Introduction
Worry is a chain of thoughts reflecting anxious apprehension about the future, which serves to prepare for and mitigate possible negative outcomes (Borkovec, Reference Borkovec, Davey and Tallis1994, Reference Borkovec2002). Although worry can help people anticipate and minimise potential threats (Tallis & Eysenck, Reference Tallis and Eysenck1994), the continuous activation of worry and the associated arousal and vigilance towards threat can have negative implications for physical and mental health (Brosschot, Gerin, & Thayer, Reference Brosschot, Gerin and Thayer2006). Chronic, excessive worry is a transdiagnostic form of repetitive negative thinking (McEvoy, Watson, Watkins, & Nathan, Reference McEvoy, Watson, Watkins and Nathan2013) and is the cardinal feature of generalised anxiety disorder (GAD; American Psychiatric Association (APA), 2013). Leading theoretical models of chronic worry, especially as expressed in GAD, have, in large part, emphasised the role of cognitive processes (e.g., unhelpful beliefs, cognitive biases, and difficulties with cognitive control) in maintaining worry (e.g., Borkovec, Alcaine, & Behar, Reference Borkovec, Alcaine, Behar, Heimberg, Turk and Mennin2004; Dugas, Gagnon, Ladouceur, & Freeston, Reference Dugas, Gagnon, Ladouceur and Freeston1998; Hirsch & Mathews, Reference Hirsch and Mathews2012; Wells, Reference Wells1995). However, there has been a growing recognition of the need to refine our understanding of the full range of behaviours associated with chronic worry, to better inform diagnosis, and cognitive–behavioural conceptualisation, assessment, and treatment (Andrews et al., Reference Andrews, Hobbs, Borkovec, Beesdo, Craske, Heimberg and Stanley2010; Mahoney et al., Reference Mahoney, Hobbs, Newby, Williams, Sunderland and Andrews2016). Notably, there was a proposal to include behavioural criteria in The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) definition of GAD (Andrews et al., Reference Andrews, Hobbs, Borkovec, Beesdo, Craske, Heimberg and Stanley2010), but these ultimately were not included.
Classical cognitive–behavioural models have conventionally characterised individuals high in worry (and anxiety) as cautious and harm- and risk-avoidant. However, there is growing evidence that impulsivity is also part of the behavioural repertoire. The objective of this study was to investigate, in individuals reporting chronic worry, behaviours associated with one form of impulsivity — negative urgency (NU), the tendency to make rash decisions when experiencing negative arousal (Cyders & Smith, Reference Cyders and Smith2008).
In the small number of studies that have considered impulsivity, some individuals high in chronic worry endorsed (e.g., on a questionnaire) taking rash and impulsive action to downregulate intense distress (e.g., Cougle, Timpano, & Goetz, Reference Cougle, Timpano and Goetz2012; Malivoire et al., Reference Malivoire, Stewart, Tallon, Ovanessian, Pawluk and Koerner2019b; Pawluk & Koerner, Reference Pawluk and Koerner2013, Reference Pawluk and Koerner2016; Salters-Pedneault, Roemer, Tull, Rucker, & Mennin, Reference Salters-Pedneault, Roemer, Tull, Rucker and Mennin2006). However, until now, there has been no known study undertaken to examine what this rash action ‘looks’ like for individuals high in chronic worry and whether the types of behaviours differ from those reported by individuals low in chronic worry. Understanding the manifestations of NU is important to situate NU within models of chronic worry and to effectively identify and target these behaviours in treatment.
Impulsivity and Worry
It is unsurprising that research on worry behaviours has focused on cautious and ‘safe’ behaviours. Historically, anxiety (which is typically accompanied by worry) has been associated with inhibitory responses, including freezing and paralysis in response to anxiety-provoking situations (Grillon et al., Reference Grillon, Robinson, O'connell, Davis, Alvarez, Pine and Ernst2017; Hagenaars, Oitzl, & Roelofs, Reference Hagenaars, Oitzl and Roelofs2014; Sagliano, Cappuccio, Trojano, & Conson, Reference Sagliano, Cappuccio, Trojano and Conson2014). Furthermore, the tendency to consistently withdraw from, avoid, and be vigilant towards novel situations is proposed to be a predictor of anxiety psychopathology (Lahat, Hong, & Fox, Reference Lahat, Hong and Fox2011; Svihra & Katzman, Reference Svihra and Katzman2004). In general, anxious people are motivated to prioritise safe environments that increase their sense of security as opposed to approaching new and unfamiliar stimuli that have the potential to involve threat and further heighten uncertainty (Dickson & MacLeod, Reference Dickson and MacLeod2004; Grupe & Nitschke, Reference Grupe and Nitschke2013; Woody & Rachman, Reference Woody and Rachman1994). Consistently, chronic worry has been conceptualised as an avoidance strategy to stave off hypothetical negative events. That is, individuals with chronic worry repeatedly focus on feared negative events and possible negative outcomes in an attempt to prevent bad events from happening or to lessen the negative impact of these events (Borkovec et al., Reference Borkovec, Alcaine, Behar, Heimberg, Turk and Mennin2004). Unsurprisingly, worry is also associated with cautious and harm-avoidant behaviours. For example, individuals high in chronic worry avoid uncertain situations out of concern that these will lead to bad outcomes (Dugas et al., Reference Dugas, Gagnon, Ladouceur and Freeston1998), or they may also experience feelings of paralysis when faced with uncertainty (Birell et al., Reference Birrell, Meares, Wilkinson and Freeston2011). Chronic worry is also associated with behaviours that reflect a tendency to be deliberate, such as checking, overplanning, controlling others’ behaviour (Mahoney et al., Reference Mahoney, Hobbs, Newby, Williams, Sunderland and Andrews2016; Mahoney, Hobbs, Newby, Williams, & Andrews, Reference Mahoney, Hobbs, Newby, Williams and Andrews2018), and overthinking in situations where a faster response would be more advantageous (e.g., taking too long to act on a potential opportunity at the risk of losing that opportunity; Pawluk & Koerner, Reference Pawluk and Koerner2013).
The first known data on the potential relationship between impulsivity and chronic worry came from studies in which impulsivity was measured, but was not the central focus of the research. For example, in a test of the emotion dysregulation model of chronic worry (Mennin et al., Reference Mennin, Heimberg, Turk and Fresco2005), individuals high in GAD symptoms endorsed a range of maladaptive coping strategies when feeling highly distressed, among them being difficulties with impulse control (Salters-Pedneault et al., Reference Salters-Pedneault, Roemer, Tull, Rucker and Mennin2006). In a study on problem-solving in individuals high in chronic worry, a correlation was observed between a self-reported habitual impulsive/careless problem-solving style and greater intensity of worry (Belzer, D'Zurilla, & Maydeu-Olivares, Reference Belzer, D'Zurilla and Maydeu-Olivares2002). In a study of correlates of interpersonal problem-solving effectiveness in people with GAD, Pawluk, Koerner, Tallon, and Antony (Reference Pawluk, Koerner, Tallon and Antony2017) found that an impulsive or careless problem-solving style was among the strongest correlates of lower-quality solutions. Together, these findings suggest that people high in chronic worry are motivated to alleviate distress quickly, which may engender impulsive decision-making and rash action. However, no known theoretical model of chronic worry or GAD elaborates on impulsivity. Furthermore, the aforementioned studies linking impulsivity to chronic worry used relatively crude measures of impulsivity consisting of broadly worded items (e.g., ‘When I'm upset, I feel out of control’). It is only recently that researchers have considered more expressly the relationship between impulsivity and chronic worry and anxiety. Much of this research focuses on one form of impulsivity, NU.
NU and Worry
NU is a construct that emerged from an analysis of the five-factor model of personality in an effort to parse the broader construct of impulsivity into its constituent dimensions (Whiteside & Lynam, Reference Whiteside and Lynam2001). NU is measured using one scale of the Urgency, (lack of) Premeditation, (lack of) Perseverance, Sensation Seeking Impulsive Behavior Scale (UPPS-P; Lynam, Whiteside, Smith, & Cyders, Reference Lynam, Whiteside, Smith and Cyders2006), a self-report measure of five forms of impulsivity. NU is suggested to be a construct similar to the impulsiveness dimension of neuroticism and low conscientiousness (e.g., lack of planning and consideration and low self-discipline; Whiteside & Lynam, Reference Whiteside and Lynam2001). Cyders and Smith (Reference Cyders and Smith2008) developed a model of NU to elucidate the relationship between the experience of negative affect and rash and risky decision-making. They proposed that the experience of intense negative emotion is likely to be coupled with pronounced needs, which can precipitate uncharacteristic and extreme behaviours that focus on addressing immediate concerns but are unhelpful in the long term. According to the theory, experiencing intense emotions reduces cognitive resources and impairs rational decision-making, increasing the likelihood of risky or imprudent behaviour. Perhaps unsurprisingly, NU has been predominately considered within the context of externalising problems classically associated with risky behaviour, including borderline personality disorder (BPD), substance abuse, and eating disorders (Anestis, Smith, Fink, & Joiner, Reference Anestis, Smith, Fink and Joiner2009; Bardone-Cone, Butler, Balk, & Koller, Reference Bardone-Cone, Butler, Balk and Koller2016; Gagnon, Daelman, & McDuff, Reference Gagnon, Daelman and McDuff2013; Kaiser, Milich, Lynam, & Charnigo, Reference Kaiser, Milich, Lynam and Charnigo2012; Peters, Upton, & Baer, Reference Peters, Upton and Baer2013; Smith & Cyders, Reference Smith and Cyders2016). However, the NU model has emerging applications to anxiety and depression psychopathology (Peters, Bowen, & Balbuena, Reference Peters, Bowen and Balbuena2019).
To date, there is only a small body of research that has expressly considered the association between NU and anxiety psychopathology in particular, with much of it focusing on chronic worry and the potential implications for GAD, given that worry is a central feature of this disorder. NU has been found to be associated with greater worry in non-clinical samples (e.g., Cougle et al., Reference Cougle, Timpano and Goetz2012; Gay, Schmidt, & Van der Linden, Reference Gay, Schmidt and Van der Linden2011; Macatee et al., Reference Macatee, Allan, Gajewska, Norr, Raines, Albanese and Cougle2016) and in samples high in GAD symptoms (Malivoire et al., Reference Malivoire, Stewart, Tallon, Ovanessian, Pawluk and Koerner2019b; Pawluk & Koerner, Reference Pawluk and Koerner2013, Reference Pawluk and Koerner2016; Salters-Pedneault et al., Reference Salters-Pedneault, Roemer, Tull, Rucker and Mennin2006). Notably, relative to other forms of impulsivity, such as sensation-seeking, lack of premeditation, and a tendency to give up prematurely, NU has shown the strongest positive association with GAD symptoms (Pawluk & Koerner, Reference Pawluk and Koerner2013), indicating that impulsivity expresses itself in a particular way in individuals high in chronic worry and anxiety.
The relationship between NU and cognitive processes has received some research attention, to gain an understanding of how NU might fit into existing conceptualisations of chronic worry and GAD, which heavily emphasise cognitive constructs. Greater NU has been shown to be associated with greater intolerance of uncertainty (IU), a set of negative beliefs about uncertainty and its implications. IU partially accounted for the relationship between GAD symptoms and NU (Pawluk & Koerner, Reference Pawluk and Koerner2016), suggesting tentatively that urgent action may be directed not only at reducing intense distress but also intense feelings of uncertainty. Furthermore, in individuals high in chronic worry, those who endorse greater NU also endorse a heightened tendency to appraise scenarios that are ambiguous, or even overtly positive/safe, as threatening. In this instance, both the urgent, rash action and inflated threat perceptions may be explained by underlying cognitive dyscontrol (Malivoire et al., Reference Malivoire, Marcos, Pawluk, Tallon, Kusec and Koerner2019a); however, it is worth noting that at least one study in a community sample showed that the relationship between NU and GAD symptoms was not stronger for those with poorer self-reported attentional control (Malivoire et al., Reference Malivoire, Stewart, Tallon, Ovanessian, Pawluk and Koerner2019b). More research is needed on the relationship between cognitive control and NU in persons high in chronic worry and anxiety.
Although there is growing evidence that individuals higher in chronic worry and GAD symptoms endorse higher NU, the question remains, what does NU ‘look like’ in people who worry uncontrollably? In other words, what are the actual behaviours that constitute NU? Behaviours associated with NU have tended to be examined in externalising psychopathology characterised by impulsivity, including addictions, eating disorders, and BPD (Anestis et al., Reference Anestis, Smith, Fink and Joiner2009; Bardone-Cone et al., Reference Bardone-Cone, Butler, Balk and Koller2016; Gagnon et al., Reference Gagnon, Daelman and McDuff2013; Kaiser et al., Reference Kaiser, Milich, Lynam and Charnigo2012; Peters et al., Reference Peters, Upton and Baer2013; Smith & Cyders, Reference Smith and Cyders2016). The model of NU predicts that rash behaviours include substance abuse, binge eating and purging, pathological gambling, risky sex, and compulsive shopping (Anestis et al., Reference Anestis, Smith, Fink and Joiner2009; Cyders & Smith, Reference Cyders and Smith2008; Deckman & DeWall, Reference Deckman and DeWall2011; Settles, Cyders, & Smith, Reference Settles, Cyders and Smith2010), and there is evidence supporting that NU, compared to other types of impulsivity, is a common risk factor for these behaviours (Dir, Karyadi, & Cyders, Reference Dir, Karyadi and Cyders2013). Furthermore, there is evidence that NU is a dispositional contributor to problematic interpersonal behaviour, including excessive reassurance-seeking (Anestis, Selby, & Joiner, Reference Anestis, Selby and Joiner2007), and interpersonal aggression, including physical (e.g., punching, pushing, and slapping others; starting fights) and verbal (e.g., swearing at or insulting others) aggression (Dvorak, Pearson, & Kuvaas, Reference Dvorak, Pearson and Kuvaas2013; Miller, Flory, Lynam, & Leukefeld, Reference Miller, Flory, Lynam and Leukefeld2003; Settles et al., Reference Settles, Fischer, Cyders, Combs, Gunn and Smith2012). It remains unknown how individuals high in chronic worry are interpreting the items that comprise NU on the UPPS-P and whether the behaviours they engage in are consistent with those reported in the model of NU. For example, one of the NU items states ‘When I feel bad, I will often do things I later regret in order to make myself feel better now’; however, when individuals high in chronic worry endorse agreement with this item, it is unknown what types of things they do to dampen the negative affect.
As noted earlier, current cognitive–behavioural models of chronic worry (and anxiety pathology more broadly) do not expressly feature NU. It is important to explore the behaviours associated with NU in chronic worry, as this could have important implications for assessment and treatment. Clinicians may not think to ask about impulsive and rash action, and individuals high in chronic worry may not think of themselves as ‘impulsive’. Therefore, it is important to understand what NU in chronic worry may look like, in order to help clinicians better target potentially maladaptive behaviours in the treatment of GAD, and other conditions in which unconstructive worry may be a feature. As such, these findings may have implications not only for GAD but also for other psychopathologies featuring chronic unconstructive worry. Furthermore, it is important to compare these behaviours to those of individuals low in worry to establish whether these behaviours are unique or qualitatively specific to people high in chronic worry. With this in mind, we investigated the types of rash behaviours that individuals high and low in worry engage in when distressed.
Method
Participants
Participants were recruited as part of a larger study examining the relationship between pathological worry, cognitive processes, and NU (see Malivoire, Stewart, et al., Reference Malivoire, Stewart, Tallon, Ovanessian, Pawluk and Koerner2019b). Participants were recruited through Amazon Mechanical Turk (MTurk), an online platform that has been demonstrated to be a valid method of data collection for social and clinical psychology research (e.g., Hauser & Schwarz, Reference Hauser and Schwarz2016; Shapiro, Chandler, & Mueller, Reference Shapiro, Chandler and Mueller2013). Eligibility criteria included: (1) being at least 18 years of age and (2) residing in Canada or the United States. Three-hundred and twenty-five participants were recruited. To ensure that the data were of adequate quality, participants had to have a track record of providing high-quality data for other MTurk studies, as evidenced by an approval rating of at least 97%. Approval rating on MTurk has been shown to be associated with the quality of data (e.g., Peer, Vosgerau, & Acquisti, Reference Peer, Vosgerau and Acquisti2014). Furthermore, we asked participants to indicate at the end of the survey whether their data were valid or whether we should discard their data (with no penalty to their compensation). In addition, participants who completed the survey in under 15 min were excluded, given that the average time to complete was 32.4 min (SD = 12.8 min). Participants were also required to answer a CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) to prevent non-human activity. Furthermore, participants whose reported date of birth did not match their reported age were excluded. In total, 51 participants were excluded due to data being deemed low or of unreliable quality according to these criteria.
Of the remaining 274 participants, 70 were considered to be experiencing chronic worry (26.6%), as evidenced by scores greater than 62 on the Penn State Worry Questionnaire (PSWQ; Behar, Alcaine, Zuellig, & Borkovec, Reference Behar, Alcaine, Zuellig and Borkovec2003; Meyer, Miller, Metzger, & Borkovec, Reference Meyer, Miller, Metzger and Borkovec1990). The participants high in worry were found to have a mean GAD Questionnaire for the DSM-IV (GAD-Q-IV) score of 9.1 (SD = 3.1), which is above the cut-off of 9.0 that is suggested to identify individuals with GAD with high sensitivity and specificity (Newman et al., Reference Newman, Zuellig, Kachin, Constantino, Przeworski, Erickson and Cashman-McGrath2002). Furthermore, the mean PSWQ (M = 71.9, SD = 5.4) score suggested a level of chronic worry that is the characteristic of anxiety psychopathology, especially GAD (Behar et al., Reference Behar, Alcaine, Zuellig and Borkovec2003; Fresco, Mennin, Heimberg, & Turk, Reference Fresco, Mennin, Heimberg and Turk2003). Of these 70 participants high in worry, 61 (87.1%) endorsed one of the NU items on the UPPS-P (Lynam et al., Reference Lynam, Whiteside, Smith and Cyders2006) chosen for follow-up analysis and were subsequently asked to answer open-ended questions about their specific experiences of NU. Of the 61 participants asked to answer open-ended questions, nine participants chose not to answer the open-ended questions. Following data exclusions for irrelevant data (see below), responses from 48 participants high in worry were ultimately coded.
Furthermore, 110 participants (40.1%) were considered to be experiencing low worry using the PSWQ cut-off score of 43 or below, which is a cut-off score that identified low worriers in a normative sample (Gillis, Haaga, & Ford, Reference Gillis, Haaga and Ford1995) and has been used in past research (e.g., Hayes, Hirsch, & Mathews, Reference Hayes, Hirsch and Mathews2008; Leigh & Hirsch, Reference Leigh and Hirsch2011). The participants low in worry were found to have a mean GAD-Q-IV score of 1.3 (SD = 1.3) and a mean PSWQ score of 30.6 (SD = 7.8). Of these 110 participants low in chronic worry, 73 (66.4%) endorsed at least one of the NU items of interest on the UPPS-P that was associated with a follow-up open-ended question. Of the 73 participants asked to answer open-ended questions, 24 participants chose not to answer the open-ended questions. Following data exclusions for irrelevant data (see below), responses from 45 participants low in worry were ultimately coded.
Of the individuals identified as being either high or low in worry (N = 180), 57 were male (31.7%), 122 were female (67.8%), and one identified as ‘other’ (0.6%). The mean age of the sample was 41.8 years (SD = 13.4), with a range of 19–75 years. Twenty-one individuals reported that they were currently enrolled in an education programme (11.7%). Almost all of the sample reported having obtained at least a high school diploma (99.4%), and 48.7% reported having an undergraduate degree or higher. Seventy-four percent of the sample reported working part- or full-time. There was a significant difference in age (p = .005) by the worry group, with low worriers having a slightly older mean age (M = 44.0, SD = 14.3) than high worriers (M = 38.3, SD = 11.3). There were no differences between groups on any other demographic variable.
Measures
The PSWQ (Meyer et al., Reference Meyer, Miller, Metzger and Borkovec1990) is a 16-item self-report measure that assesses pathological worry. The PSWQ has high internal consistency (a = .86–.95; Dear et al., Reference Dear, Titov, Sunderland, Mcmillan, Anderson, Lorian and Robinson2011; Molina & Borkovec, Reference Molina, Borkovec, Davey and Tallis1994) and good reliability and validity in non-clinical and clinical samples (Brown, Antony, & Barlow, Reference Brown, Antony and Barlow1992; Davey, Reference Davey1993; Meyer et al., Reference Meyer, Miller, Metzger and Borkovec1990). The Likert scale ranges from 1 (not at all typical of me) to 5 (very typical of me), with four items reverse scored. Greater scores indicate a greater degree of pathological worry (scores range from 16 to 80). A cut-off score of 62 can be used to identify individuals with probable GAD (Behar et al., Reference Behar, Alcaine, Zuellig and Borkovec2003). Sample items include ‘I am always worrying about something’ and ‘many situations make me worry’. In the present sample, internal consistency was excellent (a = .96).
The GAD-Q-IV (Newman et al., Reference Newman, Zuellig, Kachin, Constantino, Przeworski, Erickson and Cashman-McGrath2002) is a self-report measure of DSM-IV GAD symptoms, which are still consistent with the DSM-5 diagnostic criteria. The GAD-Q-IV has good test–retest reliability (r = .84) and good convergent validity with anxiety measures (k = .67; Newman et al., Reference Newman, Zuellig, Kachin, Constantino, Przeworski, Erickson and Cashman-McGrath2002). Higher scores indicate greater endorsement of GAD symptoms. The GAD-Q-IV was used in the present study to characterise the sample.
The UPPS-P (Lynam et al., Reference Lynam, Whiteside, Smith and Cyders2006) is a 59-item self-report measure of five dimensions of impulsivity. However, only the NU subscale was analysed in the present study. Internal consistency has been demonstrated to be good (a = .86; Whiteside & Lynam, Reference Whiteside and Lynam2001). The Likert scale ranges from 1 (agree strongly) to 4 (disagree strongly), with several items reverse scored. Higher scores indicate greater impulsivity. Sample items include ‘Sometimes when I feel bad, I can't seem to stop what I am doing even though it is making me feel worse’ and ‘It is hard for me to resist acting on my feelings’. In the present sample, internal consistency for the NU subscale was excellent (a = .91).
The Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, Reference Lovibond and Lovibond1995a) is a 42-item self-report questionnaire with three subscales measuring depression, anxiety, and stress. Only the depression scale was analysed in the present study. Internal consistency for the depression scale has been demonstrated to be excellent in non-clinical samples (a = .9) and has good convergent validity with other measures of depression (Lovibond & Lovibond, Reference Lovibond and Lovibond1995b). The Likert scale, which asks about how much each statement applied over the past week, ranges from 0 (did not apply to me at all) to 3 (applied to me very much, most of the time). Higher scores indicate greater symptom severity. Sample items include ‘I couldn't seem to experience any positive feeling at all’ and ‘I felt that I had nothing to look forward to’. In the present sample, internal consistency for the depression subscale was excellent (a = .92).
Open-ended Questions were asked in response to select NU items, as there is a high content overlap across the items. Specifically, participants who endorsed NU items (operationalised as responding with ‘agree strongly’ or ‘agree some’ to items 17, 29, 34, 44, or 51) were shown open-ended questions after completing the UPPS-P. These open-ended questions were developed by our research team. Three possible questions were shown to participants, based on which items on the NU subscale they endorsed.
Participants who endorsed items 29, 34, or 51 were asked: ‘In the previous survey you indicated that when you are upset you often act without thinking, or say things that you later regret. Please describe at least one example of something typical you might do or say when you are upset.’
Participants who endorsed item 17 were asked: ‘You indicated that “When I feel bad, I will often do things I later regret in order to make myself feel better now”. Please describe the most recent time when you were feeling bad and did something you later regretted to make yourself feel better in the moment. Please make sure to describe a situation that was relatively recent and occurred at a specific place and time’.
Participants who endorsed item 44 were asked: ‘You indicated that “I often make matters worse because I act without thinking when I am upset”. Please describe at least one action you take when ‘acting without thinking’.
Participants were included in the present thematic analysis if they endorsed at least one of these NU items. Endorsing at least one item meant that they received at least one of our three questions, depending on which NU item(s) they endorsed. Participants were able to respond to these questions with as many words as needed.
Procedure
The study was programmed in Qualtrics and was delivered via MTurk. After agreeing to the online consent form, participants were then shown a CAPTCHA code and were asked to complete demographic questions as well as the self-report measures. After completing the survey, participants were asked if their data are valid and were shown a debriefing form that included the study aims and contact information for the University Research Ethics Board. Participants were compensated $1.00 USD for their time, consistent with other research in the field at the time this study was conducted (2017) (e.g., Horton & Chilton, Reference Horton and Chilton2010; Raines et al., Reference Raines, Short, Fuller, Allan, Oglesby and Schmidt2016; Tylka, Russell, & Neal, Reference Tylka, Russell and Neal2015).
Analysing Responses to the Open-ended Questions
Two of the authors (B.L.M. and K.E.S.) first examined the responses to the open-ended questions and excluded any extraneous responses that were irrelevant to the questions of interest or that suggested the participants did not understand the questions (e.g., responses such as ‘not sure’ and ‘I don't care’). The initial inter-rater agreement between the two authors was 93%. In total, there were only seven discrepancies between raters with regard to whether or not to include a response in the analysis, which were resolved by discussion and consensus. Nine responses were ultimately excluded, and a total of 152 responses were coded for analysis (86 in the high worry group and 63 in the low worry group).
The same two authors then independently identified and coded themes in the data using Braun & Clarke's (Reference Braun and Clarke2006) guidelines. Given the exploratory quality of the present analysis, an unconstrained, bottom-up approach was considered appropriate (rather than a top-down approach that would necessitate developing themes or categories a priori) to code semantic themes in the data. Once the authors independently identified and coded themes, they discussed their coding together and arrived at final themes by consensus.
Results
Preliminary Analyses
Correlations between key study variables in the entire sample (N values ranged from 235 to 243) revealed that higher scores on the GAD-Q-IV were significantly correlated with greater scores on the PSWQ (r = .79, p < .001). Higher NU was significantly associated with greater scores on the GAD-Q-IV (r = .42, p < .001), the PSWQ (r = .41, p < .001), and the depression subscale of the DASS (r = .42, p < .001). Controlling for depression symptoms as measured by the depression subscale of the DASS, NU remained significantly positively associated with scores on the GAD-Q-IV (r = .16, p = .030) and the PSWQ (r = .20, p = .005).Footnote 1
An independent samples t-test was conducted to analyse the difference in NU between participants high and low in chronic worry. Participants high in chronic worry were found to endorse significantly greater agreement with the NU subscale items (M = 28.9, SD = 7.8) compared to participants low in chronic worry (M = 22.0, SD = 7.0), t(165) = −5.90, p < .001). For participants high in chronic worry, the mean NU was comparable to other samples high in worry and GAD symptoms (e.g., Malivoire et al., Reference Malivoire, Marcos, Pawluk, Tallon, Kusec and Koerner2019a; Pawluk & Koerner, Reference Pawluk and Koerner2013) as well as samples of people with dysregulated eating symptoms (Anestis et al., Reference Anestis, Smith, Fink and Joiner2009; Wenzel, Weinstock, Vander Wal, & Weaver, Reference Wenzel, Weinstock, Vander Wal and Weaver2014) and adolescents with elevated BPD traits (e.g., Fossati, Gratz, Maffei, & Borroni, Reference Fossati, Gratz, Maffei and Borroni2014). The mean NU for participants low in chronic worry was consistent with other studies that investigated NU in non-clinical samples (e.g., Settles et al., Reference Settles, Fischer, Cyders, Combs, Gunn and Smith2012).
Responses to the Open-ended Questions about NU
In response to the open-ended questions asking participants to describe typical responses and actions they engage in when upset or distressed, around half of the responses for both participants high in chronic worry (48.3%) and low in chronic worry (53.8%) were interpersonally oriented. For high worriers, the majority of the interpersonal responses involved saying hurtful things to others (83.7%), most of which involved name-calling (36.0%) and criticising others’ capabilities (8.0%). Other interpersonal themes included threatening to end a romantic relationship (7.0%) and yelling at another person (4.7%). The remaining responses did not fit into the above categories due to being vague (4.7%). For low worriers, the majority of the interpersonal responses also involved saying hurtful things to others (71.4%) with a third of these participants explicitly mentioning name-calling (33.3%) and a third reporting criticising others (33.3%). In addition, 2.9% of interpersonal responses mentioned yelling at another person, and none of the responses included threatening to end a relationship (0.0%). Of note, the number of vague interpersonal responses that did not fit into the above categories was higher for those low in chronic worry (25.7%). In these cases, insufficient detail was provided to confidently further categorise the response.
A number of other themes were identified: a similar percentage of participants’ responses mentioned spending money for both high (13.5%) and low worriers (12.3%), and some people explicitly stated that they did not have the money to spend or that the purchase was superfluous. Furthermore, a higher percentage of high worriers mentioned overeating (10.1%) compared to low worriers (6.2%), with frequent references to junk food, and alcohol use was reported by both high worriers (6.7%) and low worriers (4.6%). Participants described negative consequences associated with alcohol use, including drinking to the point of regret or a negative consequence (e.g., nearly passing out and having a hangover the following day). Furthermore, aggressive behaviours that were not explicitly reported to be directed at another person were reported by both high worriers (7.9%) and low worriers (12.3%), including yelling, cursing, and throwing things. Lastly, there were other infrequent themes that did not fit into those aforementioned. For those high in chronic worry (13.5%), these included making big life decisions (e.g., leaving home, moving), relationship transgressions (e.g., flirting with people other than one's partner), discarding valuable possessions, speeding, starting smoking cigarettes after having quit, and threatening self-harm. For those low in chronic worry (10.8%), other uncommon themes included smoking cigarettes, speeding, and storming out of the house.
Discussion
Consistent with growing research (e.g., Cougle et al., Reference Cougle, Timpano and Goetz2012; Gay et al., Reference Gay, Schmidt and Van der Linden2011; Malivoire et al., Reference Malivoire, Stewart, Tallon, Ovanessian, Pawluk and Koerner2019b; Pawluk & Koerner, Reference Pawluk and Koerner2013), the present study findings support a relationship between chronic worry and a disposition to act rashly and make impulsive decisions when distressed. The relationship between NU and chronic worry and GAD symptoms remained significant even after controlling for depression symptoms. Given that chronic worry is typically characterised by harm-avoidant behaviours, impulsive behaviours remain relatively unexplored in the literature on worry and anxiety psychopathology. This study addresses this gap in the literature by investigating the types of rash behaviours that people high in chronic worry engage in when distressed and whether these behaviours are unique relative to people low in chronic worry. The findings of the present study support that people high in chronic worry engage in NU behaviours to a greater extent compared to those low in chronic worry. Interestingly, the types of NU behaviours reported by participants were similar across high and low worriers, which are consistent with the view in the psychopathology literature that many so-called pathological behaviours exist on a continuum with normal behaviour (e.g., Widiger & Clark, Reference Widiger and Clark2000). Consequently, these findings suggest that we may be unable to distinguish psychopathology based on the types of NU behaviours reported; however, there may be other aspects of the behaviours that are clinically significant. This is akin to the idea that although everyone engages in the act of worrying, and worries about similar types of situations (e.g., finances, work, and problems with uncertain outcomes), chronic pathological worry may be characterised by greater ease of activation, perseveration, or persistence (e.g., Berenbaum, Reference Berenbaum2010). This may also be the case for NU behaviours. According to Cyders and Smith's (Reference Cyders and Smith2008) model of NU, more distressing emotions are likely to perpetuate greater levels of engagement in problematic and extreme impulsive behaviours. Given that people high in chronic worry experience emotions more intensely than non-anxious controls (Mennin et al., Reference Mennin, Heimberg, Turk and Fresco2005), it is possible that they have greater difficulty resisting engaging in these behaviours or regulating the use of the behaviours. It is also possible that NU-motivated behaviours are associated with different antecedents and consequences for those high in chronic worry; experience sampling methodologies may shed light on this question. To our knowledge, this is the first study to provide insight into the types of rash behaviours associated with NU in chronic worriers, and the findings suggest that the types of behaviours are consistent with those outlined in Cyders and Smith's (Reference Cyders and Smith2008) model of NU. All of the themes that were observed in this study are discussed further below in terms of their implications for individuals high in chronic worry.
Interestingly, the most prominent theme that emerged from the analysis was a proclivity to engage in problematic interpersonal behaviour in response to heightened negative affect, including insulting and criticising others and threatening to end the relationship. This finding suggests that when experiencing intense distress, individuals high in chronic worry may be inclined to self-regulate by saying hurtful things to others. It is also possible, however, that because social situations are a prominent source of distress for individuals high in chronic worry (e.g., Pincus & Borkovec, Reference Pincus and Borkovec1994; Whisman, Sheldon, & Goering, Reference Whisman, Sheldon and Goering2000), they are more likely to experience intense distress in interpersonal situations and respond by hurting others. These findings have important implications for our understanding of interpersonal dysfunction and conflict in individuals high in chronic worry and GAD symptoms. First, interpersonal relationships are a prominent source of worry for individuals with GAD (Breitholtz, Johansson, & Öst, Reference Breitholtz, Johansson and Öst1999; Roemer, Molina, & Borkovec, Reference Roemer, Molina and Borkovec1997), and consequently being hurtful towards others would likely perpetuate worry about their relationships. In addition, the tendency to criticise and insult others, or threaten to end the relationship, when distressed is likely to have negative implications for their relationships, which could reinforce interpersonal worries. Impulsive and hurtful reactions in interpersonal relationships may also have important implications for understanding interpersonal dysfunction in chronic worry, including cold and affiliative interpersonal behaviours (Eng & Heimberg, Reference Eng and Heimberg2006; Przeworski et al., Reference Przeworski, Newman, Pincus, Kasoff, Yamasaki, Castonguay and Berlin2011; Salzer et al., Reference Salzer, Pincus, Winkelbach, Leichsenring and Leibing2011). For instance, worriers who have a repertoire of cold interpersonal behaviours may be especially inclined to engage in hurtful action when distressed, which could potentially lead to the dissolution of relationships and maintain the tendency to behave in a cold manner towards others. In contrast, individuals who engage in problematic affiliative behaviours towards others may experience distress when they perceive others impede or do not reciprocate their affectionate or caretaking behaviours (Newman & Erickson, Reference Newman, Erickson and Beck2010), which may prompt rash, hurtful action towards others (e.g., criticising others and threatening to end the relationship). However, the threat of losing a relationship may engender more affiliative behaviours in an effort to restore and maintain the relationship (e.g., checking-in excessively with others and reassurance-seeking). Future research is needed to elucidate the relationship between interpersonal dysfunction and NU in individuals high in chronic worry.
A second related theme was the tendency to display aggressive behaviours, such as yelling, cursing, and throwing things. While it is possible that these responses reflected interpersonal aggression, the participants did not explicitly state that the aggressive acts were directed at others. Consequently, these acts of aggression may reflect a distinct theme from the problematic interpersonal behaviours noted above. These aggressive responses are unsurprising given that irritability and anger are associated with anxiety disorders, including GAD (APA, 2013; Deschênes et al., Reference Deschênes, Dugas, Fracalanza and Koerner2012). In addition, worriers perceive uncertainty to be unfair, which is associated with elevated anger (Fracalanza, Koerner, Deschênes, & Dugas, Reference Fracalanza, Koerner, Deschênes and Dugas2014; Sexton & Dugas, Reference Sexton and Dugas2009), and consequently may precipitate aggressive behaviours to cope.
Other themes that were identified through the analysis included making unnecessary, often expensive, purchases and excessive and uncontrollable eating that may reflect episodes of binge eating. While chronic worry may not conventionally be associated with superfluous spending and excessive eating, it is understandable that these behaviours may be used to reduce aversive arousal. Buying episodes have been shown to relieve negative emotions (Miltenberger et al., Reference Miltenberger, Redlin, Crosby, Stickney, Mitchell, Wonderlich and Smyth2003); however, any relief is suggested to be short-lived due to the onset of negative emotions, such as guilt and shame, and the associated harmful consequences, such as interpersonal conflict and financial debt (McElroy, Keck, Pope, & Smith, Reference McElroy, Keck, Pope and Smith1994; O'Guinn & Faber, Reference O'Guinn and Faber1989). Similarly, although overeating may provide relief by temporarily numbing negative emotions or through distraction, it is often followed by an increase in negative emotions and regret (Hilbert & Tuschen-Caffier, Reference Hilbert and Tuschen-Caffier2007; Whiteside et al., Reference Whiteside, Chen, Neighbors, Hunter, Lo and Larimer2007). In addition to perpetuating distress, superfluous spending and overeating are likely to maintain worry. Specifically, engaging in these behaviours may precipitate a more ‘meta-cognitive’ type of worry about what these behaviours mean about them, for example, that something is wrong with them or that they have lost control over their actions. It is also unsurprising that individuals high in chronic worry turn to alcohol to self-regulate, given that alcohol use is common among people with anxiety (Lai, Cleary, Sitharthan, & Hunt, Reference Lai, Cleary, Sitharthan and Hunt2015). Interestingly, only 6.7% of participants in this study endorsed turning to alcohol when distressed. It is possible that alcohol use may be used more typically as a deliberate and persistent coping strategy as opposed to a strategy used in the heat of the moment to mitigate elevated distress.
Lastly, there were a few unique behaviours that were reported by the minority of the sample (e.g., relationship transgressions, throwing items away, smoking cigarettes, and relocating). These responses highlight the value of asking individuals high in worry and anxiety about their experiences with impulsivity, given that there are a variety of unique and less common ways that they may act impulsively. These behaviours might be easily missed by standard questions, yet reflect underlying NU and important areas of maladaptive coping (e.g., using sexual promiscuity to reduce negative affect). Furthermore, although uncommon, these behaviours are important to consider, as some may have significant negative consequences to physical and sexual health.
Clinical Implications
The study's findings may have implications for both clinical assessment and treatment of disorders involving chronic worry. When completing clinical assessments, it may be useful to enquire about how individuals react to and cope with intense emotions to understand whether they engage in rash behaviours in response to negative affect. For example, clinicians could probe about whether patients experience emotions to be distressing and how they cope with negative emotions. It may be pertinent to specifically enquire about rash action, perhaps providing examples of potential types of behaviours, as anxious individuals may not spontaneously think of their behaviour as rash or impulsive. This could also include conducting a behavioural analysis of reactions to distress to inform case conceptualisation (Malivoire et al., Reference Malivoire, Marcos, Pawluk, Tallon, Kusec and Koerner2019a; Pawluk & Koerner, Reference Pawluk and Koerner2016). Specifically, it would be informative to understand patterns of triggers that elicit intense distress, the problematic behavioural responses, and the unwanted consequences. The present study findings shed light on the types of behaviours that are used to manage distress for individuals with chronic worry. However, given that the identified NU behaviours are not unique to high worriers, solely assessing the types of behaviours would be insufficient to determine whether the behaviours are pathological. It may be useful to query the extent to which the individual engages in the behaviours and the associated level of impairment and distress to determine whether intervention is needed.
It is possible that engaging in rash, uncalculated decisions when distressed contributes to the maintenance of worry, and consequently it may be important to target these behaviours in treatment. Specifically, individuals high in NU and worry may benefit from therapy skills aimed to improve the awareness and tolerance of unconformable emotions as well as more effective emotion regulation skills, such as dialectical behaviour therapy or emotion regulation therapy (Cougle et al., Reference Cougle, Timpano and Goetz2012; Malivoire, Reference Malivoire2020; Mennin, Reference Mennin2006; Pawluk & Koerner, Reference Pawluk and Koerner2016). The developing awareness of emotional reactions may help worriers to identify times when it would be useful to employ strategies, such as distress tolerance skills, to de-escalate their negative affect so that they can engage in adaptive emotion regulation. In addition, strategies for chronic worry, including problem-solving training, worry postponement, and attentional control training, have been proposed as strategies that may reduce the likelihood of making poor decisions in response to negative arousal (Macatee et al., Reference Macatee, Allan, Gajewska, Norr, Raines, Albanese and Cougle2016; Malivoire et al., Reference Malivoire, Marcos, Pawluk, Tallon, Kusec and Koerner2019a). Future research is needed to test the efficacy of these interventions for NU in chronically worried samples.
Given that interpersonal conflict is suggested to maintain chronic worry and GAD (Newman & Erickson, Reference Newman, Erickson and Beck2010) and contribute to low remission rates in treatment for GAD (Borkovec, Newman, Pincus, & Lytle, Reference Borkovec, Newman, Pincus and Lytle2002), it is likely an important therapy target. The finding that nearly half of participants described interpersonal conflict in response to experiencing distress provides further support for the importance of assessing and treating interpersonal dysfunction and conflict in chronic worry. Newman et al. (Reference Newman, Castonguay, Borkovec, Fisher, Boswell, Szkodny and Nordberg2011) investigated the efficacy of combining cognitive behavioural therapy (CBT) with interpersonal and emotional processing interventions to target interpersonal dysfunction in GAD. While adding interpersonal interventions to CBT led to large reductions in GAD symptoms and interpersonal dysfunction, the changes did not significantly differ from a CBT plus supportive listening comparison condition (Newman et al., Reference Newman, Castonguay, Borkovec, Fisher, Boswell, Szkodny and Nordberg2011). Further research is needed to investigate the most effective way to target dysfunction interpersonal behaviours in chronic worry. For example, it is possible that patients may benefit from interpersonal effectiveness skills to facilitate more effective communication and maintain adaptive relationships with others (Malivoire, Reference Malivoire2020; Welch, Osborne, & Pryzgoda, Reference Welch, Osborne, Pryzgoda, Sookman and Leahy2010).
Limitations and Future Directions
A limitation of the present study is the use of an analogue sample rather than a clinical sample assessed via diagnostic interview; however, given the novelty of the research question and the underlying objective of the research, exploration in an analogue sample is an important first step. The sample also consisted of predominantly well-educated females, which may limit the generalisability of these findings. In addition, given the cross-sectional nature of our analyses, causal inferences about the associations between NU and chronic worry and GAD cannot be made. Future research should employ longitudinal research designs to determine the direction of the relationship. Furthermore, given the preliminary nature of this research, we sought to understand how participants interpret the NU items as assessed using the UPPS-P to gain initial insight into the types of behaviours associated with NU in people high in chronic worry. Consequently, our research team developed open-ended questions to assess NU, which have not been validated. It is possible that the questions could be refined to more effectively assess NU. For example, it may be helpful to ask participants to describe their emotional reactions to ensure that they are providing examples of rash action that are specifically in response to experiencing negative affect. This would provide more information with respect to behavioural responses to different types of negative affect (e.g., anxiety and anger). An important future direction will be to conduct a more thorough assessment of NU behaviours, including the antecedents and consequences of the behaviours.
The present study findings suggest that there is more to understand about the unique qualities of NU behaviours in people high in chronic worry. Our findings support that people high in chronic worry engage in NU behaviours to a greater extent compared to people low in chronic worry. However, there may be other aspects of NU behaviours that distinguish pathological NU behaviours from non-pathological behaviours. As discussed, it is possible that these behaviours are more difficult to resist doing or to disengage from and consequently are difficult to regulate. It may be fruitful to understand the attitudes about these behaviours that people high in chronic worry hold, in other words, investigating the motivations underlying these behaviours. It is possible that people high in chronic worry may perceive NU behaviours to be more helpful to self-regulate, which could motivate their maladaptive use. Lastly, it is possible that people low in chronic worry have a larger repertoire of functional coping behaviours and engage in dysfunctional behaviours less frequently compared to people high in chronic worry. In contrast, for high worriers, the identified NU behaviours may be part of a larger repertoire of other dysfunctional behaviours (like chronic avoidance).
In addition, it would be useful to investigate how maladaptive behaviours associated with NU are reinforced. It is hypothesised that the reduction of negative emotion is the main mode of reinforcement; however, this may play out in different ways (e.g., winning an argument, hurting someone so they feel as bad as you, and shutting down criticism). It would be helpful to understand more about why these maladaptive behaviours are being reinforced in order to refine treatment. Furthermore, although there is research examining the association between NU and other constructs relevant to anxiety and chronic worry (e.g., cognitive processes), further research is needed to determine how NU fits into existing models of chronic worry. In addition, most of the research to date has investigated NU in the context of chronic worry and GAD, with some emerging research in obsessive-compulsive disorder (e.g., Cougle et al., Reference Cougle, Timpano and Goetz2012). However, it is likely that NU is a transdiagnostic construct, and it may be fruitful to examine NU in the context of other disorders. Lastly, future research may wish to investigate whether the types of behaviours associated with NU vary across disorders.
Conclusion
This study addresses an important gap in the literature by elucidating the types of behaviours associated with NU in individuals who worry chronically. Importantly, the findings suggest that NU behaviours exist on a continuum and it may be more clinically relevant to consider the extent to which people high in chronic worry engage in NU behaviours as opposed to the type of behaviours. The findings highlight the importance of considering maladaptive interpersonal behaviours in response to distress, which may have important implications for the maintenance of worry. Furthermore, aggressive behaviours, superfluous spending, overeating, and alcohol use are behaviours used by some individuals high in chronic worry to reduce distress quickly; however, these behaviours are likely to have negative implications. Consequently, we propose that behaviours associated with NU are important to assess and address in treatment for chronic worry.
Funding
This work was supported by the Ontario Ministry of Research and Innovation (1-51-52349).
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
