Empirical Status of the Cognitive Model 91
There is some evidence that a conscious strategic processing of threat is evident in
the form of heightened negative expectancies. Anxious individuals may be more likely
to expect that negative or threatening future events will happen to them, although more
research is needed to establish this finding. Experiments on the covariation bias indicate
that fear- related expectancies in phobic states can result in biased perceptions of envi-
ronmental contingencies (MacLeod, 1999). Whether covariation biases also operate in
the anxiety disorders requires further research. However, at this stage there is at least
some experimental support for the view that anxiety involves a biased expectancy for
future negative or threatening personal events.
Finally, the considerable research literature on explicit memory bias in anxiety has
established that a biased retrieval of threat- relevant information is evident in panic dis-
order but not in social phobia or GAD. Too few memory studies have been conducted
on individuals with OCD or PTSD to allow firm conclusions. In addition anxious indi-
viduals may have a tendency to recall personally threatening memories and this could
contribute to other elaborative processes such as anxious rumination and postevent
processing (see Hirsch & Clark, 2004). However, evidence for selective autobiographi-
cal memory for threat is very tentative at the present time.
Clinician Guideline 3.8
Considerable empirical evidence supports therapeutic interventions that seek to change the
conscious strategic information processing that is the basis of an exaggerated reappraisal of
threat. Modify intentional threat evaluations, expectancies, and memory retrieval to estab-
lish a more balanced reappraisal of immediate threat that can have a positive impact on the
automatic processes of fear activation.Hypothesis 9. Inhibited Safety Elaboration
Explicit and controlled cognitive processes in anxiety will be characterized by an inhibitory
bias for safety information relevant to selective threats such that memory retrieval, outcome
expectancies, and judgments of ambiguous stimuli will evidence fewer themes of safety in
comparison to nonanxious individuals.
If anxious individuals have a bias for consciously and effortfully processing threat-
relevant information, is it not possible that these same strategic processes might be
biased against safety- related cues? Unfortunately, very little experimental research has
addressed this possibility. Even though a number of attentional deployment studies have
shown that anxious individuals exhibit attentional avoidance of threat stimuli at longer
presentation intervals (see discussion under Hypotheses 1 and 2), there is practically no
research on whether anxious persons show a more deliberate inhibition of safety infor-
mation processing. Other researchers, such as D. M. Clark (1999), have emphasized
that safety behaviors play an important role in the persistence of anxiety, but they fail
to consider whether highly anxious individuals might actively inhibit the processing of
safety material.
In a series of experiments Hirsch and Mathews (1997) investigated the emotional
inferences that high and low anxious individuals made when primed with ambiguous