of nine scoring rules for the Clinician-Administered Posttraumatic Stress Disorder Scale. Psychological Assessment, 11, 124–133. https://doi.org/10.1037/1040-3590.11.2.124

70.

Wilkins

K. C.

Lang

A. J.

Norman

S. B.

(2011). Synthesis of the psychometric properties of the PTSD Checklist (PCL) Military, Civilian, and Specific versions. Depression and Anxiety, 28, 596–606. https://doi.org/10.1002/da.20837

71.

Wright

S. P.

(1992). Adjusted p-values for simultaneous inference. Biometrics, 48, 1005–1013. https://doi.org/10.2307/2532694

72.

Zou

G. Y.

(2007). Toward using confidence intervals to compare correlations. Psychological Methods, 12, 399–413. https://doi.org/10.1037/1082-989X.12.4.399