Implicit Adjustment vs. Debiasing: Two Types of Mental Correction

Are All Correction Efforts Deliberative and Theory-Driven? Implicit Adjustment Versus Debiasing

There has been some controversy over the manner in which people attempt to correct their psychological responses. Some theorists, such as Wilson and Brekke (1994) and Wegener and Petty (1997), stress the importance of deliberative applications of lay theories. Wegener and Petty’s (1997) flexible correction model, for example, argues that correction is driven by the conscious belief that bias has occurred and a naive theory about the direction and magnitude of the bias. Other theorists, such as Martin (1986) and Martin and Stapel (1998), argue that correction can also be a spontaneous and automatic response to the nature of the situation and stimulus.

We suggest that both camps in this debate are correct because there are different levels and types of correction. We agree with Martin and Stapel (1998), who state that a great deal of mental processing occurs at an automatic and implicit level, and that some of this processing involves a correction for information deemed inappropriate or irrelevant to the judgment at hand. In fact, a number of such automatic correction processes have been documented in the social cognition literature, such as research on self-attribution, in which people observe their behavior and, under some circumstances, infer that this behavior reflects their internal states (e.g., Bem, 1972). The attribution about the cause of one’s behavior or physiological responses sometimes involves a correction process, whereby people discount the role of one potential cause if another plausible one is present. These corrections occur automatically and nonconsciously.

Consider a classic misattribution study by Dutton and Aron (1974), in which male participants encountered an attractive woman while crossing a flimsy footbridge spanning a 200-ft gorge. The participants had to decide how much of their arousal was due to fear from the bridge versus attraction to the woman. This decision can be characterized as a correction process; people had to partial out the effects of one plausible cause from another. As it happened, people who encountered the woman on the bridge misattributed some of their arousal to attraction to the woman.

It seems clear that this correction process did not occur consciously and deliberatively. The men did not stand on the bridge scratching their heads thinking, "Hmm, is she beautiful or am I scared? Or a little of both? I’d say I’m feeling 37% fear and 63% attraction." Rather, their phenomenal experience was attraction to the woman with no conscious recognition of the processes that produced it. Although the issue of how aware people are of attribution processes was seldom discussed in the original literature on this topic, Nisbett and Wilson (1977b) later documented that participants in these studies rarely are able to verbalize the attribution processes hypothesized to have occurred.

The broader point is that mental correction occurs on a continuum of implicit and nonconscious adjustments of a response to explicit and deliberative adjustment, and researchers should keep in mind that when they use the term correction, they are sometimes referring to quite different mental processes. To help clear up this confusion, researchers should use different terms for the different ends of this continuum. We suggest that implicit adjustment be used to refer to rapid and nonconscious correction discussed by attribution theorists and Martin (1986), and debiasing (Fischhoff, 1982) be used to refer to deliberative and theory-driven correction discussed by Wilson and Brekke (1994) and Wegener and Petty (1997).

As noted by Wilson and Brekke (1994), the very fact that implicit adjustment is so common is a major source of bias. They discussed two subcategories of contamination: (1) unwanted consequences of automatic processing, whereby people process information automatically in undesired ways; and (2) source confusion, whereby people confuse two or more sources of a judgment, memory, or feeling. These types of contamination can involve implicit adjustment, such as the kind of misattribution errors just discussed.

The existence of implicit adjustment puts people in a difficult position to know whether a particular response is biased because they do not know whether a phenomenal judgment has already been “corrected.” Suppose that Michael finds Donald to be slightly aggressive. How accurate is this impression? Was it biased by irrelevant information, such as the fact that Michael had just read a newspaper article about spousal abuse, which primed the construct of aggression? There are three possibilities:

1. The newspaper article did not prime aggression, so the article did not bias Michael’s impression of Donald.

2. The newspaper article primed the construct of aggression and produced assimilation, such that Michael finds Donald to be more aggressive than he would have if he had not read the newspaper article.

3. The newspaper article primed the construct of aggression, but Michael already corrected for this fact nonconsciously. This resulted in a contrast effect, such that Michael finds Donald to be less aggressive than he would have if he had not read the newspaper article.

Obviously, the extent to which Michael attempts to consciously adjust his impression depends on which of these processes he believes has occurred. Because he has no direct access to this process, he is at the mercy of his naive theories, which might well be wrong.