The question of cognitive biases and distortions has long held interest in the study of depression with, for example, the emphasis on logical distortions and irrational beliefs in the early work of Beck (e.g., 1976) and of Ellis (e.g., 1962) (see also Chapter 6). However, as Haaga et al. (1991) have pointed out, the terms "bias" and "distortion" were not distinguished in this early work, and only recently has care been taken with this distinction. A bias is a proclivity to take one direction over another which under some conditions will lead to accuracy or realism, but under other conditions will lead to inaccuracy (see Power, 1991). In contrast, distortion is invariably wrong. For example, if a group of depressed and a group of normal individuals are given a set of positive and negative adjectives to recall, then the depressed people are typically found to recall more negative adjectives and the normal controls more positive adjectives (e.g., Derry & Kuiper, 1981). Neither group is distorting reality, but rather the depressed group are showing a negative bias and the normal group a positive bias (see also Chapter 4). They are presenting different views of the same reality in what has been called a Rashomon Effect in reference to Akira Kurosawa's film masterpiece in which four witnesses recount contrasting views of the same event, a murder in a forest.
As we shall expand upon in more detail below, the debate between proponents of the so-called "depressive realism" position versus cognitive therapy's proposed "depressive distortion" view has been an artificial one in which oranges have been compared with apples: both positions are "true", but under certain conditions. Nevertheless, even if it is now accepted that emotional disorders such as depression and anxiety are associated with cognitive biases rather than distortions, the question remains of the extent of such biases. For example, as discussed in Chapter 4, the network theories of Bower and the schema theories of Beck led to predictions for pervasive biases in a range of cognitive processes in emotional disorders which, unfortunately for the theories, have not been supported; thus, in their summaries of the evidence, Williams et al. (1988, 1997) argued that the biases most apparent in anxiety were those concerned with automatic preattentive processes (see Chapter 6), whereas those most apparent in depression were controlled mnemonic processes. Again we argue that these discrepant theories may have approached the problem of bias from the wrong angle: that whether or not bias is evident in a preattentive, mnemonic, or other cognitive process depends in part on the content and the appraised relevance of that content as we have outlined in Chapter 5.
A further key problem that has arisen with the cognitive bias approach is whether observed cognitive biases are merely state-dependent consequences of depression or whether they are vulnerability factors that are causal antecedents in the development of depression. As with many national economies, this area has been through a number of boom-and-bust cycles. Much of the initial work focused on global self-report measures such as of dysfunctional attitudes, automatic thoughts, and hopelessness, and reported that depressed individuals scored significantly higher than normal controls. Unfortunately, subsequent studies that followed up depressed individuals until they were fully recovered showed that levels of global self-reported attitudes and automatic thoughts return to the levels shown by normal controls. Such findings have led some (e.g., Coyne & Gotlib, 1983) to question the whole basis of the cognitive approach, but have led others (e.g., Clark & Beck, 1999; Power, 1987a; Williams et al., 1997) to call for the use of more sophisticated measures of cognitive vulnerability that can assess automatic as well as controlled processes. The results that have been obtained so far from this enterprise suggest a more complex view of vulnerability than the results from the self-report measures would suggest. However, we will first summarise the evidence for biases from a range of cognitive processes before returning to the theoretical implications.
Was this article helpful?