90 COGNITIVE THEORY AND RESEARCH ON ANXIETY
social phobia. In fact, most studies have been unable to find evidence of an explicit
memory bias for threat in social phobia even with information- processing manipula-
tions that map closely to real-life social experiences. Too few memory studies have been
conducted in OCD or PTSD to allow any conclusions to be drawn, although Muller
and Roberts (2005) recently concluded in their review that OCD is characterized by a
positive memory bias for threatening stimuli. Overall research on explicit memory bias
provides only modest support for Hypothesis 8.
Autobiographical Memory
If anxiety is characterized by threat- biased elaborative processing, then we would expect
anxious individuals to exhibit an elevated tendency to recall past personal experiences
of threat or danger. Selective retrieval of autobiographical memories has been demon-
strated most clearly in depression where a negative mood- congruency effect has been
found across numerous studies (for review, see D. A. Clark et al., 1999; Williams et al.,
1997). In the typical autobiographical study, individuals are asked to report the first
memory that comes to mind in response to neutral or valenced cue words. The auto-
biographical memory task has good ecological validity because it assesses individuals’
personal memories and experiences, although biased recall could be caused by a greater
number of past threatening experiences in the lives of anxious individuals (MacLeod,
1999). Thus retrieval differences may not reflect memory differences as much as differ-
ences in life experiences.
Only a few studies have investigated autobiographical memory in anxiety. Rapee
et al. (1994, Experiment 4) failed to find any differences between socially anxious and
nonanxious groups in number of positive or negative memories recalled to social or
neutral stimulus words, although Burke and Mathews (1992) produced more positive
results indicative of an autobiographical memory bias in GAD. Mayo (1989) found that
high trait anxiety was associated with recall of fewer happy and more unhappy per-
sonal memories. Wenzel, Jackson, and Holt (2002) reported that individuals with social
phobia recalled more personal memories involving negative affect in response to social
threat cues but this effect was weak, accounting for only 10% of their social threat-cued
memories. Although only a few studies of autobiographical memory in anxiety have
been published, it may emerge that this memory bias may be specific to certain anxiety
disorders such as GAD but not to others like social phobia.
Summary
Overall there is considerable empirical support for Hypothesis 8, that anxiety is charac-
terized by facilitation of threat at the elaborative, strategic stage of information process-
ing. The strongest research support is from the interpretative bias research. The most
frequent finding is biased threat- related judgments in high anxiety. This is most appar-
ent when ambiguous information is presented that is specific to the fear concerns of the
individual (e.g., body sensations for panic disorder and negative social evaluation for
social phobia). There is some indication that the interpretation bias in anxiety is persis-
tent, focuses mainly on the severity of threat, and has a causal impact on anxiety. Ques-
tions still remain on whether the interpretation bias primarily involves the exaggeration
of threat or the diminution of a positivity bias that characterizes nonanxious states.