Engaging in Risky and Impulsive Behaviors to Alleviate Distress Mediates Associations between Intolerance of Uncertainty and Externalizing Psychopathology

Abstract

Despite increasing recognition that intolerance of uncertainty is a transdiagnostic dimension of psychopathology, very little research has investigated its relevance for externalizing psychopathology and related risky/ impulsive behavior. Ninety-five unselected adults (ages 19-55, 53% men) recruited from the community completed a measure of intolerance of uncertainty, externalizing traits and problems, and risky/ impulsive behavior. Higher levels of intolerance of uncertainty were associated with greater endorsement of externalizing symptoms (e.g., aggression, alcohol/ marijuana use, problematic impulsivity) and last month risky and impulsive behaviors. Relations between intolerance of uncertainty and externalizing symptoms/ risky behaviors were mediated by a motivation to engage in these behaviors to avoid distress, but not by the motivation to experience pleasurable emotions. Findings suggest that difficulty tolerating uncertainty may confer risk for the externalizing spectrum of psychopathology by increasing the likelihood that an individual will engage in risky behaviors to alleviate distressing or unpleasant emotions.

Individuals differ in their reactions to uncertainty. Intolerance of uncertainty can be conceptualized as “… an individual’s dispositional incapacity to endure the aversive response triggered by the perceived absence of salient, key, or sufficient information, and sustained by the associated perception of uncertainty” (Carleton, 2016, pg. 16). The construct of intolerance of uncertainty was initially proposed and studied as a putative risk factor for generalized anxiety disorder and excessive worry (based on clinical experience, by Michel Dugas and colleagues), and there is now strong evidence to support this theoretical model (e.g., Ladouceur et al., 2000a, 2000b; Dugas et al, 2005a, 2005b). Yet, there has also been increasing interest and investigation in intolerance of uncertainty as a potential factor involved in other forms of psychopathology. Specifically, there is now evidence linking elevated levels of intolerance of uncertainty with a wide range of other mental health problems or disorders, including depression (Carleton et al., 2012; Yook et al., 2010), posttraumatic stress (Boelen, 2010; Fetzner et al., 2013), social anxiety (Boelen & Reijntjes, 2009; Carleton et al., 2010), obsessive-compulsive symptoms (Holaway et al., 2006; Tolin, Abramowitz, Brigidi, & Foa, 2003), eating pathology (Frank et al., 2012; Sternheim et al., 2011) and psychosis (Bredemeier et al., 2018; White & Gumley, 2010). In other words, there is mounting evidence that intolerance of uncertainty is a transdiagnostic dimension of psychopathology (Boswell et al., 2013; Carleton, 2016; Carleton et al., 2012; Einstein, 2014; Gentes & Ruscio, 2011; Mahoney & McEvoy, 2012; for evidence that intolerance of uncertainty is dimensional, see Carleton et al., 2012).

Surprisingly, the relationship between intolerance of uncertainty and externalizing problems/pathology (characterized by antisocial personality disorder, conduct problems, impulsivity, aggression, and/or substance misuse; see Achenbauch & Edelbrock, 1978 and Kendler et al., 1997) has yet to be directly examined. Some indirect evidence supports a link - for example, intolerance of uncertainty has been shown to be positively correlated with anger/aggression (Fracalanza et al., 2014; Gorka, Phan, Hosseini, Chen, & McCloskey, 2018) and drinking (alcohol) to cope or conform (Kraemer, McLeish, & O’Bryan, 2015; Oglesby et al., 2015). Relatedly, there is growing evidence that use of alcohol, cannabis, and nicotine dampens physiological reactivity to uncertain (but not certain) threat, suggesting that use of these substances may be reinforced through their impact on diminishing aversive reactivity to ambiguous threats (Gorka, Lieberman, Phan, & Shankman, 2016; Hefner et al., 2015; Hogle et al., 2010).

We propose that intolerance of uncertainty will increase risk for externalizing problems and risky or impulsive behavior broadly, as has now been demonstrated for various internalizing disorders, based on established links with neuroticism (Berenbaum et al., 2008) and avoidance motivation (Lee, Orsillo, Roemer, & Allen, 2010). Specifically, we posit that individuals with elevated levels of intolerance of uncertainty will be more likely to exhibit externalizing problems (e.g., aggression) and engage in risky behaviors (e.g., alcohol/substance use), in effort to reduce uncertainty or cope with unpleasant emotions due to uncertainty. Research on the affective motivations that trigger risky and impulsive behavior suggests that individuals engage in these types of behaviors for a variety of reasons, including to relieve distress and other negative mood states, such as extreme sadness, anxiety, or anger (“avoidance-based motivational triggers”) and enhance pleasurable affective states, such as joy and excitement (“approach-based motivational triggers”) (Horvath & Zuckerman, 1993; Kemp, Sadeh, & Baskin-Sommers, 2019; Leyro et al., 2010; Nock, 2010). There are of course other reasons that individuals may act rashly (e.g., social influences, cognitive deficits; Franken, van Strien, Nijs, &Muris, 2008; Strack, & Deutsch, 2004; Weitzman, Nelson, & Wechsler, 2003); however, these triggers represent two broad dimensions of affective motivation that have been linked to adverse mental health outcomes, including substance use, aggression, and other externalizing behaviors (Arnett, Smith, & Newman, 1997; Carlson, Pritchard, & Dominelli, 2013).

Distress arising from high intolerance of uncertainty could increase the propensity to engage in risky behaviors based on research showing that risk-taking can be reinforced by temporarily alleviating negative affect (e.g., Cooper, Agocha, & Sheldon, 2000; MacPherson et al., 2012). Evidence to support this proposed link between intolerance of uncertainty and externalizing problems may have important clinical implications, by identifying a potentially novel target for interventions to address externalizing and risky/impulsive behavior. Notably, interventions to reduce intolerance of uncertainty have been developed (see Dugas & Robichaud, 2012) and proven efficacious for individuals seeking treatment for generalized anxiety disorder (e.g., Ladouceur et al., 2000a; Dugas et al., 2003, 2010), social phobia (Mahoney & McEvoy, 2012, and other internalizing disorders (Boswell et al., 2013).

To provide an initial test of this proposal, the present study was designed to examine associations between self-reported levels of intolerance of uncertainty, externalizing tendencies/behavior, and (recent) risky or impulsive behaviors in an (unselected) sample of adults. Both externalizing and risky/impulsive behavior were conceptualized and measured as broad (general) constructs, consistent with the extant evidence that diverse behavioral manifestations of these constructs often co-occur and reflect latent traits (e.g., see Kreuger et al., 2002; Krueger, Markon, Patrick, Benning, & Kramer, 2007; Sadeh & Baskin-Sommers, 2017; Venables & Patrick, 2012). We hypothesized that intolerance of uncertainty would be significantly and positively correlated with both externalizing and risky/impulsive behavior. Moreover, we hypothesized that the relationship between intolerance of uncertainty and externalizing would be mediated by motivation to engage in risky/impulsive behaviors to avoid unpleasant emotions (but not motivation to experience pleasant emotions).

Methods

Participants

Participants consisted of 95 adults (ages to 19-55, M_age = 32.33, SD = 9.68) who were recruited from the general community via online postings (e.g., Craigslist) and flyers advertising a study on risky behavior. Individuals who met the following criteria were eligible to participate in the study: age 18-55, fluent in English, no history of psychosis, not imminently at risk for suicide, and no major medical problems (e.g., epilepsy, more than three head injuries with loss of consciousness greater than 30 minutes). Half of the sample identified as men (53.2%), 44.7% as women, and 2.1% as another gender. The sample was ethnically diverse, with 55.9% identifying as White, 35.5% identifying as Black or African-American, 5.4% identifying as Asian, 3.3% identifying as biracial, and 13.8% identifying as Hispanic. Over half of the sample reported a history of outpatient mental health treatment (59.6%), 20.4% a prior psychiatric hospitalization, and 50.5% involvement in the criminal justice system (e.g., incarceration, probation, parole).

Procedures

Following initial screening for eligibility over the phone, individuals were invited to participate in a lab session that involved completing a battery of questionnaires, clinical diagnostic interviews, and brief estimates of cognitive functioning. Written and verbal informed consent was obtained by research staff following a detailed description of the study. All procedures were approved by the University of Delaware Institutional Review Board.

Measures

Intolerance of Uncertainty.

The Intolerance of Uncertainty Scale – 12 (IUS-12; Carleton, Norton, & Asmundson, 2007) is a 12-item measure that indexes how people react emotionally and behaviorally to the uncertainties of life. Participants were asked to rate how characteristic each statement was of them on a 5-point scale form “Not at All Characteristic of Me” (1) to “Entirely Characteristic of Me” (5). Despite its brevity, the IUS-12 has strong evidence for reliability and construct validity (Carleton et al., 2007; Fergus & Wu, 2013; Hong & Lee, 2015; McEvoy & Mahoney, 2011). In the present sample, internal consistency for the total score was excellent (Cronbach’s alpha = .90). Scores were calculated by averaging the items for the total score (M = 2.66, SD = 0.87, Min/Max = 1.17/5.00), with higher scores reflecting greater levels of intolerance of uncertainty.

Externalizing Psychopathology.

The Externalizing Spectrum Inventory-Brief (ESI-Brief; Hall, Bernat, & Patrick, 2007) is a 100-item self-report measure that assesses a range of behavioral and personality characteristics associated with the externalizing spectrum of psychopathology on both broad- and individual-factor levels. Participants are asked to choose which option describes them best in regard to each statement: “True” (1), “Mostly True” (2), “Mostly False” (3), or “False” (4). Of the 18 subscales, examples include: physical aggression, blame externalization, boredom proneness, drug use, and empathy. Each is scored as the sum of the questions asked within that facet, with higher scores indicating greater levels of externalizing. Total score in the sample ranged from 110-336 (M = 191.54, SD = 49.20, Cronbach’s alpha = 0.96).

Risky and Impulsive Behavior.

The Risky, Impulsive, and Self-destructive behavior Questionnaire (RISQ; Sadeh & Baskin-Sommers, 2017) is a 38-item self-report questionnaire composed of a total score that measures general tendencies to engage in risky and self-destructive behaviors, and eight domain-specific factors measuring specific risky and impulsive behavioral categories. We used participants reports of how many times they engaged in the behavior in the last month. To reduce positive skewness, we categorized responses into 5 bins that constrained the range of possible responses at the high end of the distribution: 0, 1-10, 11-50, 51-100, >100 times (Sadeh & Baskin-Sommers, 2017). Participants reported engaging in a variety of harmful behaviors in the month prior to the assessment, with 47.4% reporting drug use behavior, 18.9% reporting aggressive behavior, 29.5% reporting gambling behavior, 21.1% reporting risky sexual behavior, 27.4% reporting alcohol abuse, 5.3% reporting self-harm behavior, 43.2% reporting dysregulated eating behavior, and 72.6% reporting reckless driving/spending behavior. We analyzed the RISQ last month total score (M = 6.15, SD = 5.87, Min/Max = 0/30, Cronbach’s alpha = 0.82) as an overall index of recent risk/ impulsive behavior. We used a normal score transformation (Blom) to reduce kurtosis in the total score (resulting skewness/ kurtosis = .10/ −.34).

The RISQ also provides measures of the affective context that is motivating an individual’s tendency to engage in risky or impulsive behavior. Specifically, participants were also asked to rate on a 5-point Likert-type scale (0 = “Strongly Disagree” to 4 = “Strongly Agree”) how much they agreed with the following for each behavior endorsed: “I do this behavior to stop feeling upset, distressed, or overwhelmed” and “I do this behavior to feel excitement, to get a thrill, or to feel pleasure.” These questions were designed to assess avoidance and approach motivational triggers for each behavior, respectively. Responses to these questions were averaged across items to create a Motivation to Avoid Negative Emotions score (M = 1.81, SD = 0.88, Min/Max = 0.00/3.39) and a Motivation to Approach Positive Emotions score (M = 2.11, SD = 0.76, Min/Max = 0.17/3.67).

Psychiatric Disorder Symptoms.

Lifetime symptoms of mental disorders were assessed with the Structured Clinical Interview for DSM-5 (SCID 5; First, Williams, Karg, & Spitzer, 2015). A composite measure of internalizing disorders symptoms was created by totaling the number of threshold symptoms for the most severe lifetime episode (generalized anxiety, panic, agoraphobia, social anxiety, major depressive, and bipolar disorders). Although the SCID was not designed as a dimensional scale, calculating this score allowed us to include an estimate of overall internalizing psychopathology in our analyses as a covariate (Cronbach’s alpha = .63). Lifetime prevalence rates for common mental disorders are presented in Table 1.

Table 1.

Lifetime Psychiatric Diagnoses

Psychiatric Disorder	Lifetime Diagnosis
Major Depressive Disorder	50.6%
Bipolar Disorder	7.4%
Alcohol Use Disorder	37.8%
Substance Use Disorder	45.8%
Gambling Disorder	4.9%
Panic Disorder	4.9%
Agoraphobia	3.8%
Social Anxiety Disorder	14.8%
Generalized Anxiety Disorder	10.0%
Posttraumatic Stress Disorder	26.6%
Antisocial Personality Disorder	8.5%
Borderline Personality Disorder	16.0%

Data Analysis

Analyses were conducted with the software programs SPSS version 25 (Armonk, NY: IBM Corp.) and Mplus 8.0 (Muthén & Muthén, 2017). Bivariate relations among study variables were assessed with Pearson correlations. Hierarchical linear regressions were conducted with covariates entered in the first block (age, gender, and lifetime internalizing disorder symptoms), and explanatory variables (intolerance of uncertainty total score) entered in the subsequent block. Given that internalizing and externalizing disorders are often correlated (Kotov et al., 2017), we included lifetime internalizing disorder symptoms as a covariate in analyses to test whether intolerance of uncertainty explained variance in externalizing psychopathology and risky behaviors over and above the variance externalizing disorders share with internalizing disorders. Mediation analyses were conducted using the robust maximum likelihood estimator in Mplus. The indirect effect of intolerance of uncertainty on externalizing traits and behaviors via affective motivational triggers for risky/ impulsive behavior was assessed using the ‘model indirect’ procedure.

Results

Intolerance of Uncertainty Relates to Externalizing Psychopathology & Risky Behavior

First, we examined associations between intolerance of uncertainty and externalizing psychopathology and risky/ impulsive behavior by conducting a series of hierarchical linear regressions. To contextualize the regression analyses, bivariate associations among the study variables are presented in Table 2. We included age, gender, and lifetime internalizing disorder symptoms in our models to examine whether intolerance of uncertainty accounted for variance in externalizing psychopathology and risky/ impulsive behavior over and above these variables.

Table 2.

Bivariate Correlations between the Study Variables.

Variable	1	2	3	4	5	6	7
1 Intolerance of Uncertainty (IUS-12)	---
2 Externalizing Psychopathology (ESI-Brief)	.30**	---
3 Last Month Risky/ Impulsive Behavior Total (RISQ)	.28*	.60**	---
4 Motivation to Avoid Negative Emotions (RISQ)	.41**	.48**	39**	---
5 Motivation to Approach Positive Emotions (RISQ)	.11	.45**	.38**	.41**	---
6 Internalizing Disorder Symptoms (SCID-5)	.25*	.16	.09	.23*	−.12	---
7 Age	−.17	−.04	−.21*	−.25*	−.12	−.01	---
8 Gender	−.18	.09	.01	−.12	.23*	−.20	.09

Note. N = 95. IUS-12 = Intolerance of Uncertainty Scale – 12. ESI-BF = Externalizing Spectrum Inventory-Brief. RISQ = Risky, Impulsive, and Self-destructive behavior Questionnaire. SCID-5 = Structured Clinical Diagnostic Interview for DSM-5.

^**p < .01,

^*p < .05.

For our first model, we conducted a linear regression with the ESI-Brief total score as the dependent variable to examine overall externalizing psychopathology. Age, and gender were not associated with externalizing psychopathology (|βs| < .17, ps >.11). Lifetime internalizing symptoms was positively associated with the ESI-Brief total score (β = .25, SE = .67, p =.021). However, the addition of IUS-12 total score in the second block of the regression explained significant variance in externalizing traits (β = .29, SE = 5.91, p =.007, ΔR² = .08), such that greater levels of intolerance of uncertainty were associated with higher levels of externalizing psychopathology. To follow-up on this association, we examined bivariate correlations between the IUS-12 total score and the ESI-Brief subscales. These associations are presented in Table 3 for descriptive purposes in order to illustrate the nature of the association between intolerance of uncertainty and features of externalizing psychopathology.

Table 3.

Intolerance of Uncertainty Associations with Externalizing Psychopathology.

Externalizing Spectrum Inventory-Brief	Intolerance of Uncertainty Scale-12 Total Score
Alcohol Problems	.24*
Drug Use	.09
Marijuana Problems	.22*
Marijuana Use	.12
Physical Aggression	.28*
Destructive Aggression	.22*
Relational Aggression	.25*
Impulsive Urgency	.26*
Problematic Impulsivity	.29*
Excitement Seeking	.12
Boredom Proneness	.30**
Blame Externalization	.40**
Irresponsible	.18
Rebelliousness	.18
Lacks Empathy	.26*
Lacks Honesty	.07
Theft	.03
Fraud	.17

Note. N = 95. ESI-BF = Externalizing Spectrum Inventory-Brief. IUS-12 = Intolerance of Uncertainty Scale – 12.

^**p < .01,

^*p < .05.

Given that the ESI-Brief indexes trait-like tendencies toward externalizing, we next tested whether intolerance of uncertainty also related to recent engagement in risky and impulsive behaviors that may or may not be stable over time. Linear modeling followed the same sequence as above, with the exception that RISQ last month total score was entered as the dependent variable. Gender and internalizing symptoms were unrelated to recent risky/ impulsive behaviors (|βs| < .11, ps >.31), and age was inversely associated (β = −.22, SE = .01, p =.041). Paralleling the findings for externalizing psychopathology, IUS-12 total score was positively related to engagement in recent risky/ impulsive behaviors (β = .24, SE = .12, p =.028, ΔR² = .05).

Mediation Analyses with Affective Motivational Triggers

Externalizing Psychopathology.

Based on the foregoing, we tested whether motivational triggers for risky and impulsive behavior mediated the relationship between intolerance of uncertainty and externalizing psychopathology. Path models examining these associations are displayed in Figure 1. First, we tested the hypothesis that the relationship between intolerance of uncertainty and externalizing psychopathology is mediated by the tendency for individuals who score high on intolerance of uncertainty to engage in risky or impulsive behavior to avoid negative emotions. The direct effects of intolerance of uncertainty on avoidance motivation (β = .41, SE = .08, p <.001), and from RISQ avoidance motivation to externalizing psychopathology (β = .43, SE = .10, p <.001) were both significant; the direct path from intolerance of uncertainty to externalizing psychopathology became non-significant with avoidance motivation in the model (β = .13, SE = .09, p =.161). Notably, the indirect effect of intolerance of uncertainty on externalizing psychopathology via avoidance motivation was significant (standardized β = .17, SE = .06, p = .002), indicating that the tendency for individuals to engage in risky/ impulsive behaviors to avoid negative emotions mediated the effect of intolerance of uncertainty on externalizing psychopathology. The model explained 24.5% of the variance in externalizing psychopathology (p = .015).

Figure 1. Motivation to Avoid Negative Emotions Mediates Intolerance of Uncertainty Association with Externalizing Psychopathology.

A) Depicts mediation model with the RISQ Avoidance Motivation scale entered as a mediator of the association between intolerance of uncertainty and externalizing psychopathology. B) Depicts mediation model with the RISQ Approach Motivation scale entered as a mediator of the association between intolerance of uncertainty and externalizing psychopathology. Solid paths: p<.05.

Next, we tested whether intolerance of uncertainty influences externalizing psychopathology indirectly via the tendency for individuals to engage in risky/ impulsive behavior to approach positive emotions. In this model, the association between intolerance of uncertainty and approach motivation was not significant (β = .11, SE = .11, p =.324), although approach motivation was positively associated with externalizing psychopathology (β = .40, SE = .08, p <.001). The direct path from intolerance of uncertainty to externalizing psychopathology remained significant with RISQ approach motivation in the model (β = .26, SE = .09, p =.003), and the indirect effect of intolerance of uncertainty on externalizing psychopathology was non-significant (standardized β = .04, SE = .05, p =.336), indicating approach motivation did not explain the association between intolerance of uncertainty and externalizing psychopathology. The model explained 24.8% of the variance in externalizing psychopathology (p = .003).

Recent Risky and Impulsive Behavior.

A parallel mediation analysis conducted with last month risky and impulsive behaviors as the outcome variable and RISQ avoidance motivation entered as the mediator produced converging results to those obtained for externalizing psychopathology. The direct effect from RISQ avoidance motivation to last month risky/ impulsive behavior was significant (β = .32, SE = .11, p =.003), and the direct path from intolerance of uncertainty to last month risky/ impulsive behavior became non-significant with avoidance motivation in the model (β = .15, SE = .14, p =.148). The indirect effect of intolerance of uncertainty on risky/ impulsive behavior via avoidance motivation was also significant (standardized β = .13, SE = .05, p = .014). The model explained 16.2% of the variance in last month risky/ impulsive behavior (p = .041).

We next tested a similar model with RISQ approach motivation entered as the mediator. Consistent with the above findings for externalizing psychopathology, the direct effects from RISQ approach motivation to last month risky/ impulsive behavior was significant (β = .35, SE = .09, p <.001). The direct path from intolerance of uncertainty to last month risky/ impulsive behavior remained significant with approach motivation in the model (β = .24, SE = .10, p =.010), and the indirect effect of intolerance of uncertainty on risky/ impulsive behavior via approach motivation was not significant (standardized β = .04, SE = .04, p = .336). The model explained 20.0% of the variance in last month risky/ impulsive behavior (p = .002).

Discussion

Understanding the psychological processes that confer risk for externalizing behaviors may offer new intervention targets for these difficult-to-treat behaviors. The present study was designed to provide an initial test of the hypothesis that intolerance of uncertainty is a psychological process that is relevant for explaining symptoms of externalizing psychopathology and engagement in other risky and impulsive behavior. Results suggest that higher levels of intolerance of uncertainty predict greater reports of externalizing symptoms, as well as recent risky and impulsive behavior more broadly. Further, the tendency to use risky behaviors to avoid experiencing negative emotions, but not positive emotions, mediated the association between intolerance of uncertainty and externalizing psychopathology. These findings are consistent with previous research linking intolerance of uncertainty with the use of emotional avoidance coping behaviors (e.g., Lee et al., 2010), and extends this work by, for the first time, demonstrating that difficulty tolerating uncertainty may confer risk for the externalizing spectrum of psychopathology by increasing the likelihood that an individual will engage in risky behaviors to alleviate distressing or unpleasant emotions.

Interest in the transdiagnostic relevance of intolerance of uncertainty has grown, and empirical evidence now links intolerance of uncertainty to diverse forms of psychopathology, including posttraumatic stress (Boelen, 2010, Fetzner et al., 2013), obsessive-compulsive (Holaway et al., 2006, Tolin et al., 2003), eating (Frank et al., 2012, Sternheim et al., 2011, and psychotic disorders (Bredemeier et al., 2018; White & Gumley, 2010). Extending this literature, our findings provide new evidence that intolerance of uncertainty is a psychological process that may also confer risk for externalizing psychopathology, such as alcohol and substance use disorders, and antisocial personality disorder. This association, though relatively novel, can be understood in the context of research showing that externalizing psychopathology is marked by high levels of negative emotionality (Venables & Patrick, 2012), a characteristic that is common across psychological disorders associated with intolerance of uncertainty (e.g., Weinstock & Whisman, 2006). Unlike the majority of the disorders previously associated with intolerance of uncertainty, however, the externalizing spectrum is also marked by elevations on measures of disconstraint (Venables & Patrick, 2012), which may influence the ways in which externalizing individuals react to perceptions of uncertainty.

The cross-diagnostic relevance of intolerance of uncertainty makes sense given that theories focus on the situations and events that trigger negative reactions, but not how an individual will cope with perceptions of uncertainty and the distressing feelings they trigger (Carleton, 2016). Some individuals may rely on cognitive strategies (e.g., worry, rumination; Buhr & Dugas, 2009; Gentes & Ruscio, 2011; Ladouceur et al., 2000b) to manage aversive feelings associated with uncertainty, whereas others might turn to maladaptive behavioral strategies (e.g., checking behaviors, alcohol use; Gorka et al., 2016; Kraemer et al., 2015; Tolin et al., 2003) to cope with these experiences. Thus, the particular strategies employed by an individual to endure uncertainty will likely determine how intolerance of uncertainty manifests and its cognitive, behavioral, and psychopathological correlates. The present findings extend previous work on this topic by identifying novel cognitive (e.g., blame externalization) and behavioral (e.g., aggression, impulsive risk-taking) strategies associated with externalizing psychopathology that people may use to cope with uncertainty. Consistent with these findings, a recent study by Gorka and colleagues (2018) reported a positive association between intolerance of uncertainty and trait aggression in a sample of adults with intermittent explosive disorder, and other research has identified links between intolerance of uncertainty and the use of alcohol as a coping mechanism (Gorka et al., 2016; Kraemer et al., 2015; Oglesby, Albanese, Chavarria, & Schmidt, 2015). In combination, these studies provide growing evidence that some individuals high on intolerance of uncertainty cope with the discomfort that accompanies uncertainty by engaging in impulsive, externalizing behaviors.

More specifically, present results suggest that the tendency to use risky and impulsive behaviors to alleviate unpleasant and distressing emotions may explain the observed association between intolerance of uncertainty and externalizing psychopathology. Similar to research on internalizing disorders that has identified neuroticism (Berenbaum et al., 2008) and avoidance motivation (Lee et al., 2010) as psychological processes linked to intolerance of uncertainty, the current findings highlight the role that emotion-based coping behaviors play in understanding intolerance of uncertainty and its associations with psychopathology. Although we found that externalizing psychopathology was also associated with the tendency to engage in risky and impulsive behaviors to experience pleasurable emotions, this type of affective trigger was unrelated to intolerance of uncertainty in this study, suggesting specificity in the valence of the emotional experience that is relevant to understanding intolerance of uncertainty-externalizing associations. However, the cross-sectional nature of the data prohibits making conclusions about the direction of the associations between intolerance of uncertainty and externalizing disorders. Longitudinal data on temporal changes in perceptions of uncertainty, emotional arousal, and emotional valence would be valuable for determining how intolerance of uncertainty impacts emotional responding in externalizing individuals and for identifying factors that differentiate intolerance of uncertainty relations with internalizing vs. externalizing disorders.

In addition to motivation to avoid unpleasant emotions, another factor that we hypothesize may help explain the association between intolerance of uncertainty and externalizing psychopathology is information processing biases. Specifically, some behavioral evidence suggests that intolerance of uncertainty shapes processing of ambiguous information (Dugas, et al., 2005a), and there is a large body of research showing a link between certain facets of externalizing and judgement biases (e.g., hostile attribution biases and aggression; De Castro et al., 2002). Behavioral measures of information processing biases were not administered in the present study, but could be examined as another potential mediator between intolerance of uncertainty and externalizing in future research.

Yet another potentially important factor in the link between intolerance of uncertainty and externalizing that was not explored in the current study is impaired decision-making. Unlike most research on intolerance of uncertainty that suggests individuals high on this trait are particularly cautious when making decisions and desire more, rather than less, information before rendering a judgment (Ladouceur, Talbot, & Dugas, 1997), externalizing psychopathology is characterized by poor impulse control and rash decision-making (Patrick et al., 2013). Although these results are seemingly contradictory, our finding that intolerance of uncertainty is positively associated with risky and impulsive behaviors converges with the idea that delaying decision-making engenders more uncertainty about the future than taking immediate action. For example, a study by Luhmann et al. (2011) found that intolerance of uncertainty was positively related to shorter wait times before making decisions and more frequent selection of immediate, riskier rewards in a sample of 50 undergraduate students. The authors interpreted these findings as evidence that, if delaying uncertainty increases distress among individuals high on intolerance of uncertainty, then they will be more likely to make riskier, immediate decisions to alleviate unpleasant emotions. The tendency to take immediate action may be more pronounced among individuals with a desire for certainty combined with other individual difference factors that increase impulsivity, such as an approach motivation (Elliot & Thrash, 2002) or deficits in executive functioning (Nigg, 2017). Risky/ impulsive behavior may also be heightened in individuals who are intolerant of uncertainty in situations where there is a significant disparity between an individual’s present state (e.g., anxiety related to ambiguity) and his/her desired state (e.g., feeling certain about the future), and lower-risk options are unlikely to meet this need to reduce uncertainty (Mishra & Lalumiere, 2010). For example, individuals high on intolerance of uncertainty may act aggressively in situations where not doing so may lead to serious harm (e.g., when being provoked into a fight with a weapon), and the effectiveness of less-risky options for reducing the potential harm (e.g., resolving the conflict with words) is uncertain. This theory is based on literature showing that individuals tend to shift from risk-aversion to risk-proneness decision-making in situations where low-risk options are unlikely to meet the individual’s needs (Mishra & Lalumiere, 2010). The extent to which decision-making and motivational tendencies moderate whether intolerance of uncertainty relates to impulsive, externalizing behaviors or a preference for cautious action requires further study, although extant research suggests emotional avoidance may play a role in either case.

This study has several strengths including recruitment of a diverse, unselected sample of adults from the general community, examination of an understudied research question, and findings with novel research and clinical implications. As with any study, there are also limitations to the study design. First, the cross-sectional nature of the data prohibits drawing strong conclusions about the direction of the relationships between intolerance of uncertainty, externalizing psychopathology, and emotional triggers for risky/ impulsive behavior. The mediation models were based on theoretical relationships among the variables, and longitudinal studies are needed to verify the proposed temporal ordering. Second, this represents an initial investigation into the relationship of intolerance of uncertainty with externalizing psychopathology, and thus, replication is necessary to establish the reliability of the findings. Third, the present study relied exclusively on self-report measures — future studies examining this relationship should incorporate behavioral measures. Fourth, future research should test whether the observed links between intolerance of uncertainty and externalizing behaviors could be accounted for by related constructs, including fear of the unknown (Carleton, 2016) and distress tolerance. Current findings are consistent with research conducted on the related, but broader, construct of distress tolerance that show elevated levels of distress intolerance are related to externalizing psychopathology (Cummings et al., 2013). Intolerance of uncertainty has been theorized to represent a domain of distress tolerance (Zvolensky, Vujanovic, Bernstein, & Leyro, 2010) and further study is needed to determine whether our mediation models replicate using measures of distress tolerance, and whether intolerance of uncertainty explains unique variance over and above the distress tolerance construct. Finally, although there was strong representation of externalizing in this sample, examining the generalizability of the findings in a sample recruited for their externalizing traits is an important next step.

Despite these limitations, this study contributes new evidence that difficulty tolerating uncertainty may confer risk for the externalizing spectrum of psychopathology by increasing the likelihood that an individual will engage in risky behaviors to alleviate distressing or unpleasant emotions and highlights a novel mechanism that may promote externalizing behaviors in individuals high on intolerance of uncertainty. Results provide a basis for future research to explore intolerance of uncertainty as an understudied, yet potentially influential, psychological mechanism for understanding externalizing psychopathology and related risky/ impulsive behavior.