MMPI-3 Predictors of Anxiety Sensitivity and Distress Intolerance

Anxiety Sensitivity and Distress Intolerance

Although extensive evidence in the MMPI-3 Technical Manual supports MMPI-3 scale score validity, continued examination of empirical correlates of scale scores is important to comprehensively document clinical phenomena that can be measured using MMPI-3 scales. Practically, this research practice can expand the pool of probabilistic statements clinicians can make about a test-taker based on their MMPI-3 scale score elevations. Moreover, this practice could also provide researchers with a more nuanced understanding of the constructs reflected by MMPI-3 scale scores and, in turn, facilitate use of the instrument to examine important empirical questions related to personality and psychopathology.

There are no published data that speak to the ability of MMPI-3 scale scores to assess anxiety sensitivity (AS) and distress intolerance (DI), two closely related and commonly studied constructs reflecting sensitivity to and intolerance of various affective experiences (Bernstein et al., 2009; Zvolensky et al., 2011). The first of these constructs, AS reflects the proclivity to experience fear in response to arousal associated with fear and anxiety experiences (i.e., “fear of fear;” Reiss, 1991; Reiss & McNally, 1985). Individuals with elevated AS tend to interpret fear sensations as threatening, indicative of future harm, or otherwise aversive which, in turn, intensifies these fear experiences. The construct has been conceptualized as multidimensional, with subdimensions reflecting fear of the physical, cognitive, or social consequences of experiencing anxiety or fear (Taylor et al., 2007). Meanwhile, DI reflects the tendency to interpret distressing emotional experiences as overwhelming, unacceptable, and intolerable (Simons & Gaher, 2005; Zvolensky et al., 2011). Like AS, individuals with elevated standing on DI hold inflexible views of emotional distress as aversive, see themselves as unable to regulate their emotions, and become reluctant to experience distressing emotions (Clen et al., 2011). As seen in these definitions, AS and DI share some features and may be best conceptualized as facets of a global affect intolerance dimension (Conway et al., 2020; Juarascio et al., 2020). However, important distinctions exist between the two constructs, with anxiety-related interoceptive and somatic sensations at the core of AS, but not DI (Zvolensky et al., 2011). Moreover, previous empirical evidence supports the distinctiveness of the two constructs, as they possess some unique empirical correlates when considered conjointly (i.e., AS with anxious arousal and fear disorder symptomology, DI with anhedonic depression and substance use disorder symptomology; Allan et al., 2015; Bernstein et al., 2009).

Past research has supported the relevance of AS and DI to several different psychopathological difficulties. AS and DI have been theorized to contribute to psychological dysfunction through similar processes, with interpretations of fear, anxiety, or distressing emotional experiences as harmful or intolerable leading to difficulties in emotion regulation, avoidance behaviors, and additional negative emotions that “amplify” the original emotional experiences (Reiss, 1991; Zvolensky et al., 2011). Given these consequences of AS and DI, it is not surprising that these constructs have been implicated across numerous psychological disorders, including those characterized by internalizing dysfunction (e.g., depressive, anxiety, and fear disorders; Allan et al., 2014; Allan et al., 2015; Cox et al., 2001; Keough et al., 2010; Michel et al., 2016; Naragon-Gainey, 2010) and externalizing dysfunction (e.g., substance use, impulsivity; Allan et al., 2015; Kaiser et al., 2012; Marshall-Berenz et al., 2011; Schmidt et al., 2007). Moreover, preliminary evidence from longitudinal studies of AS and DI suggest that the constructs are not only related to these diagnoses but also contribute etiologically to disorder development (Cummings et al., 2013; Schmidt et al., 1997).

Given the importance of AS and DI in a variety of psychopathological dysfunctions, targeting these constructs in treatment settings may be an effective method for simultaneously treating comorbid psychological problems. Indeed, several treatments designed to reduce AS and DI through interoceptive exposure to feared experiences, increasing acceptance of distress, and modifying interpretations of negative emotions have been developed and empirically investigated (e.g., Bornovalova et al., 2012; Boswell et al., 2013). Individual modules serving similar functions also appear in manuals for commonly used treatment modalities, such as Dialectical-Behavior Therapy (Linehan, 2015) and the Unified Protocol for Transdiagnostic Treatment of Emotional Disorders (Allen et al., 2008). Initial evidence suggests that these interventions not only reduce standing on AS and DI but also lead to reductions in comorbid disorder symptomology (Bornovalova et al., 2012; Boswell et al., 2013). Thus, individuals with maladaptive reactions to fear, anxiety, or distress experiences may benefit from interventions explicitly aimed at altering these interpretations.

Measuring AS and DI

Given the importance of AS and DI to the etiology and treatment of various psychological difficulties, the ability to assess for AS and DI using MMPI-3 scales could be beneficial to practicing clinicians and researchers alike. The MMPI-3 may particularly well-suited for measuring AS and DI for several reasons. First, the instrument indexes conceptually indicated constructs that are adjacent to AS and DI, including both broad and narrow phenomena in the domains of internalizing, somatic, and interpersonal dysfunction. Next, the MMPI family of instruments are also commonly used in clinical practice (Wright et al., 2017). Last, several instruments do exist for directly measuring AS and DI, such as the Anxiety Sensitivity Index–3 (ASI-3; Taylor et al., 2007), Distress Tolerance Scale (DTS; Simons & Gaher, 2005), and Distress Intolerance Index (DII; McHugh & Otto, 2012). However, although the sound psychometric properties of scores on these measures are well documented (McHugh & Otto, 2012; Simons & Gaher, 2005; Taylor et al., 2007), they lack normative samples and formal interpretive guidelines to facilitate identification of clinically significant levels of AS and DI. As such, these instruments do not meet standards for psychological testing in applied settings (American Educational Research Association, 2014). In contrast, MMPI-3 supporting materials contain clear guidelines for interpretations of clinically elevated scores, allowing for ease of use in clinical settings. In all, practicing clinicians may be able to assess for possible elevated AS or DI using only MMPI-3 scale scores (i.e., without the use of a collateral instrument).

Current Study

Given this gap in the literature, the goals of the current study were to determine whether conceptually relevant MMPI-3 scales could predict an individual’s standing on the constructs of AS and DI. We planned to investigate this initial research question by examining associations between scores on MMPI-3 scales and commonly used measures of AS and DI, as well as evaluating the amount of variance accounted for in models predicting AS and DI from MMPI-3 scale scores. The second aim of the current study was to identify the specific MMPI-3 scales that serve as the strongest individual predictors of each construct when considered concurrently with other MMPI-3 scales.

Regarding hypotheses, the core features of AS and its subdimensions are fear of physical, cognitive, and social consequences of anxiety and fear experiences. Moreover, the construct has been related to various internalizing phenomena, such as depression, anxiety, panic, and social phobia, as previously described. As such, we hypothesized that a combination of internalizing, somatic, and interpersonal MMPI-3 scales at the H-O, RC, SP, and PSY-5 level would best predict AS. Specifically, given the constructs indexed by MMPI-3 scales (see Table 1 for scale descriptions), we hypothesized that scores on scales such as Emotional/Internalizing Dysfunction (EID), Somatic Complaints (RC1), Dysfunctional Negative Emotions (RC7), Anxiety-Related Experiences (ARX), and Negative Emotionality/Neuroticism (NEGE) would emerge as the strongest predictors of the global AS construct across scale families. Additionally, because of the unique features of each AS subdimension (i.e., the focus on physical, cognitive, or social consequences of fear), we also hypothesized that scores on scales such as RC1 and Neurological Complaints (NUC) would best predict physical concerns at the RC and SP levels, RC1 and Cognitive Complaints (COG) would most strongly predict cognitive concerns, while RC7 and Shyness (SHY) would best predict social concerns. Last, DI is defined as intolerance of negative emotional experiences, but is not operationalized as containing the specific physical or social features of AS. As such, we predicted that DI would be best predicted by the MMPI-3’s internalizing scales, although we included the same combination of internalizing, somatic, and interpersonal MMPI-3 scales used to predict AS in the prediction of DI to allow direct comparisons of predictors of each construct. Given previously observed correlations between MMPI-2-RF scale scores and DI measures (e.g., Kremyar et al., 2020), as well as the constructs underlying MMPI-3 scales, we predicted that scores on scales such as EID, Demoralization (RCd), RC7, STR, and NEGE would offer the strongest prediction of scores on DI measures.

OPEN IN VIEWER

Table 1.

Descriptive Statistics for MMPI-3 Scales and Measures of Anxiety Sensitivity and Distress Intolerance.

Scales	Intended to assess	Items	α	M	SD
MMPI-3
EID	Generalized emotional distress and internalizing dysfunction	42	.88	14.57	9.30
RCd	Demoralization; pervasive unhappiness with poor coping	17	.90	6.23	4.98
RC1	Somatic complaints; physical symptoms	21	.84	4.91	4.33
RC2	Anhedonia; denial of positive emotional experiences	14	.74	3.20	2.76
RC7	Various negative emotional experiences; fear, anxiety, anger	19	.85	9.88	4.83
MLS	Nonspecific sense of poor physical health	7	.65	2.84	1.76
NUC	Physical symptoms such as dizziness and numbness	10	.61	2.32	1.95
COG	Cognitive problems; impaired memory, poor concentration	11	.86	4.40	3.29
SUI	Suicidal ideation and behaviors	7	.78	0.58	1.16
HLP	Perceived inability to overcome problems, achieve goals	7	.65	1.03	1.35
SFD	Self-doubt; poor self-confidence	7	.84	2.41	2.28
NFC	Perceived inability to make decisions, handle crises	9	.76	4.46	2.58
STR	Difficulties handling stress; stress reactivity	6	.63	3.17	1.71
WRY	Proclivity to excessively worry	7	.76	3.92	2.18
CMP	Compulsive and repetitive behaviors	8	.74	4.48	2.30
ARX	Anxiety, panic, and trauma-related experiences	15	.84	5.80	3.89
ANP	Tendency to experience angry emotions and behaviors	12	.86	4.82	3.55
BRF	Fear restricting activities of daily living	7	.57	1.29	1.41
DOM	Assertiveness; social self-confidence	9	.65	6.23	2.06
SAV	Avoidance of and ambivalence toward social situations	9	.83	2.40	2.58
SHY	Fear, embarrassment, and discomfort in social situations	7	.77	3.43	2.26
NEGE	Tendency to experience negative affect and emotions	15	.87	7.51	4.25
INTR	Tendency to be detached and experience anhedonia	14	.80	3.56	3.20
AS
ASI-3	Fear of consequences of anxiety sensations; “fear of fear”	18	.90	18.19	12.52
Physical	Fear of physical anxiety symptoms	6	.85	5.03	5.04
Cognitive	Fear of mental/cognitive anxiety symptoms	6	.88	3.82	4.54
Social	Fear of social consequences of anxiety symptoms	6	.78	9.34	5.40
DI
DTS	Perceived ability to endure negative emotions	15	.89	3.23	0.73
DII	Perceived ability to endure negative somatic/emotional states	10	.82	26.43	7.09
FDS–EI	Beliefs about controllability and aversiveness of distress	7	.82	18.69	5.74

Note. α = Cronbach’s alpha; MMPI-3 = Minnesota Multiphasic Personality Inventory–3; EID = Emotional/Internalizing Dysfunction; RC = Restructured Clinical; RCd = Demoralization; RC1 = Somatic Complaints; RC2 = Low Positive Emotions; RC7 = Dysfunctional Negative Emotions; SP = Specific Problems; MLS = Malaise; NUC = Neurological Complaints; COG = Cognitive Complaints; SUI = Suicidal/Death Ideation; HLP = Helplessness/Hopelessness; SFD = Self-Doubt; NFC = Inefficacy; STR = Stress; WRY = Worry; CMP = Compulsivity; ARX = Anxiety-Related Experiences; ANP = Anger Proneness; BRF = Behavior-Restricting Fears; DOM = Dominance; SAV = Social Avoidance; SHY = Shyness; PSY-5 = Personality Psychopathology–5; NEGE = Negative Emotionality/Neuroticism; INTR = Introversion/Low Positive Emotionality; AS = Anxiety sensitivity; ASI-3 = Anxiety Sensitivity Index–3; DI = Distress intolerance; DTS = Distress Tolerance Scale; DII = Distress Intolerance Index; EI = Emotional Intolerance; FDS = Frustration Discomfort Scale.

Method

Measures

Descriptive statistics for all examined MMPI-3 scales and measures of AS and DI are presented in Table 1. The distribution of scores on all measures did not strongly deviate from normality (apart from Suicidal/Death Ideation [SUI] on the MMPI-3; skewness = 2.59, kurtosis = 7.19) and demonstrated sufficient variability. Internal consistencies for scores on all scales were adequate to excellent (αs = .57-.90; see Table 1). MMPI-3 scale score reliabilities were comparable to those of scale scores in the MMPI-3 normative sample (Ben-Porath & Tellegen, 2020).

MMPI-3

As described previously, the MMPI-3 (Ben-Porath & Tellegen, 2020) is a 335-item, true–false inventory of personality, psychopathology, and social and behavioral functioning. In the current study, H-O, RC, SP, and PSY-5 scales from the somatic/cognitive, internalizing, and interpersonal domains were examined (see Table 1 for a list of scales and their descriptions). Intercorrelations between MMPI-3 scale scores, presented in Table 2, were comparable to those observed in the MMPI-3 normative sample (Ben-Porath & Tellegen, 2020).

OPEN IN VIEWER

Table 2.

Intercorrelations Between Examined MMPI-3 Scales.

	EID	RCd	RC1	RC2	RC7	MLS	NUC	COG	SUI	HLP	SFD	NFC	STR	WRY	CMP	ARX	ANP	BRF	DOM	SAV	SHY	NEGE
RCd	.93
RC1	.63	.59
RC2	.74	.65	.46
RC7	.74	.66	.54	.31
MLS	.67	.64	.57	.49	.50
NUC	.43	.41	.79	.28	.42	.42
COG	.53	.57	.44	.34	.54	.43	.39
SUI	.41	.42	.23	.33	.26	.20	.15	.26
HLP	.60	.61	.41	.48	.39	.37	.30	.37	.51
SFD	.87	.90	.53	.64	.59	.56	.32	.51	.41	.57
NFC	.71	.74	.50	.38	.69	.51	.38	.56	.26	.44	.63
STR	.66	.53	.45	.33	.62	.42	.28	.33	.20	.27	.50	.53
WRY	.69	.67	.54	.31	.70	.47	.37	.44	.26	.37	.62	.63	.60
CMP	.37	.33	.36	.13	.58	.19	.34	.26	.21	.20	.30	.38	.39	.47
ARX	.71	.66	.68	.34	.75	.56	.49	.50	.28	.40	.63	.61	.64	.70	.44
ANP	.49	.44	.29	.18	.67	.36	.22	.41	.17	.33	.39	.41	.35	.37	.33	.41
BRF	.30	.27	.31	.11	.46	.32	.24	.27	.02	.17	.21	.35	.30	.30	.23	.38	.28
DOM	−.32	−.28	−.25	−.50	.00	−.20	−.17	−.16	−.06	−.16	−.31	−.25	−.22	−.14	.04	−.14	.15	−.01
SAV	.52	.41	.32	.62	.25	.32	.20	.16	.17	.20	.37	.31	.30	.26	.16	.24	.13	.02	−.38
SHY	.65	.53	.38	.53	.49	.38	.25	.30	.15	.27	.51	.52	.39	.42	.18	.40	.24	.18	−.36	.52
NEGE	.78	.70	.62	.36	.81	.54	.43	.48	.27	.40	.66	.67	.73	.87	.48	.85	.47	.36	−.15	.29	.47
INTR	.58	.47	.36	.75	.25	.40	.22	.20	.21	.26	.42	.33	.26	.24	.15	.24	.17	.02	−.38	.93	.51	.28

Note. All values represent Pearson zero-order correlation (r) coefficients. EID = Emotional/Internalizing Dysfunction; RCd = Demoralization; RC1 = Somatic Complaints; RC2 = Low Positive Emotions; RC7 = Dysfunctional Negative Emotions; MLS = Malaise; NUC = Neurological Complaints; COG = Cognitive Complaints; SUI = Suicidal/Death Ideation; HLP = Helplessness/Hopelessness; SFD = Self-Doubt; NFC = Inefficacy; STR = Stress; WRY = Worry; CMP = Compulsivity; ARX = Anxiety-Related Experiences; ANP = Anger Proneness; BRF = Behavior-Restricting Fears; DOM = Dominance; SAV = Social Avoidance; SHY = Shyness; NEGE = Negative Emotionality/Neuroticism; INTR = Introversion/Low Positive Emotionality.

All coefficients ≥ |.21| significant at p < .001; all coefficients ≤ |.20| and ≥ |.16| significant at p < .01; all coefficients ≥ |.13| significant at p < .05.

ASI-3

The ASI-3 (Taylor et al., 2007) is an 18-item instrument measuring AS. Test-takers indicate the extent of their agreement with each item using a 5-point Likert-type scale ranging from 1 (very little) to 5 (very much). In line with contemporary conceptualizations of AS as a multidimensional construct, items are summed to generate both a composite AS score and subscale scores reflecting physical, cognitive, and social concerns. Among nonclinical samples, composite and subscale scores on the ASI-3 have previously demonstrated meaningful associations with scores on collateral measures of somatization, interpersonal sensitivity, and symptoms of depressive, anxiety, and fear disorders, supporting the convergent validity of these scores (Osman et al., 2010).

DTS

The DTS (Simons & Gaher, 2005) is a 15-item instrument measuring DI. Test-takers indicate their agreement with items using a 5-point Likert-type scale ranging from 1 (strongly agree) to 5 (strongly disagree). In the current study, items on the DTS were reverse coded, such that higher composite scores were indicative of greater DI (i.e., less distress tolerance). Among nonclinical samples, DTS scores have previously demonstrated moderate to strong associations with scores on measures of emotion regulation and affective lability, supporting the convergent validity of scores on the measure (Simons & Gaher, 2005).

EI Subscale of the Frustration Discomfort Scale (FDS)

The FDS (Harrington, 2005) is a 28-item measure of frustration intolerance based around the principles of rational-emotive behavior therapy. Of relevance to the current study, the seven-item EI subscale of the measure has previously been examined as a measure of DI (Zvolensky et al., 2011). Test-takers rate the extent of their agreement with items using a 5-point Likert-type scale ranging from 1 (absent) to 5 (very strong), with higher scores indicating greater DI. Among nonclinical samples, scores on EI have demonstrated moderate to strong associations with scores on measures of anger, depression, and anxiety, supporting the convergent validity of these scores (Harrington, 2006).

DII

The DII (McHugh & Otto, 2012) is a 10-item measure of DI composed of select items from the DTS, ASI, and FDS. Test-takers rate the extent of their agreement with items using a 5-point Likert-type scale ranging from 1 (very little) to 5 (very much), with higher scores indicating greater DI. Scores on the DII have previously demonstrated strong associations with scores on measures of constructs such as emotion dysregulation and experiential avoidance, supporting the validity of these scores (McHugh et al., 2012).

Intercorrelations between scores on examined measures of AS and DI are presented in Table 3. Observed correlation coefficients between scores on measures of AS and DI were comparable to those in previous examinations of these constructs (i.e., Allan et al., 2015; Keough et al., 2010) and are consistent with conceptualizations of AS and DI as two related yet distinct constructs.

OPEN IN VIEWER

Table 3.

Intercorrelations Between Measures of Anxiety Sensitivity and Distress Intolerance.

	ASI-3: Total	ASI-3: Physical Concerns	ASI-3: Cognitive Concerns	ASI-3: Social Concerns	DTS	DII
ASI-3: Physical Concerns	.85
ASI-3: Cognitive Concerns	.81	.58
ASI-3: Social Concerns	.84	.56	.50
DTS	.54	.41	.45	.49
DII	.64	.49	.51	.60	.89
EI Subscale of FDS	.55	.46	.48	.45	.54	.75

Note. All values represent Pearson zero-order correlation (r) coefficients. ASI-3 = Anxiety Sensitivity Index–3; DTS = Distress Tolerance Scale; DII = Distress Intolerance Index; EI = Emotional Intolerance; FDS = Frustration Discomfort Scale.

All coefficients significant at p < .001.

Results

Anxiety Sensitivity Analyses

Results of zero-order correlational, regression, and dominance analyses predicting scores on the ASI-3 from MMPI-3 scales are presented in Table 4. First, as seen in the first block of Table 4, at the H-O level, EID scores were strongly associated with composite ASI-3 scores (r = .57) and Social Concerns subscale scores (r = .54) and moderately associated with Physical Concerns (r = .42) and Cognitive Concerns (r = .47) subscale scores. In regression models, EID scores also predicted ASI-3 composite and subscale scores with similar magnitudes (R²s = .18-.33).

OPEN IN VIEWER

Table 4.

Predicting Anxiety Sensitivity and Associated Subdimensions From MMPI-3 Scale Scores.

	ASI-3: Total			ASI-3: Physical Concerns			ASI-3: Cognitive Concerns			ASI-3: Social Concerns
	r	β	DW	r	β	DW	r	β	DW	r	β	DW
HO		R2 = .33***			R2 = .18***			R2 = .22***			R2 = .29***
EID	.57 ***	.57***	—	.42 ***	.42***	—	.47 ***	.47***	—	.54 ***	.54***	—
RC		R2 = .43***			R2 = .28***			R2 = .29***			R2 = .33***
RCd	.56 ***	.28***	.130	.43 ***	.14	.078	.49 ***	.29***	.121 a	.49 ***	.19*	.091
RC1	.57 ***	.32***	.148 a	.48 ***	.32***	.128 a	.45 ***	.21**	.092	.49 ***	.25***	.103 b
RC2	.33 ***	−.06	.032	.25***	—	—	.28***	—	—	.30 ***	−.01	.027
RC7	.54 ***	.20**	.117	.42 ***	.15*	.074	.43 ***	.13	.077	.50 ***	.24**	.107 b
SP		R2 = .47***			R2 = .30***			R2 = .36***			R2 = .41***
MLS	.40 ***	−.04	.021	.35 ***	.04	.022	.29***	—	—	.34 ***	−.04	.016
NUC	.35 ***	.03	.019	.30 ***	.04	.015	.28***	—	—	.31 ***	.02	.014
COG	.45 ***	.11	.037	.34 ***	.05	.020	.50 ***	.30***	.100 a	.30 ***	−.05	.011
SUI	.17**	—	—	.07	—	—	.21***	—	—	.15*	—	—
HLP	.37 ***	.06	.024	.28***	—	—	.36 ***	.09	.032	.30 ***	.05	.014
SFD	.54 ***	.11	.054	.40 ***	.07	.030	.49 ***	.14	.061	.46 ***	.01	.035
NFC	.52 ***	.01	.042	.42 ***	.04	.031	.40 ***	−.09	.030	.47 ***	.04	.038
STR	.50 ***	.11	.050	.39 ***	.07	.030	.36 ***	.06	.027	.50 ***	.12	.056
WRY	.59 ***	.20**	.081 b	.47 ***	.17*	.054 b	.45 ***	.14	.050	.56 ***	.21**	.078 a
CMP	.39 ***	.07	.028	.31 ***	.07	.020	.29***	—	—	.36 ***	.07	.027
ARX	.60 ***	.18*	.080 b	.47 ***	.09	.047	.48 ***	.13	.057	.56 ***	.23**	.080 a
ANP	.25***	—	—	.17**	—	—	.24***	—	—	.23***	—	—
BRF	.32 ***	.08	.020	.31 ***	.12*	.026	.21***	—	—	.27***	—	—
DOM	−.21***	—	—	−.18**	—	—	−.11	—	—	−.24***	—	—
SAV	.23***	—	—	.14*	—	—	.15*	—	—	.27***	—	—
SHY	.34 ***	.02	.016	.24***	—	—	.20**	—	—	.41 ***	.16**	.041
PSY-5		R2 = .40***			R2 = .24***			R2 = .24***			R2 = .35***
NEGE	.63 ***	.63***	—	.49 ***	.49***	—	.49 ***	.49***	—	.59 ***	.59***	—
INTR	.24***	—	—	.16**	—	—	.17**	—	—	.25***	—	—

Note. Bold text indicates r ≥ .30. ASI-3 = Anxiety Sensitivity Index–3; r = Pearson zero-order correlation; β = Standardized regression weight; DW = Dominance weights; R² = Coefficient of determination; H-O = Higher-Order; EID = Emotional/Internalizing Dysfunction; RC = Restructured Clinical; RCd = Demoralization; RC1 = Somatic Complaints; RC2 = Low Positive Emotions; RC7 = Dysfunctional Negative Emotions; SP = Specific Problems; MLS = Malaise; NUC = Neurological Complaints; COG = Cognitive Complaints; SUI = Suicidal/Death Ideation; HLP = Helplessness/Hopelessness; SFD = Self-Doubt; NFC = Inefficacy; STR = Stress; WRY = Worry; CMP = Compulsivity; ARX = Anxiety-Related Experiences; ANP = Anger Proneness; BRF = Behavior-Restricting Fears; DOM = Dominance; SAV = Social Avoidance; SHY = Shyness; PSY-5 = Personality Psychopathology–5; NEGE = Negative Emotionality/Neuroticism; INTR = Introversion/Low Positive Emotionality.

a

Complete dominance over remaining unmarked scales in a given scale family. ^bConditional dominance over remaining unmarked scales in a given scale family.

*

p < .05. ^**p < .01. ^***p < .001.

Next, as seen in the second block of Table 4, at the RC level, RCd, RC1, and RC7 were moderately to strongly associated with ASI-3 composite and subscale scores (rs = .42-.57), while Low Positive Emotions (RC2) was moderately associated with ASI-3 composite scores (r = .33) and Social Concerns subscale scores (r = .30). Regression models predicting ASI-3 composite and subscale scores from these scale scores accounted for a large amount of variance in each ASI-3 scale score (R²s = .28-.43). Standardized regression coefficients and DWs were examined to evaluate the strongest predictor of scores on each scale. For composite ASI-3 scores, RC1 was the strongest predictor (β = .32; DW = .148) and demonstrated complete dominance over other RC scales, including RCd (β = .28; DW = .130) and RC7 (β = .20; DW = .117). For the Physical Concerns subscale scores, RC1 was again the strongest predictor (β = .32; DW = .128) and demonstrated complete dominance over other RC scales, including RC7 (β = .15; DW = .074). Cognitive Concerns subscale scores were best predicted by RCd (β = .29; DW = .121), which demonstrated complete dominance over remaining RC scales, including RC1 (β = .21; DW = .092). Lastly, for Social Concerns subscale scores, RC1 (β = .25; DW = .103) and RC7 (β = .24; DW = .107) offered comparable strengths in prediction and were conditionally dominant over remaining RC scales, including RCd (β = .19; DW = .091).

At the SP level, as seen in the third block of Table 4, a variety of SP scale scores demonstrated moderate to strong associations with ASI-3 composite scores, including somatic (Malaise [MLS], NUC, COG), internalizing (HLP, SFD, Inefficacy [NFC], STR, WRY, CMP, ARX, Behavior-Restricting Fears [BRF]), and interpersonal scales (SHY; rs = .32-.59). When considered conjointly, these scale scores predicted a large amount of variance in ASI-3 composite scores (R² = .47), with WRY (β = .20; DW = .081) and ARX (β = .18; DW = .080) offering comparable strengths of prediction and demonstrating conditional dominance over remaining SP scales. Next, similar somatic (MLS, NUC, COG) and internalizing (SFD, NFC, STR, WRY, CMP, ARX, BRF) scale scores demonstrated moderate to strong associations with Physical Concerns subscale scores (rs = .31-.47). These scales also predicted a large amount of variance in Physical Concerns subscale scores in regression analyses (R² = .30), with WRY again emerging as the strongest predictor and demonstrating conditional dominance over remaining SP scales (β = .17; DW = .054). Third, COG and various internalizing SP scale scores (HLP, SFD, NFC, STR, WRY, ARX) were moderately to strongly associated with Cognitive Concerns subscale scores (rs = .36-.50). In regression analyses, these scales together predicted a large amount of variance (R² = .36), with COG scores emerging as the strongest predictor of Cognitive Concerns subscale scores and demonstrating complete dominance over all remaining SP scales (β = .30; DW = .100). Last, somatic (MLS, NUC, COG), internalizing (HLP, SFD, NFC, STR, WRY, CMP, ARX), and interpersonal (SHY) SP scale scores demonstrated moderate to large associations with Social Concerns subscale scores (rs = .30-.56). When entered into regression analyses, these scale scores accounted for a large amount of variance in Social Concerns subscale scores (R² = .41). ARX (β = .23; DW = .080) and WRY (β = .21; DW = .078) emerged as the strongest predictors and demonstrated complete dominance over all remaining SP scales, including SHY (β = .16; DW = .041).

Finally, at the PSY-5 level in the fourth block of Table 4, NEGE was moderately to strongly associated with ASI-3 composite and subscale scores (rs = .49-.63). In regression models, NEGE scores also accounted for a moderate to large amount of variance in each ASI-3 scale score (R²s = .24-.40).

Distress Intolerance Analyses

Results of zero-order correlational and regression analyses predicting scores on various measures of DI from MMPI-3 scales are presented in Table 5. First, at the H-O level in the first block of Table 5, EID scores were strongly associated with scores on the DTS, DII, and EI (rs = .54-.64) and accounted for a large amount of variance in scores on each measure (R²s = .29-.41).

OPEN IN VIEWER

Table 5.

Predicting Distress Intolerance From MMPI-3 Scale Scores.

	DTS			DII			EI Subscale of FDS
	r	β	DW	r	β	DW	r	β	DW
H-O		R2 = .36***			R2 = .41***			R2 = .29***
EID	.60 ***	.60***	—	.64 ***	.64***	—	.54 ***	.54***	—
RC		R2 = .38***			R2 = .46***			R2 = .37***
RCd	.56 ***	.36***	.146 a	.60 ***	.31***	.155	.53 ***	.26**	.118
RC1	.44 ***	.11	.065	.51 ***	.16**	.095	.46 ***	.16**	.080
RC2	.31 ***	−.05	.030	.34 ***	−.04	.034	.30 ***	−.03	.027
RC7	.54 ***	.27***	.134	.62 ***	.34***	.180 a	.55 ***	.30***	.141 a
SP		R2 = .43***			R2 = .54***			R2 = .39***
MLS	.40 ***	.03	.026	.46 ***	.06	.033	.42 ***	.07	.030
NUC	.29***	—	—	.32 ***	−.03	.013	.33 ***	.03	.017
COG	.41 ***	.09	.031	.45 ***	.09	.034	.43 ***	.12	.038
SUI	.27***	—	—	.26***	—	—	.29***	—	—
HLP	.39 ***	.14*	.036	.41 ***	.13*	.033	.39 ***	.14*	.035
SFD	.50 ***	.06	.046	.52 ***	−.02	.040	.45 ***	−.06	.031
NFC	.50 ***	.07	.046	.57 ***	.13*	.061	.51 ***	.11	.051
STR	.54 ***	.26***	.088 b	.60 ***	.28***	.100 b	.46 ***	.13	.043
WRY	.52 ***	.10	.057	.60 ***	.16*	.074	.51 ***	.14	.053
CMP	.28***	—	—	.34 ***	−.01	.016	.33 ***	.05	.020
ARX	.56 ***	.15	.071	.62 ***	.16*	.081	.53 ***	.10	.053
ANP	.30 ***	−.03	.012	.37 ***	.02	.020	.36 ***	.05	.022
BRF	.28***	—	—	.33 ***	.05	.019	.27***	—	—
DOM	−.07	—	—	−.11	—	—	−.07	—	—
SAV	.17**	—	—	.18**	—	—	.11	—	—
SHY	.31 ***	−.03	.013	.31 ***	−.08	.012	.24***	—	—
PSY-5		R2 = .37***			R2 = .45***			R2 = .30***
NEGE	.60 ***	.60***	—	.67 ***	.67***	—	.55 ***	.55***	—
INTR	.19**	—	—	.21***	—	—	.16**	—	—

Note. Bold text indicates r ≥ .30. DTS = Distress Tolerance Scale; DII = Distress Intolerance Index; EI = Emotional Intolerance; FDS = Frustration Discomfort Scale; r = Pearson zero-order correlation; β = Standardized regression weight; DW = Dominance weights; R² = Coefficient of determination; H-O = Higher-Order; EID = Emotional/Internalizing Dysfunction; RC = Restructured Clinical; RCd = Demoralization; RC2 = Low Positive Emotions; RC7 = Dysfunctional Negative Emotions; SP = Specific Problems; MLS = Malaise; NUC = Neurological Complaints; COG = Cognitive Complaints; SUI = Suicidal/Death Ideation; HLP = Helplessness/Hopelessness; SFD = Self-Doubt; NFC = Inefficacy; STR = Stress; WRY = Worry; CMP = Compulsivity; ARX = Anxiety-Related Experiences; ANP = Anger Proneness; BRF = Behavior-Restricting Fears; PSY-5 = Personality Psychopathology–5; NEGE = Negative Emotionality/Neuroticism; INTR = Introversion/Low Positive Emotionality.

*

p < .05. ^**p < .01. ^***p < .001.

a

Complete dominance over remaining unmarked scales in a given scale family. ^bConditional dominance over remaining unmarked scales in a given scale family.

Next, at the RC level in the second block of Table 5, RCd and RC7 scores were strongly related to scores on each DI measure (rs = .53-.62). Scores on RC1 were strongly related to DII scores (r = .51) and moderately related to DTS (r = .44) and EI scores (r = .46). RC2 scores were moderately related to scores on each measure (rs = .30-.34). When considered conjointly in regression analyses, RCd, RC1, RC2, and RC7 together accounted for a large amount of variance in each DI measure’s scores (R²s = .37-.46). Standardized regression weights and DWs indicated that RCd was the strongest predictor of DTS scores and had complete dominance over remaining RC scales (β = .36; DW = .146), including RC7 (β = .27; DW = .134). In contrast, RC7 was the strongest predictor of DII scores (β = .34; DW = .180) and EI scores (β = .30; DW = .141) with complete dominance over remaining RC scales for both measures, including RCd (β = .31 and DW = .155 for DII; β = .26 and DW = .141 for EI) and RC1 (β = .16 and DW = .095 for DII; β = .16 and DW = .080 for EI).

Third, at the SP scale level in the third block of Table 5, several somatic and internalizing scale scores, including MLS, COG, HLP, SFD, NFC, STR, WRY, ARX, and ANP, demonstrated moderate to strong associations with scores on all DI measures (rs = .30-.60). Additionally, NUC was moderately associated with DII (r = .32) and EI scores (r = .33), CMP was moderately associated with DII (r = .34) and EI scores (r = .33), and BRF was moderately associated with DII scores (r = .33). When considered conjointly in regression analyses, these scale scores accounted for a large amount of variance in scores on each DI measure (R²s = .39-.54). As indicated by standardized regression weights and DWs, STR was the strongest predictor of DTS scores (β = .26; DW = .088) with conditional dominance over remaining SP scales, including HLP (β = .14; DW = .036). STR also emerged as the strongest predictor of DII scores (β = .28; DW = .100) with conditional dominance over remaining SP scales, including ARX (β = .16; DW = .081), WRY (β = .16; DW = .074), NFC (β = .13; DW = .061), and HLP (β = .13; DW = .033). No SP scales demonstrated complete or conditional dominance over remaining scales in predicting EI scores, with HLP emerging as the only significant predictor of these scores (β = .14; DW = .035).

Last, at the PSY-5 level in the fourth block of Table 5, NEGE demonstrated strong associations with all DI measure scores (rs = .55-.67). In regression models, the scale also accounted for a large amount of variance in scores on each DI measure (R²s = .30-.45).

Discussion

The first goal of the current study was to examine whether conceptually indicated MMPI-3 scales were associated with a test-taker’s standing on the constructs of AS and DI. It was hypothesized that internalizing MMPI-3 scale scores would demonstrate meaningful associations with and account for meaningful variance in AS, its subdimensions, and DI, while somatic and interpersonal scale scores would also account for variance in AS and its subdimensions. Results indicated that H-O, RC, SP, and PSY-5 scales in these domains, in combination with other MMPI-3 scales from the same scale family, predicted a moderate to large amount of variance in scores on each AS and DI measure, supporting this initial hypothesis.

The second goal of the current study was to identify individual MMPI-3 scale scores that most strongly predict AS and DI when considered conjointly with other scales from the same family. Regarding AS, we hypothesized that scales such as EID, RC1, RC7, ARX, and NEGE would emerge as the strongest predictors of the global AS construct, while RC1 and NUC, RC1 and COG, and RC7 and SHY would best predict physical, cognitive, and social concerns respectively at the RC and SP levels. In support of these hypotheses, correlation and regression coefficients and DWs suggested that EID, RC1, and NEGE scores best predicted both the global AS construct and AS subdimension reflecting fear of physical consequences of anxiety at the H-O, RC, and PSY-5 levels. Contrary to hypotheses, however, WRY and ARX scores were similar in their prediction of the global AS construct at the SP level, while WRY scores emerged as the strongest among SP scales for predicting physical concerns rather than NUC. Although some of the results just described were unexpected, each of these scales are theoretically indicated in the prediction of AS. Specifically, internalizing scales such as EID and NEGE contain items capturing the fearfulness and emotional reactivity components of AS, while WRY reflects the proclivity to engage in worry in response to physical symptoms (e.g., “When my stomach is upset, I worry that I might be seriously ill” from the ASI-3; Ben-Porath & Tellegen, 2020). Scores on RC1 tap into the preoccupation with physical health concerns characteristic of individuals with elevated AS.

Next, in support of our hypotheses, the AS subdimension reflecting cognitive concerns was best predicted by EID, COG, and NEGE in their respective scale families. However, contrary to hypotheses, RCd scores emerged as the strongest predictor of cognitive concerns among RC scales rather than RC1. Although this finding also deviates from our expectations, the cognitive AS subdimension has been consistently linked to depression symptoms and suicidality in past research (Capron et al., 2015; Naragon-Gainey, 2010; Oglesby et al., 2015). As such, it conceptually follows that scales such as EID, RCd, and NEGE that index phenomena related to distressing emotional experiences (including the demoralization and dissatisfaction with life that characterize depression; Ben-Porath & Tellegen, 2020) would emerge as the strongest predictors. Additionally, individuals high in this subdimension of AS are likely to frequently experience difficulties with focus, racing thoughts, and feeling “spaced out,” while subsequently interpreting these problems as indicative of “losing one’s mind.” This focus on disrupted cognition may help explain why COG scores, which reflect similar difficulties with memory, concentration, and clear thinking (Ben-Porath & Tellegen, 2020), emerged as the strongest SP scale predictor of the construct.

Last, as hypothesized, fear of social consequences of anxiety was best predicted by EID and NEGE in the H-O and PSY-5 scale families. However, although RC7 and SHY were hypothesized as the strongest predictors of social concerns in their respective scale families, both RC1 and RC7 demonstrated comparable prediction of social concerns at the RC level, while WRY and ARX also provided comparable prediction this subdimension at the SP level. It seems likely RC1 emerged as one of the strongest RC scale predictors of the construct, as it indexes frequent experience of various somatic symptoms, including those associated with fear and anxiety (i.e., nausea, trembling hands, dizziness, agitated speech; Ben-Porath & Tellegen, 2020). The social concerns AS subdimension is operationalized as the fear that outwardly observable symptoms of fear and anxiety, such as those captured in RC1, will provoke social rejection and ridicule (Taylor et al., 2007). Moreover, the internalizing scales implicated here as the strong predictors of the social concerns AS subdimension contain items reflecting frequent experiences of anxiety and worry, social comparisons involving negative self-evaluation, and sensitivity to criticism (Ben-Porath & Tellegen, 2020). Taken together, the scales identified here appear to tap into both the somatic and social aspects of this AS subdimension.

Regarding DI, we hypothesized that the construct would be best predicted by internalizing MMPI-3 scales, including EID, RCd and RC7, STR, and NEGE in their respective scale families. Indeed, EID and NEGE scores offered the best prediction at the H-O and PSY-5 levels across DI measures. These associations make sense given their assessment of broad experiences of emotional distress, internalizing dysfunction, and negative affect (Ben-Porath & Tellegen, 2020). However, we observed discrepancies from our hypotheses in prediction across DI measures at the RC and SP scale levels. At the RC scale level, DTS scores were best predicted by RCd scores, followed by RC7 scores. In contrast, scores on the DII and EI scales were best predicted by RC7 scores, followed by RCd scores. The inconsistent evidence for the superiority of either RCd or RC7 scores in predicting DI may be due to the conceptualization of the construct. In other words, the intolerable emotional distress that characterizes DI is not specific to a particular subtype of emotional dysfunction in terms of its definition (e.g., Zvolensky et al., 2011), measurement (e.g., Simons & Gaher, 2005), or empirical correlates (e.g., its association with broad emotion dysregulation, symptoms of both anxiety and depression; Keough et al., 2010; Leyro et al., 2010; Simons & Gaher, 2005). As such, DI as a construct may exist in the interstitial space between demoralization and negative affect, plausibly explaining the emergence of both RCd and RC7 as predictors. This finding also converges with previous studies relating scores on MMPI-2-RF scales and measures of DI, which demonstrated comparable strengths of association to those observed in the current study (Kremyar et al., 2020).

At the SP scale level, STR scores most strongly predicted scores on the DTS and DII, supporting our hypotheses. However, contrary to our predictions, no SP scale score met our criteria for the single strongest predictor of the EI subscale (i.e., a moderate zero-order correlation, significant regression weight, and complete or conditional dominance over remaining scales). The emergence of STR as the strongest predictor of DTS and DII scores, but not EI scores, might be explained by differences in the operationalization of DI underlying each criterion measure. For example, the DTS defines DI in terms of distress experiences being wholly overwhelming (e.g., “My feelings of distress are so intense that they completely take over”), as well as comparisons of one’s own ability to handle distress to that of others (e.g., “Other people seem to be able to tolerate feeling distressed or upset better than I can;” Simons & Gaher, 2005). This conceptualization may relate to the stress reactivity construct underlying STR scores, which is similarly defined by an inability to handle stress and evaluations of others as less stress-reactive than the test-taker (Ben-Porath & Tellegen, 2020). In contrast, the EI scale defines DI as an inability to withstand “disturbing” emotions (e.g., “I must be free of disturbing feelings as quickly as possible; I can’t bear if they continue”), a desire for thought control (e.g., “I can’t bear to have certain thoughts”), and a perceived lack of ability to persevere through difficult experiences (e.g., “I can’t get on with my life, or be happy if things don’t change”). As such, the discrepant cognitions and affective phenomena tapped by the EI scale may have resulted in a diffuse pattern of prediction across SP scales, such that no one scale emerged as superior to others. The DII, meanwhile, consists of six DTS items, three EI items, and one ASI item, combining these related yet somewhat disparate constructs (McHugh & Otto, 2012). Thus, the emergence of STR as the strongest SP scale predictor of scores on both the DTS and DII may be attributable to the prominent item crossover and subsequent construct overlap between the two measures.

In summation, results of the current study demonstrate that MMPI-3 scale scores can be used to predict one’s standing on AS and DI, with theoretically implicated scales across MMPI-3 scale families offering the strongest prediction. Notably, we also observed unique patterns of prediction between AS and DI at the RC and SP scale levels. Specifically, scores on scales such as RC1, WRY, COG, and ARX were found to most strongly predict AS, but not DI, aligning with the conceptualization of AS as future-focused worry and fear (as captured by WRY and ARX) of anxiety- and fear-related physiological arousal (as captured by RC1 and COG). In contrast, scores on STR were found to most strongly predict DI, but not AS, converging with the definition of DI as the inability to tolerate or cope with distressing emotions.

Although replication is needed, we think there are several additional second-order inferences that could be made about an individual based on their MMPI-3 scale elevations, as supported by the strong, positive associations between MMPI-3 scale scores and AS and DI measure scores across correlational, regression, and dominance analyses. First, at the H-O and PSY-5 levels, we propose that test-takers with elevated scores (i.e., T ≥ 65) on EID and NEGE are likely to experience broad difficulties with AS and DI. More specific nuances in the experiences of AS and DI are likely captured at the RC level. Namely, high scorers on RCd are likely to be distress intolerant and may experience cognitive aspects of AS (e.g., interpret racing thoughts as being indicative of “losing one’s mind”). Other aspects of AS are likely better reflected in high scores on RC1 and RC7. High RC1 scorers may be likely to experience fear of anxiety-related physiological sensations and the social consequences of outwardly expressing fear or anxiety. High scorers on RC7, on the other hand, may experience problems with the social dimension of AS, perceive distress as intolerable, or desire escape from disturbing emotions and cognitions. Finally, the most nuanced interpretations about AS and DI are derived from SP scales. Results suggest test-takers with elevated scores on COG are likely to experience difficulties with the cognitive dimension of AS, while those with such scores on WRY and ARX may face problems with the physical and social dimensions of AS. Test-takers with elevated scores on STR are likely to view distress as intolerable.

The current study was subject to several important limitations. First, the DI measures examined in the current study contain discrepancies in terms of the specific cognitions, emotions, or behaviors seen as indicative of DI. These differences were accounted for in the current study by individually predicting scores on each of these commonly used measures. However, these inconsistencies obfuscate whether scores on these measures truly index identical constructs, reflecting a broader issue in the empirical study of DI that hinders synthesis and comparison of results across studies, as described by previous authors (i.e., Bardeen et al., 2013; McHugh & Otto, 2012). Future studies may look to not only create more unitary measures of DI (cf. McHugh & Otto, 2012) but also establish consensus as to specific features that should be considered and excluded in defining and modeling the construct. Second, in instances where multiple scales emerged as similar in their strength prediction for a given criterion, we suggested that an elevation on both scales together may be indicative of elevated standing on AS or DI (e.g., T ≥ 65 for both RC1 and RC7 being indicative of elevated standing on the social concerns AS subdimension). However, MMPI-3 scale elevations in the context of other scale elevations should receive continued empirical examination in future studies. For example, future studies might compare differences in scores on AS and DI measures across individuals who produced elevations on one, both, or neither of the indicated scales. Regarding methodological limitations, participants who identified as white and/or female were overrepresented in the current study, hindering the generalizability of these findings. Our sample was also comprised entirely of a nonclinical college student sample. Future research should replicate the methodology of the current study using a clinical sample and with a more diverse sample to determine whether these patterns of prediction hold true for these groups as well. We also relied entirely on self-report measures, which can inflate observed correlations due to shared method variance. Given the availability of behavioral measures of AS and DI (Lejuez et al., 2003; Rabian et al., 1999; Renna et al., 2018), future studies might examine the ability of MMPI-3 scales to predict performance on these measures as well.

Despite these limitations, the current study is the first to demonstrate that MMPI-3 scale scores can strongly predict a test-taker’s standing on AS, its subdimensions, and DI and to identify specific scales at the H-O, RC, SP, and PSY-5 levels that most strongly predict these constructs. These findings not only expand on the extensive literature on empirical correlates of MMPI-3 scales but also suggest additional interpretive statements that could be made about a test-taker from their pattern of MMPI-3 scale elevations. These statements, in turn, may help mental health practitioners identify individuals with elevated AS or DI using an empirically grounded broadband measure of psychological dysfunction that is appropriate for clinical practice.