Uncertainty as an anxiety cue at high and low levels of threat

Highlights

• Participants rated anxiety in scenarios in which uncertainty and threat level were manipulated.

• Higher threat situations provoked higher ratings of anxiety and safety behavior urge.

• Explicit (vs. implicit) uncertainty provoked higher ratings of anxiety and safety behavior urge.

• Explicit uncertainty increased anxiety at low, but not at high, levels of threat.

Abstract

Background and objectives

Intolerance of uncertainty and the overestimation of threat contribute to the maintenance of anxiety; however, the interaction between uncertainty and threat perception has not been examined empirically. The current study examined the extent to which explicitness of uncertainty is involved in perceptions of, and responses to, scenarios about threatening situations.

Methods

A series of systematically varied scenarios were used to examine whether manipulating uncertainty (implicit vs. explicit) and threat level (high vs. low) altered the perception of a situation as anxiety-inducing. Undergraduate participants (n = 373) responded to vignettes about common situations (e.g. taking an elevator) with ratings of anxiety and desire to perform a safety behavior.

Results

Results revealed that higher threat situations, and those in which uncertainty was made explicit, provoked higher ratings of anxiety and urge to perform a safety behavior. In addition, explicit uncertainty significantly increased anxiety and urge to perform a safety behavior at low, but not at high, levels of threat.

Limitations

Participants rated (via self-report) their hypothetical feelings as induced by vignettes, rather than actually experiencing these situations in vivo.

Conclusions

We found evidence for “uncertainty-based reasoning,” in which an individual perceives a situation as more anxiety-provoking (and is more likely to have the urge to perform a safety behavior) when the uncertain aspects of a situation are obvious or explicit, than when such uncertainty is merely implied or tacit. Implications for the understanding of “uncertainty-based reasoning” are discussed.

Introduction

The cognitive model of emotion proposes that emotions arise from beliefs and appraisals (i.e., cognitions) about situations and stimuli, rather than from the situations or stimuli themselves (e.g., Beck, 1976). Moreover, this model specifies that certain types of cognitions give rise to particular emotions. For example, overly negative beliefs about loss and personal degradation (e.g., “I am a failure”) lead to depression. Rigid beliefs about the importance of obeying rules and standards (e.g., “Others must not disobey me”) lead to frustration and anger. A large body of literature indicates that anxiety is associated with beliefs, perceptions, and automatic thoughts concerning threat and the perceived inability to cope with negative outcomes (Beck, Emery, & Greenberg, 1985). In fact, the tendency to overestimate threat appears to play a role in the etiology and maintenance of clinical anxiety and fear (Beck, 1976, Eysenck, 1992, Mathews, 1990, Mathews and MacLeod, 1994, Mathews and MacLeod, 2002, Williams et al., 1988). Such “catastrophic” thinking is also associated with the performance of “safety behaviors” (e.g., carrying a benzodiazepine in one's pocket) that serve as an escape from anxious feelings.

Research also suggests that intolerance of uncertainty (IU) is a cognitive feature of anxiety, worry, and mood disorders (e.g., Carleton, 2012, Carleton et al., 2012, Mahoney and McEvoy, 2012a). IU refers to a set of maladaptive beliefs about the necessity of having guarantees and the incapacity to cope with unpredictability or ambiguity (Obsessive Compulsive Cognitions Working Group, 1997). IU is considered a transdiagnostic factor that has been implicated in conditions such as generalized anxiety disorder (Dugas, Buhr, & Ladouceur, 2004), social anxiety disorder (e.g., Carleton et al., 2010, Mahoney and McEvoy, 2012b), illness anxiety (e.g., Fergus, 2013, Fergus and Bardeen, 2013), obsessive compulsive disorder (e.g., Tolin, Abramowitz, Brigidi, & Foa, 2003), and mood disorders (e.g., Carleton et al., 2012, Yook et al., 2010). A meta analysis confirmed that IU is a broad construct not specific to any particular anxiety disorder per se, nor to this class of problems as a whole; in fact, IU correlates strongly with general negative affect (Gentes & Ruscio, 2011).

Examples of IU-related cognitions include thoughts that uncertainty is bad, reflects poorly on a person, and should be avoided. IU is also associated with the tendency to perceive ambiguous situations as threatening and with difficulty functioning in uncertain or ambiguous situations (Buhr and Dugas, 2002, Krohne, 1993). Most people, for instance, are tolerant of acceptable levels of ambiguity or risk and regard many mundane activities that hold at least some implicit uncertainty as fundamentally safe. Consider, as an example, using the oven in your home. Although the probability of starting a house fire is acceptably low, it is not zero. On the other hand, someone who is less tolerant of uncertainty might explicitly focus on the uncertain aspects of the situation (even in acceptably low risk situations) and focus on the fact that there is always some risk of causing a house fire.

As with the literature on threat perception, high levels of IU are associated with the tendency to overestimate the probability of negative events (Dugas et al., 2004, Dugas et al., 2005, Koerner and Dugas, 2008, Ladouceur et al., 1997). In fact, experimentally increasing IU (e.g., by providing information about the probability of winning money in a gambling simulation) has been shown to lead to increased worry, suggesting a possible causal association between IU and symptoms of anxiety (de Bruin et al., 2006, Grenier and Ladouceur, 2004, Ladouceur et al., 2000, Rosen and Knäuper, 2009). That is, the experience of uncertainty itself serves as a threat cue for some individuals (i.e., “uncertainty based reasoning”). Findings from a study examining IU in relation to social anxiety suggested that the relationship persists independent of anxiety sensitivity, negative affect, and positive affect (Carleton et al., 2010). The extent to which peoples' levels of fear and anxiety are jointly influenced by their perception of threat and the degree to which they attend to uncertainty, however, has not been examined.

It is possible that the experience of anxiety in a given situation is the result of an interaction between one's attention to the uncertain aspects of the situation and one's perception of threat. Accordingly, the present study aimed to investigate how the experience of uncertainty might affect anxiety in situations of varying threat levels. We were also interested in the strength of urges to perform anxiety-reducing (i.e., safety seeking) behavior. We hypothesized that increasing the threat level within the situation would lead to higher ratings of anxiety and urges to perform safety behaviors. Similarly, we hypothesized that making uncertainty explicit would heighten ratings of perceived anxiety and urges to perform safety behaviors. Finally, given that existing cognitive models of anxiety favor threat perception as a key maintenance factor (e.g., Barlow, 2000, Barlow, 2002, Barlow et al., 2014, Clark, 1999), we predicted that at higher levels of threat, uncertainty would play less of a role in generating anxiety and producing urges to perform safety behaviors. Yet at lower levels of threat, we hypothesized that individuals would rely on their experience of uncertainty when judging their expected anxiety and desire to perform anxiety-reducing behaviors (i.e., “uncertainty-based reasoning”).

We elected to use a nonclinical sample for this study. The use of such a sample as an apt analogue for clinical anxiety is predicated on the assumptions that the experience of anxiety is prevalent in both clinical and non-clinical populations, is phenomenologically similar across both populations, and is associated with the same developmental and maintenance factors in clinical and nonclinical individuals. Research to date suggests that these assumptions have been met (e.g., Barlow, 2002).