66 COGNITIVE THEORY AND RESEARCH ON ANXIETY
tion that is mood- incongruent, such as cues of nonthreat or safety. We predict that
information signifying safety or absence of threat would be a stimulus category most
likely to be ignored in anxiety states because it is highly incongruent with this intense
focus on a narrow band of threatening information.
Two questions are relevant for this second hypothesis. First, do highly anxious indi-
viduals exhibit significantly reduced processing of relevant safety information? Second,
do nonanxious individuals show an enhanced processing bias for safety cues? Two other
issues that are related but less central to this hypothesis are whether nonanxious indi-
viduals automatically shift their attention away from threat and whether highly anxious
individuals eventually avoid threatening cues in an effort to intentionally compensate or
suppress the earlier automatic hypervigilance for threat and danger (Mathews & Mack-
intosh, 1998; Mogg & Bradley, 2004; Wells & Matthews, 1994).
High Anxiety: Reduced Safety Signal Processing
As noted in Chapter 2 inhibited processing of safety information is an important faulty
information- processing characteristic of anxiety. Diminished processing of safety might
be a cognitive factor that underlies the propensity of anxious individuals to engage in
safety- seeking behavior, an important factor in the persistence of anxiety (i.e., Rach-
man, 1984a; Salkovskis, 1996a, 1996b; Salkovskis, Clark, Hackmann, Wells, & Gelder,
1999). This is because avoidance and other safety behaviors (e.g., holding on to objects,
venturing out only when accompanied, having immediate access to medication, reassur-
ance seeking, checking) deprive individuals of opportunities to disconfirm their cata-
strophic beliefs. For example, a person with panic disorder who will only go to a store
with a close family member fails to learn that she will not have a heart attack from chest
pain (i.e., the catastrophic fear belief) even though she may feel intense anxiety when
alone in the store. The catastrophic belief, then, persists despite the nonoccurrence of
heart attacks because the person engages in safety- seeking behavior (avoids stores or
takes a friend) that averts the dreaded outcome and reduces anxiety, but it also prevents
the person from learning that the belief is groundless (Salkovskis, Clark, & Gelder,
1996).
Research has shown a link between safety- seeking behavior, catastrophic beliefs, and
persistent anxiety. A questionnaire study of panic disorder (Salkovskis et al., 1996) found
evidence of the predicted associations between threat beliefs and actual safety- seeking
behavior when individuals were questioned about their responses during their most pan-
icky or anxious episodes. In addition, brief treatment analogue studies have shown that
decreases in safety- seeking behavior lead to greater reductions in catastrophic beliefs
and anxiety (Salkovskis et al., 1999; Sloan & Tech, 2002; Wells et al., 1995). If anxious
individuals exhibit less rapid and efficient processing of safety information, this would
leave them with a narrowed intense focus on the threatening aspects of a situation. This
hypervigilance for threat combined with diminished processing of mood- incongruent
safety cues might promote more extreme and effortful attempts to reestablish a sense of
security through safety- seeking behavior (see Figure 3.3 for proposed relationships).
Only a few studies have investigated information processing of safety cues in anxi-
ety. Mansell and D. M. Clark (1999) found that socially anxious individuals exposed
to a social- threat manipulation (give a short speech) recalled significantly fewer positive
public self- referent trait adjectives and Amir, Beard, and Prezeworski (2005) reported