the given context. Moreover, increasing individuals' motivation to arrive at a defens-
ible judgment does not attenuate reliance on normatively irrelevant information
presented by the researcher. To the contrary, it increases their eorts to ®nd meaning
in the material presented to them (e.g. Tetlock & Boettger, 1996).
As a consequence, the procedures typically used in psychological research are likely
to result in an overestimation of the size and the pervasiveness of judgmental biases.
Note, however, that this analysis does not imply that violations of conversational
norms are the sole source of judgmental biases. Like most robust phenomena, judg-
mental biases are likely to have many determinants. If we are to understand their
operation in natural contexts, however, we need to ensure that their emergence in
laboratory experiments does not re¯ect the operation of determinants that are
unlikely to hold in other settings (see Schwarz, 1996). How exactly we can accomplish
this, however, is a tricky and largely unresolved issue that has been insuciently
addressed in this line of work.
We now know that we can (a) reliably exaggerate many biases by violating con-
versational norms; can (b) reliably attenuate them by undermining the assumption
that the experimenter is a cooperative communicator; or can (c) often avoid their
emergence by being cooperative communicators who do not present misleading
information in the ®rst place. But where do we go from here? How can we study, for
example, how individuals select, weight, and use information of dierential diag-
nosticity when every piece of information we present is rendered relevant by the sheer
fact that we present it, thus biasing the very processes we want to explore? One
potentially promising answer is that we need to heed the socially situated nature of
cognition, paying closer attention to the conditions under which people encounter the
tasks of interest in daily life (for a related discussion see Wyer & Gruenfeld, 1995). As
an example, consider a study on leading questions in eyewitness testimony reported by
Dodd and Bradshaw (1980). Like numerous other researchers, Dodd and Bradshaw
observed that leading questions biased eyewitness memory, yet they only obtained this
eect when the misleading question was asked by the experimenter, not when it was
asked by the defendant's lawyer. Presumably, their participants assumed that the
experimenter was a cooperative communicator, whereas their real world knowledge
about the adversarial nature of courtroom proceedings entailed that the defendant's
lawyer may very well not be cooperative. Hence, they drew on the information implied
by the misleading question in the former but not in the latter case, in contrast to the
assumption that misleading question eects are due to context independent
distortions of memory (e.g. Loftus, 1979). Findings of this type illustrate how real
world knowledge about the social context aects cognitive performance and suggest
that we may miss the boat by limiting our investigations to a social setting that renders
much of this real world knowledge inapplicable: the psychological laboratory.
In fact, the sheer knowledge that one participates in a psychological study may itself
aect participants' judgments. For example, Norenzayan and Schwarz (1999)
presented a newspaper account of a mass murder case on the letterhead of an
`Institute for Personality Research' or an `Institute for Social Research' and asked
Dostları ilə paylaş: