INTROSPECTIVE ACCESS -- DO PEOPLE KNOW WHY THEY DO WHAT THEY DO?


INTROSPECTIVE ACCESS -- DO PEOPLE KNOW WHY THEY DO WHAT THEY DO?

 (based in part on the site below)
http://www.brunel.ac.uk/~hsstbbp/soclec10.htm
For more information see my notes.

Nisbett and Wilson's (1977) claims about the limitations of introspective access are essentially:

1. "People often cannot report accurately on the effects of particular stimuli on higher order, inference-based responses. ... The accuracy of subjective reports is so poor as to suggest that any introspective access that may exist is not sufficient to produce generally correct or reliable reports."
2. "When reporting on the effects of stimuli, people may not interrogate a memory of the cognitive processes that operated on the stimuli; instead, they may base their reports on implicit a priori theories about the causal connection between stimulus and response" (p. 233).
(See Linde, 1993, ch. 6 pp. 184-7, ch.7 & 8 for sources of these "a priori" theories in common culture -- professional ideas leaked into popular culture, religion, self-help notions, etc..)

Experimental evidence for lack of introspective access

Failure to report influence of effective stimuli

Nisbett and Wilson primed certain word associations using a word-pair learning task, e.g., participants were asked to remember a list of word-pairs including "ocean-moon" and were later asked to name a washing powder. The "ocean-moon" word-pair successfully primed the response Tide, but very few participants reported being aware that the memory task had influenced their subsequent associations.
In another study, participants were asked to choose the best quality pair of stockings from a range of samples ostensibly for consumer research purposes. Because of a strong order effect, participants selected the pair of stockings that was the furthest right in the display four times as often as the pair that was furthest left. However, participants never referred to position as a factor in their choice, but instead explained their selection behavior in terms of the stockings' inherent qualities.
They report studies by Maier on "functional fixedness" -- the classic pendulum problem where subjects are told to tie two strings hanging from the ceiling together. Solution involves tying objects to each string and swinging them so that they come close enough for the subject to grasp them. The trick is to use the tools provided, not in their usual function (hammer, pliers) but simply as weights. The dependent variables are numbers of solutions and time of solution.
Against this baseline, Maier examined effects of cues and labels. Relevant to Nisbett and Wilson, Maier compared a version of the task where he nudged "inadvertently" one of the strings. This greatly facilitated solutions yet subjects did not report noticing the swinging but confabulated various stories about why they thought of a pendulum. 

Reports of influence of actually ineffective stimuli

In another of Nisbett and Wilson's studies, participants read an emotionally upsetting passage of prose. Different versions of this passage were prepared with certain graphic descriptive sections taken out. Although none of the missing sections actually made any statistical difference to the rated emotional impact of the stories, participants who had read the full version reported that the relevant text had intensified their reaction, and participants who had read the shortened version similarly believed that including the extra material would have made a difference. In a related study, judgements of a sad film were actually unaffected by the noise of an electric power-saw coming from the corridor outside, yet participants still reported that this distraction had interfered with their rated response to the film.

Absence of differences between reports by actors and observers

Nisbett and Bellows (1977) directly compared actors' and observers' estimates of the influence of various factors on judgement to determine if people really do have privileged access to information about the processes controlling their own behavior. Participants read through a portfolio describing a candidate for a job in a crisis intervention center, and made judgements about the candidate's suitability for the post. Different participants received different patterns of information concerning the candidate and what happened during her interview, some of which was intended to influence judgement without participant awareness, and some of which was intended to be a plausible but actually ineffective factor in evaluation.
Actual and reported effects of the various factors were compared, and it was found that participants were often wrong about their relative influence. For example, the only factor that actually had a significant impact on liking for the candidate was information about whether she had accidentally spilt some coffee during the interview, but participants reported that this factor was the least influential on their judgement.
One case in which actor participants were correct in their estimates of influences on judgement concerned the genuine impact of information about the candidate's academic record on ratings of her intelligence. However, observer participants (who were only told what kind of information was presented in the study without actually making the ratings or reading the full portfolio) also judged this academic information to be influential, so the actor participants' accuracy may simply have been attributable to their fortuitous possession of an accurate implicit theory about the factors underlying judgements of intelligence.
 

Criticisms and Alternative explanations of Nisbett and Wilson's results

(JL) Both of these have merit but do not negate N&W's overall point that in at least some -- if not most-- cases introspection does not access the mental machinery directly.

1. Failure of memory

Ericsson and Simon (1980) suggested that participants' mistaken explanations may often be due to their inability to remember the relevant processes rather than lack of awareness of the processes at the time when they actually took place. We know that masking and subsequent processing can rapidly (< one second) eliminate awareness of stimuli whose impact is obvious. For example, ambiguous words are quickly reduced to the contextually appropriate sense even though momentarily both senses are available to subjects.

2. Failure of descriptive ability

We may be fully aware of certain images and feelings that guide our behavior but have problems in translating these into words when asked to give an explanation. We may not even have a word -- that is why new words are created. Remember our discussion of the invention of "belief."