Empirical Status of the Cognitive Model 69
There is now considerable evidence that an automatic threat appraisal process is
involved in the preattentive threat bias in anxiety. Mathews and Mackintosh (1998)
proposed that the representation of potential threat is dependent on activation of a
threat evaluation system (TES). The TES represents the threat value of a previously
encountered stimulus and is computed automatically at an early stage of information
processing. During heightened anxiety, the output from the TES increases so that a
lower threshold of stimulus intensity is required for threat valuation. Thus Mathews
and Mackintosh argue that a hypervigilant attentional threat bias occurs in response
to a prior preconscious automatic threat appraisal. Mogg and Bradley (1998, 1999a,
2004) also proposed that threat stimulus evaluation is a critical part of the automatic
information processing that occurs in anxiety (see also the self- regulatory executive
function model proposed by Wells, 2000). Recent theoretical accounts of fear and anxi-
ety derived from a conditioning perspective propose that information is first analyzed
by feature detectors and a preconscious “significance evaluation system” that results in
a quick judgment of the fear relevance of stimuli (Öhman, 2000). Thus our contention
that automatic threat appraisal is a critical component of primal threat mode activation
is entirely consistent with other cognitive and behavioral models of fear and anxiety.
Implicit memory tasks offer an excellent experimental paradigm for investigating
the presence of automatic threat evaluation in anxiety. These tasks involve memory
retrieval in which some previously encoded information causes enhanced performance
on a subsequent task even though the individual has no awareness or recollection of the
relation between the prior experience and the task at hand (Schacter, 1990; Sternberg,
1996). In other words, previous exposure to a stimulus passively facilitates subsequent
processing of the same stimuli and this “priming effect” is thought to reflect the degree of
integrative processing that occurs during stimulus encoding (MacLeod & McLaughlin,
1995). Implicit memory more likely reflects automatic information processing, whereas
explicit memory, a deliberate and effortful retrieval of stored information, maps more
closely onto controlled, strategic processes (Williams et al., 1997).
Word Stem Completion
Implicit memory was first investigated with the word completion task. In this task indi-
viduals are presented with a list of anxiety- relevant (e.g., disease, attack, fatal) and
neutral (e.g., inflated, daily, storing) words. After a filler task, individuals are given a set
of word fragments, such as the first three letters of a word, and are asked to complete
the fragment with the first word that comes to mind. A tendency to complete the word
fragment with a less common word that was included in a previously presented word
list would be an example of implicit memory. In the following example a threat- priming
effect would be evident by completing the word fragment with a previously presented
threat word rather than with a more common neutral word.
Encoded List Word Fragment Possible Response
coronary cor coronary vs. corn
attack att attack vs. attend
fatal fat fatal vs. father