92 COGNITIVE THEORY AND RESEARCH ON ANXIETY
sentences after they read about and imaged being interviewed. The main difference
between groups occurred with the nonanxious group, who showed a quicker latency
to make positive inferences after a positive prime. The high anxiety group failed to
show this positivity bias in their online inferences. The authors concluded that biased
judgments in anxiety may be better characterized in terms of an absence of a protective
positive bias that characterizes healthy individuals (see also Hirsch & Mathews, 2000).
If we extend this deficit inferential processing of positive information to include safety
material, then these results might suggest that nonanxious individuals have a propensity
to elaborate safety- relevant information whereas individuals with social anxiety may
lack such a deliberate, strategic processing bias.
Self- report measures can also be used to assess whether anxious individuals are
less likely to deliberately process safety or corrective information. Researchers at the
Center for Cognitive Therapy in Philadelphia developed a 16-item questionnaire called
the Attentional Fixation Questionnaire (AFQ) to assess whether individuals with panic
disorder fixate on distressing physical symptoms and ignore corrective information dur-
ing panic attacks (Beck, 1988; Wenzel, Sharp, Sokol, & Beck, 2005). A number of the
AFQ items deal with safety issues such as “I am able to focus on the facts,” “I can
distract myself,” “I can think of a variety of solutions,” or “I remember others’ advice
and apply it.” Fifty-five patients with panic disorder completed the questionnaire at four
time intervals: pretreatment, 4 weeks, 8 weeks, and termination. Patients who contin-
ued to have problems with panic attacks scored higher on the AFQ than individuals
with panic disorder who no longer had panic attacks, and treatment improvement was
associated with large pre– posttreatment differences on the ATQ. While only suggestive,
these results are consistent with Beck’s (1988) contention that during a panic attack
individuals are less able to consciously process safety or corrective information.
Summary
At this point it is unknown whether the interpretation threat bias in anxiety also affects
the processing of safety cues. We might expect that safety information would not be
encoded as deeply if the information- processing apparatus is oriented toward threat.
However, to date there is only suggestive evidence for inhibited or diminished elabora-
tive processing of safety information in anxiety, with a present lack of critical research
on this issue.
Clinician Guideline 3.9
Treatment of anxiety might benefit from training that improves deliberate and effortful
processing of safety and corrective information during periods of anticipatory and acute
a n x ie t y.Hypothesis 10. Detrimental Cognitive Compensatory Strategies
In high anxiety states worry has a greater adverse effect by enhancing threat salience
whereas worry in low anxiety states is more likely to be associated with positive effects such