Psychometric Properties of the Defense Mechanisms Rating Scales-Self-Report-30 (DMRS-SR-30): Internal Consistency, Validity and Factor Structure

Abstract

Assessment of defense mechanisms has a longstanding history within the clinical psychology and psychopathology literature. Despite their centrality to clinical practice, there are few self-report measures that assess defenses and, those that do exist, have limitations in addressing individual defenses and levels of defensive functioning. To address this need, we investigated the psychometric properties of the Defense Mechanisms Rating Scale - Self-Report − 30 item (DMRS-SR-30) with a global, community sample of 1,539 participants who responded to an online survey about distress and coping. Exploratory factor analysis found a three-factor model for the DMRS-SR-30 – mature, mental inhibition and avoidance, and immature-depressive. Internal consistency was high for the Overall Defensive Functioning (ODF) and the three extracted factors with coefficient alphas ranging from .75 to .90. Examination of concurrent validity with a commonly used measure of defensive functioning found significant relationships in the predicted directions. The group of immature defenses had the strongest concurrent validity (r = .50). Finally, correlations with external criteria – including psychological distress and adverse childhood experiences – supported the convergent and discriminant validity of the DMRS-SR-30. The three factor structure of the DMRS-SR-30 has good psychometric properties. Limitations and directions for future research, as well as clinical implications, are described.

Introduction

More than a century of research demonstrates that defense mechanisms are automatic psychological processes that aid individuals in dealing with internal conflicts and stressful situations (American Psychiatric Association, 1994; Freud, 1984). Cramer considers them cognitive processes that function to protect the individual from experiencing excessive anxiety or other negative emotions, loss of self-esteem, and loss of self-integration (Cramer, 2008). Defenses are a central part of personality structure (Kernberg, 1967). They operate in a way that is largely unconscious, automatic, protective, reversible, discriminable, and either adaptive or maladaptive, depending on the situation in which they are activated (Cramer, 1991; Cramer & Davidson, 1998; Vaillant, 1971). Defense mechanisms can be empirically assessed with different approaches, such as laboratory procedures, projective tests, self-report, and observer-rated techniques, each of which show specific strengths and limitations (Guldberg et al., 1993; MacGregor & Olson, 2005; Perry & Lanni, 1998). These varied methods often result in discrepant findings due to differences in data perspective, conceptual bases and measurement characteristics. This variability understandably can generate confusion among scholars and results in difficulties with interpreting and comparing findings (Di Giuseppe et al., 2018). There is increasing awareness of the potential usefulness of valid self-report measures to assess defenses in empirical research and to inform clinical work (Davidson & MacGregor, 1998; Perry & Lanni, 1998; Ruggero et al., 2019). One commonly used measure, the Defense Styles Questionnaire (DSQ-40; Andrews et al., 1993), has an unstable factor structure (Prout et al., 2018). Additionally, the DSQ-40 has well-documented problems with face validity, with expert raters unable to correctly identify items linked to specific defenses (Chabrol et al., 2005; Trijsburg et al., 2000). Despite these limitations, the DSQ-40 remains the most widely used self-report measure of defense mechanisms.

In this study, we tested validity and internal consistency of a novel 30-item questionnaire based on the gold-standard, empirically derived hierarchical organization of defense mechanisms (American Psychiatric Association, 1994; Vaillant, 1992). This hierarchy was first operationalized with the observer-rated Defense Mechanisms Rating Scales (DMRS; Perry, 1990). The present study reports on the internal consistency, validity, factor structure, and overall psychometric properties of a self-report measure based on the DMRS, the Defense Mechanisms Rating Scale-Self-Report-30, (DMRS-SR-30; Di Giuseppe et al., 2020).

Developed by George Vaillant fifty years ago (Vaillant, 1971, 1977, 1992), the hierarchy of defense mechanisms was introduced in the Diagnostic and Statistical Manual, 4^th Edition (DSM-IV) as an axis known as the Defense Functioning Scale (DFS; American Psychiatric Association, 1994; Perry et al., 1998; Skodol & Perry, 1993). The DFS is based on a version of the DMRS (DeFife & Hilsenroth, 2005; Perry, 1990), which is now considered the gold-standard measure for the assessment of defense mechanisms (Bond & Perry, 2004; Conversano & Di Giuseppe, 2021). The DMRS is an observer-rated measure that provides definitions of 30 defense mechanisms, hierarchically organized into seven levels of adaptiveness which are further organized into three defensive categories. Ranked from immature to mature, these levels and categories are: Immature defensive category, including Action (Level 1), Major Image Distortion (Level 2), Disavowal (Level 3), Minor Image-Distortion (Level 4) defense levels; Neurotic defensive category, including Neurotic (Level 5), further divided into Hysterical (Level 5a) and Other Neurotic (Level 5 b) sublevels, and Obsessional (Level 6) defense levels; and Mature defensive category, corresponding to High Adaptive (Level 7) defenses. The DMRS Manual describes the definition and intrapsychic function, accompanied by examples which aid in the qualitative discrimination of a defense from near-neighbor defenses (Perry, 1990). Defense mechanisms are then identified in transcripts of clinical interviews or therapy sessions by trained raters who apply either a qualitative or quantitative scoring system (Perry & Henry, 2004). The DMRS quantitative scoring is more frequently used in research and provides scores for: (1) overall defensive functioning (ODF), an index of defensive maturity that can be used as an outcome measure (Siefert et al., 2006); (2) three hierarchically ordered defensive categories of mature, neurotic and immature defenses, largely corresponding to Vaillant’s tripartite organization of defense mechanisms; (3) seven defense levels, each of which is comprised of defenses that have overlapping functions; and (4) 30 individual defense mechanisms included in the hierarchy. Psychotic defenses were not originally included, although others have added them. Validity and reliability of the DMRS have been widely demonstrated (Perry et al., 2020; Perry & Cooper, 1989; Perry & Henry, 2004; Perry & Høglend, 1998). The DMRS led to the development the DMRS Q-sort version (DMRS-Q; Di Giuseppe et al., 2014).

The DMRS-Q is a computerized observer-rated Q-sort for the assessment of defense mechanisms based on the DMRS manual (Di Giuseppe et al., 2014). It provides the same three levels of quantitative scoring described for the DMRS as well as a Defensive Profile Narratives (DPN), which is a qualitative description of the patient’s defensive profile based on the most representative defensive patterns (Di Giuseppe & Perry, 2021). The DMRS-Q requires raters to rank-order 150 items into a 7-level forced distribution; it takes approximately 15 minutes for expert trained raters to complete. The rating procedure, which can be used with clinical interviews or therapy sessions, is available at https://webapp.dmrs-q.com/login. Preliminary validation studies have found good convergent validity and reliability of quantitative scores in both trained and untrained raters (Békés et al., 2021). Correlations between the DMRS and DMRS-Q range from acceptable to excellent (0.72 to 0.92) for both the ODF and the three overarching categories of defenses (Di Giuseppe et al., 2014). Inter-rater reliability is good for the ODF and defense levels (intraclass R values > 0.80), decreasing to acceptable for individual defenses (median ICC= 0.62). Building on the strength of the DMRS and the DMRS-Q, we developed a self-report measure of defense mechanisms, the DMRS-SR-30.

Development of the DMRS-SR-30

There is ample support indicating that DMRS-based measures are comprehensive and valid methods for assessing defense mechanisms in both research and clinical settings (Perry et al., 2009, 2013, 2019). However, there are evident limitations in applying either the DMRS or DMRS-Q to large samples due to the time required for data collection and coding procedures and costs of personnel involved. Despite their well-documented limitations (Davidson & MacGregor, 1998), self-report measures are sometimes preferred for use in large research protocols due to their ease of administration, low cost, and ability to anonymize respondent identities. While much of defensive activity occurs out of awareness, studying psychotherapy sessions reveals that patients are sometimes aware of the result of their defenses, so-called conscious derivatives of defenses. Self-report items that describe these phenomena are readily identified by individuals and serve as the basis for self-report measures of defenses. However, they do not reflect the part of defensive functioning that occurs fully out of awareness. One result is that observer ratings should capture some aspects of defensive functioning that self-report does not, which should yield some differences in what each method predicts.

There is an urgent need for assessing distress and psychological resources with large samples during the COVID-19 pandemic (Conversano et al., 2020; Prout et al., 2020). We developed the DMRS-SR-30, a 30-item self-report questionnaire based on the DMRS, to meet this need. An Italian version of the measure was validated with a large sample (Di Giuseppe et al., 2020).

The DMRS-SR-30 is a self-administered measure that includes the whole hierarchy of defense mechanisms as described in the DMRS manual (Perry, 1990). Items were developed in reference to the 150 DMRS-Q statements, from which we selected one item for each defense mechanism that best described the definition and function of 28 defenses (we combined self-and other image distorting defenses to yield one defense each for idealization and devaluation). One more item was selected for describing particular forms of passive aggression and dissociation, respectively. To test the accuracy of a priori selected items, we compared the DMRS-SR-30 set with 30 items randomly extracted from the remaining 120 DMRS-Q statements (one per defense, plus one more item for passive aggression and dissociation). Psychometric properties (internal consistency, concurrent validity, convergent and discriminant validity) were higher for the DMRS-SR-30 that for the randomly extracted set. Thus, we concluded that the theory-based selection was the most appropriate one. The DMRS-SR-30 scoring system is similar to other DMRS methods and provides quantitative scoring for ODF, defense levels, and individual defenses. As suggested by Hopwood and Bornstein (2014), who described the advantage to multi-method assessment of personality constructs, the development of complementary self-report system of the DMRS observer-rated method allows for examining defenses using a self-report and clinician-rated approach that is based on a similar conceptualization of the construct. In the present study we examine the psychometric properties of the English version of the DMRS-SR-30. Data were collected during the COVID-19 pandemic, a time when stress levels and psychological distress have been high (Prout et al., 2020) and therefore a prime opportunity to study this important construct. Specifically, we tested for: 1) factor structure, using exploratory factor analysis on both DMRS-SR-30 defensive categories and defense levels; 2) internal consistency of the DMRS-SR-30 subscales; 3) concurrent validity with the 40-item Defense Style Questionnaire (DSQ-40), another widely used self-report for defense mechanisms; and 4) convergent and discriminant validity, with outcome measures of anxiety, depression, and post-traumatic symptoms.

Methods

Participants

The sample in the current study was drawn from a larger research study examining the psychological impact of the COVID-19 pandemic (Prout et al., 2020). This online, cross sectional study was advertised via social media and email listservs with data collected between March 25, 2020 and April 22, 2020. The number of participants who provided informed consent was 3,192. Only those participants (n = 1,539) who completed the DSQ-40 – an optional study measure used to test concurrent validity of the DMRS-SR-30 – were included in the current study. Chi-square tests of independence indicated that women were more likely than men to complete the optional measure (χ²(2) = 10.79, p = .005). There were also differences between those who completed the DSQ-40 and those who did not for level of education (χ²(7) = 31.34, p = .00002); specifically, those with only a high school education were less likely than expected to opt in to the DSQ-40 portion of the study. Additionally, there were differences based on race/ethnicity (χ²(7) = 58.69, p < .001), with the proportion of Whites completing the DSQ-40 higher than expected and the proportion of Asian Americans lower than expected. There were no significant differences for marital or socioeconomic status.

The final set of participants was primarily White (n = 1,334; 86.7%), female (n = 1,266; 82.3%), had an average age of 21.63 (SD = 1.59), and lived in the United States (n = 1,072; 69.7%). About half of the participants were married or cohabitating (n = 831; 54.0%) and many had children (n = 800; 53.3%). There was a range of education levels, with approximately 26% having completed a four-year degree, 32% with less education, and 42% with advanced degrees. Socioeconomic status was also variable, with 43.4% (n = 668) reporting middle class status (Prout et al., 2020).

Measures

To test the concurrent validity of the DMRS-SR-30, another measure assessing defense mechanisms, the Defense Style Questionnaire-40 (DSQ-40; Andrews et al., 1993), was included. To assess convergent and discriminant validity DMRS-SR-30, we included measures of psychological symptoms and traumatic experiences during childhood – the Patient Health Questionnaire (PHQ; Spitzer et al., 1999), Impact of Event Scale – Revised (IES-R; Weiss & Marmar, 2004), and the Adverse Childhood Experiences questionnaire (ACE; Felitti et al., 1998).

Defense mechanisms rating scales-self report-30 (DMRS-SR-30; Di Giuseppe et al., 2020)

The DMRS-SR-30 is a 30 item 5-point scale assessing the whole hierarchy of defense mechanisms as described in the DSM-IV and empirically developed in the Defense Mechanisms Rating Scale (DMRS; Perry, 1990; Perry & Henry, 2004). DMRS-SR-30 items were developed from the Q-sort version of the DMRS (DMRS-Q; Di Giuseppe et al., 2014) and adapted for self-report. The DMRS-SR-30 provides several levels of scoring: an index of defensive maturity (ODF), a proportional score for each of the three defense categories and two-subcategories; a proportional score for each of the seven hierarchically ordered defense levels, and a proportional score for each of the 28 hierarchically ordered defense mechanisms. As the total sum of all 30 scores varies widely across respondents, we have found that total sum scores below 8 tend to be associated with outlier results, such as, having one defense responsible for 50% of a respondent's defensive functioning, or ODF = 1 or 7. Whenever the total sum scores are 8 or above, this tends not to occur. Preliminary validation of the Italian version of the questionnaire demonstrated good reliability for ODF and defensive categories with median ICC of .77, ranging from .68 to .89 (Di Giuseppe et al., 2020). Good criterion (ranging from .77 to .59) and concurrent (ranging from .63 to .27) validity also emerged from comparison of self-report and observer-rated versions of the DMRS (Di Giuseppe et al., 2020). SPSS syntax for scoring the DMRS-SR-30 is provided in supplement material Appendix B.

Defense style questionnaire-40 (DSQ-40; Andrews et al., 1993)

The DSQ-40 is a widely used questionnaire assessing defense mechanisms. The DSQ-40 contains 40 statements about individual action under stress, which reflect 20 different defense mechanisms. The DSQ-40 provides subscale scores for each of three styles of psychological adaptiveness: immature, neurotic, and mature. The DSQ-40’s psychometric properties have been demonstrated (Andrews et al., 1993), however there are questions about its face validity (Saint-Martin et al., 2013; Spinhoven et al., 1995) and varying factor structure across different samples (Prout et al., 2018).

Patient health questionnaire (PHQ; Spitzer et al., 1999)

The PHQ is a diagnostic tool for mental health disorders that assesses general psychological distress. It includes 26 self-administered items that screen for five of the most common groups of disorders in primary care: depressive, anxiety, alcohol, somatoform, and eating disorders. The PHQ and its modules for the various diagnostic categories have been used in thousands of studies and its reliability and validity are well-documented in the literature (Kroenke et al., 2010; Spitzer et al., 1999). In this study, scores from the Patient Health Questionnaire for Depression (PHQ-9; Kroenke et al., 2001) the Generalized Anxiety Disorder Scale (GAD-7; Spitzer et al., 2006), and the somatization subscale, the PHQ-15 Kroenke et al., 2002), were used. Participants completed the GAD-7 (65% of the total sample) only if they endorsed being bothered in the past four weeks by “feeling nervous, anxious, on edge, or worrying about a lot of different things.” Cronbach’s alphas for the PHQ-9, GAD-7, and PHQ-15 were .90, .80, and .76, respectively.

Impact of event scale – revised (IES-R; Weiss & Marmar, 2004)

The IES-R is a 22-item five-point scale self-report measure that assesses subjective distress caused by traumatic events. It is a revised version of the older 15-item IES (Horowitz et al., 1979). Items correspond directly to 14 of the 17 DSM-IV symptoms of post-traumatic stress disorder. Respondents in this sample were asked to respond to items with regard to the pandemic as a specific stressful life event, indicating how much they were distressed or bothered during the past seven days by each difficulty listed. The IES-R yields a total score (ranging from 0 to 88) and subscale scores can also be calculated for Intrusion, Avoidance, and Hyperarousal. Internal consistency for the total score in this sample was .93.

Adverse childhood experiences questionnaire (ACE; Felitti et al., 1998)

The ACE is a measure assessing the history of traumatic experiences based on a large epidemiological study of adults which demonstrated the long-term deleterious physical and mental health effects of child maltreatment (Dube et al., 2003; Felitti et al., 1998). The 10-item ACE questionnaire (Dube et al., 2003; Felitti et al., 1998) asks about an individual’s history of abuse, neglect, and household dysfunction prior to age 18. The test-retest reliability for emotional abuse, physical abuse, and sexual abuse is 0.66, 0.55, and 0.69, respectively (Dube et al., 2004). In the present study, internal consistency for the ACE was 0.77.

Procedures

This study was advertised via social media (e.g. Facebook, Twitter, Instagram) and email listservs worldwide. Adult respondents were invited to an online survey hosted on Qualtrics. After providing informed consent through the website, participants were then directed to complete the surveys online. The average completion time was 20 minutes. All the study procedures were conducted according to ethical standards outlined by the World Medical Association’s Declaration of Helsinki and were approved by the Institutional Review Board.

Statistical analyses

The factor structure of the DMRS-SR-30 was tested using SPSS principal axis factoring (PAF) extraction method with varimax rotation and Kaiser normalization. We used Velicer’s minimum average partial (MAP) test for determining the number of components. The test focuses on the common variance in a correlation matrix. Descriptive statistics were calculated for ODF, extracted factors, defensive categories, and defense levels. Cronbach’s alpha was used to test the internal consistency of the DMRS-SR-30 factors and subscales. Concurrent validity was analyzed using Pearson correlations between the DMRS-SR-30 and DSQ-40 on the three defensive categories of mature, neurotic and immature defenses. Finally, convergent and discriminant validity were examined by comparing both the DMRS-SR-30 and the DSQ-40 with psychological distress, posttraumatic symptoms, and child traumatic experiences.

Results

DMRS-SR-30 factor structure

There were three factors with eigen values over one. Table 1 presents the rotated factor loadings of three fixed factors. The loading range for all items in the triple-factor extraction was 0.32 − 0.69. Item 28 had a factor loading of < 0.3. According to the MAP test the optimal number of factors was three. In addition, the three-factor model largely confirmed the hierarchical organization of defense mechanisms based on the DMRS theory and therefore it was considered appropriate to retain three factors. The results of Velicer’s MAP Test are shown in Figure S1 in the online supplemental materials.

Table 1. Factor Structure of the DMRS-SR-30.

	Factor 1 Mature	Factor 2 Mental Inhibition and Avoidance	Factor 3 Immature-Depressive
Suppression	.52
Sublimation	.52
Self-observation	.62
Self-assertion	.68
Humor	.46
Anticipation	.48
Altruism	.61
Affiliation	.42
Isolation of Affects		.61
Intellectualization		.42
Undoing		.46
Repression		.57
Dissociation (item 3)		.56
Dissociation (item 24)		.69
Reaction Formation		.34
Displacement		.50
Devaluation			.54
Idealization			.41
Omnipotence	--	--	--
Denial		.67
Rationalization			.36
Projection			.46
Autistic Fantasy		.50
Projective Identification			.61
Splitting of other’s image			.36
Splitting of self-image			.32
Passive aggression (item 14)			.42
Passive aggression (item 18)			.54
HRC			.57
Acting out			.56

Note: Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization. Factor loadings < .3 are suppressed, total variance explained = 40.2%.

The first factor corresponds to the mature defensive category, with relatively high loadings ranging from .42 − .68. This maturity factor contains the eight defense mechanisms included in the high-adaptive defense level: affiliation, altruism, anticipation, humor, self-assertion, self-observation, sublimation, and suppression. The second extracted factor – described as mental inhibition and avoidance – contains obsessional and neurotic defense mechanisms with the addition of two disavowal defenses, denial and autistic fantasy. Factor loadings ranged from .34 to .69. From the most to the least adaptive, defense mechanisms included in the mental inhibition and avoidance factor were: isolation of affects, intellectualization, undoing, repression, dissociation, reaction formation, displacement, denial, and autistic fantasy. Finally, the third factor contains all depressive defenses, with the addition of two other immature defenses, rationalization and idealization. Factor loadings ranged from .32 to .62. From the most to the least adaptive, defense mechanisms included in the immature-depressive factor were: idealization, devaluation, rationalization, projection, splitting of self-image, splitting of object-image, projective identification, passive aggression, help-rejecting complaining, and acting out.

Descriptive statistics and internal consistency for the three-factor model, ODF, and defense levels

Table 2 displays mean scores of self-reported defensive functioning assessed with the DMRS-SR-30. The mean ODF fell in the high-neurotic range and the three defensive categories followed a normal distribution for non-clinical samples (Di Giuseppe et al., 2021; Perry et al., 2015; Prout et al., 2020). In descending order of magnitude, their prevalence was: mature defenses, mental inhibition and avoidance defenses, and immature-depressive defenses.

Table 2. Descriptive statistics and internal consistency for DMRS-SR-30 ODF, defensive categories and defense levels (N = 1,539).

	Mean (SD)	Range	Cronbach’s Alpha	Number of Items
ODF	5.43 (0.61)	3.53-7.00	.90	30
DMRS-SR-30 Factors
Mature	49.10 (16.14)	0-100	.75	8
Mental inhibition and avoidance	28.08 (9.97)	0-60	.84	10
Immature-depressive	20.49 (11.13)	0-56	.85	11
Factors Median			.81
Defensive Categories
Mature	49.10 (16.14)	0-100	.75	8
Neurotic	23.15 (7.93)	0-80	.81	8
Immature	27.75 (12.52)	0-60	.85	14
Depressive	16.10 (9.95)	0-56	.83	9
Non-depressive	11.65 (6.53)	0-60	.61	5
Categories Median			.77
Defense Levels
High-adaptive	49.10 (16.14)	0-100	.75	8
Obsessional	9.50 (5.04)	0-40	.58	3
Neurotic	13.66 (5.64)	0-38	.73	5
Minor image-distorting	6.93 (5.47)	0-60	.44	3
Disavowal	9.24 (5.81)	0-39	.68	4
Major image-distorting	6.92 (5.24)	0-50	.60	3
Action	4.65 (4.30)	0-26	.66	4
Levels Median			.63

Table 2 also shows internal consistency of the DMRS-SR-30 three-factors model and subscales (i.e., ODF, defense categories, and defense levels). Internal consistency was excellent for the total score (ODF) and ranged from good to very good for the three hierarchically ordered factors. Similarly, defensive categories also showed good internal consistency. As expected, lower Cronbach’s alpha values were found for the seven DMRS-SR-30 defense levels, which included fewer items in each subscale. Defense levels’ internal consistency was good and ranged from .58 to .75, with only minor image-distorting defenses showing very low reliability. The median value of internal consistency fell in the acceptable range, slightly above the median value reported in the Italian version of the questionnaire (Di Giuseppe et al., 2020).

Concurrent validity

Table 3 shows a comparison of the DMRS-SR-30 and DSQ-40 on tripartite defensive categories. All DMRS-SR-30 factor scores were in consistent and significant relationship with the corresponding DSQ-40 defense style scores, with the immature-depressive factor showing the highest correlation coefficient (r = .50, p = .000). Similar results were found comparing the DMRS-SR-30 defensive categories with the DSQ-40 defense styles. As expected, negative correlations were found between the maturity factor and immature defense style, indicating good concurrent validity between DMRS-SR-30 and the DSQ-40.

Table 3. Pearson Correlations between the DMRS-SR-30 and DSQ-40 (N = 1,549).

	DSQ
DMRS-SR-30 Scales	Mature	Neurotic	Immature
ODF	.29*	-.23*	-.58*
DMRS-SR-30 Factors
Mature	.25*	-.27*	-.60*
Mental Inhibition and Avoidance	-.13*	.24*	.40*
Immature-Depressive	-.27*	.20*	.50*
Defensive Categories
Mature Defenses	.25*	-.27*	-.60*
Neurotic Defenses	-.11*	.26*	.35*
Immature Defenses	-.25*	.20*	.56*
Depressive	-.31*	.16*	.47*
Non-Depressive	-.01	.14*	.35*

* p<.001.

Note: Depressive defense category includes acting out, help-rejecting complaining, passive aggression, splitting of self and other’s image, projective identification, projection, and devaluation. Non-depressive defense category includes denial, rationalization, autistic fantasy, omnipotence, and idealization.

Convergent and discriminant validity

Pearson correlations between the ODF and the seven defense levels assessed with the DMRS-SR-30 and measures assessing depression (PHQ-9), anxiety (GAD-7), posttraumatic symptoms (IES-R), somatization (PHQ Somatization), and childhood trauma (ACE) are presented in Table 4. There was good convergent validity between defense mechanisms and psychiatric symptoms among all immature and neurotic DMRS-SR-30 subscales, with the exception of minor image-distorting and other neurotic defenses which were unsurprisingly unrelated to self-reported symptoms. Conversely, ODF and high adaptive defenses were negatively correlated with psychological distress (ranging from r = −.27 to −.50). Good discriminant validity was demonstrated by small correlations between the DMRS-SR-30 and ACE scores across all defense levels (ranging from r = .006 to .17).

Table 4. Convergent and discriminant validity of DMRS-SR-30 ODF and Defense levels (N = 1,539).

	Depression (N = 1531)	Anxiety (N = 1229)	PTSS (N = 1539)	Somatization (N = 1515)	ACE (N = 1491)
ODF	-.51*	-.43*	-.44*	-.28*	-.15*
Mature Defenses	-.53*	-.46*	-.48*	-.30*	-.16*
Obsessional Defenses	.19*	.15*	.24*	.10*	.01
Neurotic Defenses	.42*	.41*	.39*	.34*	.19*
Hysterical	.51*	.50*	43*	.42*	.22*
Other Neurotic	.08*	.05	.12*	.05	.04
Minor I-D Defenses	-.07	.01	-.09*	-.12*	-.05
Disavowal Defenses	.44*	.35*	.43*	.27*	.17*
Major I-D Defenses	.38*	.27*	.27*	.19*	.13*
Action Defenses	.24*	.21*	.22*	.14*	.03

* p <.001.

Note: I-D = Image-distorting; PTSS = post-traumatic stress symptoms; ACE = adverse childhood experiences. Hysterical defense sublevel includes repression and dissociation. Other neurotic defense sublevel includes reaction formation and displacement.

Table 5 shows further analyses of convergent and discriminant validity tested among the three hierarchically ordered factors of the DMRS-SR-30 and the corresponding defense styles of the DSQ-40. Higher convergent validity of the DMRS-SR-30 was found for 10 of 12 correlations as compared to the DSQ-40. In particular, the DMRS-SR-30 mature factor resulted in larger negative correlations with psychiatric symptoms (ranging from r = −.29 to −.50) as compared to the DSQ-40 mature factor (ranging from r = −.07 to −.19). Small correlation coefficients were found between ACE scores and the DMRS-SR-30 and DSQ-40.

Table 5. Convergent and discriminant validity - Comparison of DMRS-SR-30 and DSQ-40 (N = 1,539).

	Depression	Anxiety	PTSS	Somatization	ACE
Mature Defenses
DMRS-SR-30 Factor 1	-.53*	-.46*	-.48*	-.30*	-.16*
DSQ	-.20*	-.15*	-.08	-.10*	.02
Neurotic Defenses
DMRS-SR-30 Factor 2	.48*	.42*	.49*	.33*	.16*
DSQ	.25*	21*	.30*	.21*	.11*
Immature Defenses
DMRS-SR-30 Factor 3	.41*	.33*	.33*	.21*	.11*
DSQ	.39*	.34*	.32*	.27*	.21*

* p <.001.

Note: DMRS-SR-30 Factor 1= Mature; DMRS-SR-30 Factor 2= Mental Inhibition and Avoidance; DMRS-SR-30 Factor 3= Immature-Depressive; PTSS = post-traumatic stress symptoms; ACE = adverse childhood experiences.

Discussion

This study adds to preliminary findings about the reliability and validity of the Italian DMRS-SR-30 (Di Giuseppe et al., 2020), a novel self-report measure reflecting the hierarchical organization of defense mechanisms theorized by George Vaillant (Vaillant, 1971, 1977, 1992) and further operationalized in the DMRS manual (Perry, 1990). We examined the psychometric properties of the English version of the DMRS-SR-30 including internal consistency, concurrent, convergent and discriminant validity. Overall, findings suggest strong psychometric properties of the measure, similar to those of the Italian version (Di Giuseppe et al., 2020), highlighting its utility for use in future research studies.

The factor structure that emerged from exploratory PAF analyses was generally in line with expectations and provided further confirmation of Vaillant’s theory of the tripartite organization of non-psychotic defense mechanisms (1977). The three extracted factors differed only slightly from the three DMRS defensive categories. Consistent with the DMRS hierarchy of defense mechanisms, the mature factor included all defenses belonging to the high-adaptive defense level (synonymous with the mature defensive category), which leads individuals to the best adjustment and possible resolution of internal and external stressors. Mature defenses can reach the subject’s awareness and, for this reason, they are often associated with coping strategies (Cramer, 1998).

Neurotic defenses were all included in Factor 2 with the addition of denial and autistic fantasy. Due to the evident similarities with the Vaillant’s classification of defense mechanisms (American Psychiatric Association, 1994; Vaillant, 1971, 1977), we labeled this factor as mental inhibition and avoidance in order to capture the defensive function of all included defenses. These defense mechanisms protect the self by partially or completely removing the internal conflict or stressor from conscious awareness. This unawareness allows the individual to maintain distance from charged feelings (obsessional defenses), unacceptable ideas (neurotic defenses), or from the emotional and cognitive aspects of internal or external stressors (disavowal defenses).

As expected, immature defenses naturally grouped into the third factor, labeled as immature-depressive due to the presence of all depressive defenses with the addition of two more immature defenses (for detailed description of the hierarchical organization of defense mechanisms see Di Giuseppe & Perry, 2021; Perry, 2014). These defense mechanisms keep emotional and cognitive aspects of internal or external stressors outside of awareness in order to protect the individual from experiencing unmanageable feelings, desires, and thoughts, which are handled as follows: (1) blowing off steam by expressing impulses directly and without prior thought, (2) circumventing inhibitions (action defenses); (3) preventing the integration of contrasting perceptions of the self and others that provoke intolerable conflicting emotions (major image-distorting defenses); (4) protecting of the subject from uncomfortable feelings and thoughts by distorting some aspects of reality (disavowal defenses); and (5) shoring up vulnerable self-esteem by distorting self and others’ images (minor image-distorting defenses). These types of immature defenses are common in patients with severe psychiatric distress, including depression (McDonald et al., 2020; Perry et al., 2020; Tanzilli et al., 2021). They can result in poor adaptation, maladaptive behaviors, problematic relationships, and short-term gratification coupled with exacerbation of internal conflicts and external stressors.

Finally, omnipotence was the only defense mechanism excluded from any factor. This may be accounted for by the low base rate of this defense on the survey; however, it may have also been due to the unique lockdown conditions that participants were experiencing at the time of data collection. The diffuse feeling of powerlessness and vulnerability among all respondents, independent of their usual defensive functioning, could have somewhat influenced the low self-reported use of omnipotence and consequently biased results. Another possibility is self-selection bias, that is, individuals high on omnipotence did not feel vulnerable to the pandemic and chose not to participate in the survey. Further investigation is needed to evaluate the potential bias of the self-reported use of omnipotence in the DMRS-SR-30.

Internal consistency analyses showed excellent reliability for the overall defensive functioning (ODF) score, which includes all 30 items of the measure. Acceptable to good internal consistency was found for the three factors of the DMRS-SR-30 and the three defense categories, and the depressive defense subcategory. Consistent with a preliminary validation of the DMRS-SR-30 (Di Giuseppe et al., 2020), our findings confirmed greater reliability for subscales including four or more items as compared to 3-item subscales. In particular, minor image-distorting defenses had questionable reliability, possibly due to environmental conditions during the time of data collection (described above). Moreover, the absence of items discriminating between self and other image distortion could have influenced internal consistency, which suggests the need for further testing and the possible inclusion of additional items in this subscale. According to these findings, the DMRS-SR-30 three-factor model, generally similar to the tripartite organization of mature, neurotic, and immature defensive categories, is a reliable alternative to the classic seven-level scoring approach (Perry, 2014; Perry & Henry, 2004). It remains to be seen whether this same factor structure will replicate using the original DMRS observer-based method.

We tested concurrent validity by comparing ODF, and defensive categories and factors of the DMRS-SR-30 with defense styles of the DSQ-40. All correlations were in the expected direction along the table’s diagonal, and almost all relationships were significant. The DMRS-SR-30 domain with the strongest bivariate correlation with the DSQ-40 was the immature defense domain. Similarly, depressive and non-depressive immature defenses were moderately correlated with the DSQ-40 immature defense style. While the off-diagonal correlations between the two measures were generally lower than those on the diagonal, one puzzling finding was the magnitude of the negative correlation (r = −.60) between the DMRS-SR-30 Mature defenses and the DSQ-40 Immature defense Style. An earlier study, comparing observer-coded methods of defense assessment, demonstrated strong concurrent validity between the DMRS-SR-30 and the observer-rated DMRS and DMRS-Q with very large to extremely large effect sizes for the DMRS (Cohen’s d range 1.06-2.41) and moderate to extremely large effect sizes for the DMRS-Q (Cohen’s d range 0.56-1.61) (Di Giuseppe et al., 2020). The current study examined the relationship between two self-report measures of defenses and found significant correlations between the three DMRS-SR-30 factors and their corresponding factors on the DSQ-40, indicating modest to good concurrent validity between the two.

Finally, correlations with external criteria – depression, anxiety, post-traumatic stress, and somatization, and adverse childhood experience – supported the convergent and discriminant validity of the DMRS-SR-30. DMRS-SR-30 scores were associated with various presenting problems and childhood experiences in a conceptually expected manner. Pearson correlations confirmed that ODF and the mature defense level were negatively associated with depression, anxiety, posttraumatic symptoms, and somatization, whereas immature and neurotic defenses were positively associated with the same measures. In particular, less adaptive immature defenses generally showed stronger relationships to symptoms than did the neurotic defenses which are in the middle of the hierarchy of adaptation. However, hysterical defense mechanisms (Level 5a) showed strong correlations with depression, anxiety, posttraumatic symptoms, and somatization whereas other neurotic defenses (Level 5 b) were only slightly related to depression. In line with a previous study (Di Giuseppe, Perry, Conversano et al., 2020), our findings provide further confirmation for the non-specificity of minor image-distorting and other neurotic defense mechanisms in relation to psychological distress, post-traumatic symptoms, somatization, and anxiety. However, this statement should be tempered by our assessment concerns noted above, regarding the minor image-distorting items combining self and other assessments. Conversely, defense mechanisms were unrelated to adverse child experiences, indicating good discriminant validity of the DMRS-SR-30. Convergent and discriminant validity were further demonstrated by comparing the three hierarchically ordered factors of the DMRS-SR-30 and the three defense styles of DSQ-40. The associations between the DMRS-SR-30 and the psychological measures considered in this study were substantially higher than the correlations between the DSQ-40 the same measures. Both measures were unrelated to traumatic experiences during childhood. These findings demonstrate both convergent and discriminant validity of DMRS-SR-30, while indicating superior convergent validation for the DMRS-SR-30 and its ability to capture the construct of defense mechanisms in comparison to a widely used self-report measure for assessing defense mechanisms.

The present study has also several limitations. First, the use of a community sample may not have yielded sufficient prevalence of individuals with psychiatric disorders, which may limit the generalizability of study results. Individuals with high proportions of immature defenses were less well represented. Further studies including participants with physical and mental symptoms and disorders and an equal proportion of women and men should be conducted to replicate and extend these findings. Second, the absence of other DMRS-based measures did not allow to further test criterion validity. Future research would benefit from including both self-report and observer-rated measures based on the DMRS manual in order to confirm the good criterion validity of the questionnaire (Di Giuseppe et al., 2020). Third, internal consistency of individual defenses could not be tested, since 26 of the 28 individual defenses are single-item scales. However, as other authors found in personality research (Widiger, 2017; Widiger & Crego, 2019) the low number of items included in the DMRS-SR-30 is also one of the strengths of this measure, which takes just five minutes to complete (see supplement material Appendix A). Finally, future research should collect new data and run CFA on the factors to show unidimensionality, and also examine the test-retest reliability of the scales.

Ease of administration for this questionnaire may facilitate the rapid assessment of defense mechanisms in both research and clinical settings and may serve as a useful tool to delineate the role of dynamic factors in overall psychological functioning.

Conclusion

The psychometric properties of the English version of the DMRS-SR-30 have been demonstrated in this study and confirm previous findings tested on the Italian version of the questionnaire (Di Giuseppe et al., 2020). A strength of the DMRS-SR-30 is its strong theoretical background and its ability to provide a comprehensive assessment of the whole hierarchy of defense mechanisms (Di Giuseppe & Perry, 2021; Perry, 1990). Moreover, it uses a rapid, multi-level scoring system that includes a global assessment of defensive functioning, defense levels, and specific individual defense scores (see supplement material Appendix B). The DMRS-SR-30 represents a valuable contribution to the measurement of self-reported defense mechanisms in both community and clinical populations. In particular, it may be a useful tool to differentiate subgroups based on defensive functioning, which may respond to tailored interventions in personalized medicine (Chen et al., 2020; Conversano, 2021; Conversano & Di Giuseppe, 2021; Delgadillo et al., 2020). Future studies should examine the stability of the DMRS-SR-30 factor structure in other samples and degree to which changes in defensive functioning on the DMRS-SR-30 can predict changes in other aspects of mental functioning, such as psychiatric symptoms, personality, quality of life, and adherence to treatment (Flückiger et al., 2020; Helmich et al., 2020; Webb et al., 2020).

Ethics approval statement

All the study procedures were conducted according to ethical standards outlined by the World Medical Association’s Declaration of Helsinki and were approved by the Institutional Review Board.

Disclosure statement

No potential conflict of interest was reported by the authors.