The influence of prior theories on new information leads us to a
confirmation bias. We seek out evidence that would confirm our
(favored) hypothesis rather than testing it against evidence that would
disconfirm it (see chapter 20 by Plous, 1993, for research on this theme).
These two strategies are nicely mirrored in the philosophy of science debate
between schools of verification (e.g., Vienna Circle) and schools of
falsification (e.g., Karl Popper).
A variation on this theme is the feature positive effect. It describes
our tendency to pay much more attention to the instances of something happening
(e.g., two events co-occur) than to the counter-instances of the same thing not
happening. Examples include "always being in the slower line in the grocery
store," "always thinking of mom right before she called," "since the new people
came to town there has been more crime..." or "all welfare recipients drive
expensive cars."
Logicians use a trick to encourage falsification. They call it a reductio
ad absurdum, in which a theorem is proven by showing that its
opposite leads to contradictions. To be critical about our judgments,
then, we should sometimes assume the opposite of our opinion and see whether or
not it accounts at least as well or better for the evidence at hand.
When are people likely to adopt a confirmatory bias? When their
hypothesis is salient or desirable When are people less likely to adopt that
bias? When they are able to consider an alternative (e.g., when that
alternative is more desirable). Compare testing your own hypothesis vs.
testing someone else's...
Guiding theme in judgment research
The amount and speed of input to the human mind is greater than the
capacity and speed of the mind as an information processor. This limitation
has invited much research on the strategies that the mind has developed to deal
with those limits and on the biases associated with these strategies.Formation of beliefs
Most of our beliefs are formed through perception (directly through perceiving
events or indirectly through communication). As a default, we believe what we
perceive; only through additional efforts do we correct those beliefs. Gilbert
(1991) gives plenty of examples and evidence on the powerful impact of this
close link between perceiving and believing. Perceptual illusions (e.g., the
St. Louis arch) show how much that link can persist even in light of perfectly
evident knowledge to the contrary.Giving up
Dawes (1988) likens the tight connection between perceiving and
believing to our reluctance to give up things we have accepted--be they
beliefs, attitudes, or behaviors. Giving up beliefs is particularly
hard, in part because we have to exert effort to cast doubt on them. Thomas
Kuhn describes the motions the scientific community goes through when giving up
a dear theory. Giving up theories or beliefs is easier if we have something
else to adopt--being offered a new belief (or hypothesis) makes it easier to
drop the old one. Craig Anderson demonstrated that changing people's beliefs
works best by providing them with causal explanations for the new
belief. We find it hard to suspend judgment or accept new data without a theory
that explains those data. (In scientific journals, too, you can rarely publish
"mere findings" that contradict an old theory--you must provide a plausible
alternative. )Confirming versus disconfirming
Our beliefs, attitudes, theories influence how we perceive and judge the
world. This is far from trivial if this tendency leads us to distort and
select perceptions, as described nicely in chapter 1 by Plous (1993). Lord,
Ross, and Lepper (1979) provided one of the most powerful illustrations of this
phenomenon when they showed that people with attitudes for or against capital
punishment find all kinds of "flaws" in scientific studies that oppose their
attitude and applaud the high quality of those studies that support their
attitude. (No wonder that it takes a scientific community so long until it
accepts a new theory...)