Saying Is Believing: This has been called the "saying is believing"
paradigm. That is, dissonance theory predicts that we begin to believe
our own lies—but only if there is not abundant external justification
for making the statements that run counter to our original attitudes.

...the greater the reward for compliance, the greater the probability
that a person will comply.

...When it comes to producing a lasting change in attitude, the
greater the reward, the less likely any attitude change will occur.

If all I want you to do is recite a speech favoring Fidel Castro, the
Marx brothers, socialized medicine, or anything else, the most
efficient thing for me to do would be to give you the largest possible
reward. This would increase the probability of your complying by
making that speech.

But suppose I have a more ambitious goal: Suppose I want to effect a
lasting change in your attitudes and beliefs. In that case, just the
reverse is true. The smaller the external reward I give to induce you
to recite the speech, the more likely it is that you will be forced to
seek additional justification for delivering it by convincing yourself
that the things you said were actually true. This would result in an
actual change in attitude rather than mere compliance. The importance
of this technique cannot be overstated. If we change our attitudes
because we have made a public statement for minimal external
justification, our attitude change will be relatively permanent; we
are not changing our attitudes because of a reward (compliance) or
because of the influence of an attractive person (identification).
http://tinyurl.com/459wx

We are changing our attitudes because we have succeeded in convincing
ourselves that our previous attitudes were incorrect. This is a very
powerful form of attitude change.

...Leon Festinger and J. Merrill Carlsmith. These investigators asked
college students to perform a very boring and repetitive series of
tasks—packing spools in a tray, dumping them out, and then refilling
the tray over and over, or turning rows and rows of screws a quarter
turn and then going back and turning them another quarter turn. The
students engaged in these activities for a full hour. The experimenter
then induced them to lie about the task; specifically, he employed
them to tell a young woman (who was waiting to participate in the
experiment) that the task she would be performing was interesting and
enjoyable.

Some of the students were offered $20 for telling the lie; others were
offered only $1 for telling the lie.

After the experiment was over, an interviewer asked the liars how much
they enjoyed the tasks they had performed earlier in the experiment.

The results were clear-cut:

Those students who had been paid $20 for lying — that is, for saying
the spool packing and screw turning had been enjoyable — rated the
activity as dull. This is not surprising — it was dull.

But what about the students who had been paid only $1 for lying? They
rated the task as enjoyable. In other words, people who received
abundant external justification for lying told the lie but didn't
believe it, whereas those who told the lie in the absence of a great
deal of external justification moved in the direction of believing
that what they said was true.

Research support for the "saying is believing" phenomenon has extended
beyond relatively unimportant attitudes like the dullness of a
monotonous task. Attitude change has been shown on a variety of
important issues.

...in one experiment, Arthur R. Cohen induced Yale college students to
engage in a particularly difficult form of counterattitudinal
behavior. Cohen conducted his experiment immediately after a student
riot in which the New Haven police had behaved in a rather brutal
manner toward the students.

The students (who strongly believed the police had behaved badly) were
asked to write an essay in support of the actions taken by the police.
They were urged to write the strongest, most forceful defense of the
police actions they could muster. Before writing the essay, students
were paid for their efforts.

There were four conditions: Some students were paid $10; others, $5;
still others, $1; and a fourth group, the paltry sum of 50 cents.
After writing his essay, each student was asked to indicate his own
private attitudes about the police actions. The results were perfectly
linear: The smaller the reward, the greater the attitude change. Thus,
students who wrote in support of the New Haven police for the meager
sum of 50 cents developed a more favorable attitude than did those who
wrote the essay for $1; the students who wrote the essay for $1
developed a more favorable attitude toward the actions of the police
than did those who wrote the essay for $5; and those who wrote the
essay for $10 remained the least favorable. The less the external
justification in terms of money, the greater the attitude change.

Let's look at race relations and racial prejudice — surely one of our
nation's most enduring problems. Would it be possible to get people to
endorse a policy favoring a minority group — and then see if their
attitudes become more favorable toward that group? In an important set
of experiments, Mike Leippe and Donna Eisen-stadt induced white
college students to write an essay demonstrating counter-attitudinal
advocacy: publicly endorsing a controversial proposal at their
university—to double the amount of funds available for academic
scholarships for African-American students. Because the total amount
of scholarship funds were limited, this meant cutting by half the
amount of funds available for scholarships for white students. As you
might imagine, this was a highly dissonant situation. How might the
students reduce dissonance? The best way would be to convince
themselves that they really believed deeply in that policy—that,
taking the big picture into consideration, it was only fair to offer
more financial aid to African-Americans. Moreover, it is reasonable to
suggest that dissonance reduction might generalize beyond the specific
policy— that is, the theory would predict that their general attitude
toward African-Americans would become more favorable and much more
supportive. And that is exactly what Leippe and Eisenstadt found.

What constitutes external justification? As I mentioned a moment ago,
external justification can and does come in a variety of forms. People
can be persuaded to say things or do things that contradict their
beliefs or preferences if they are threatened with punishment or
enticed by rewards other than monetary gain—such as praise or the
desire to please. Furthermore, most of us would consider doing
something that we otherwise wouldn't do if a good friend asked us to
do it as a favor. To take a farfetched example, suppose a friend asked
you to eat an unusual food she or he had recently learned to prepare
in an "exotic foods" cooking class. And just to make things
interesting, let's say the food in question was a fried grasshopper.
Now, imagine the reverse situation—that someone you didn't like very
much asked you to sink your teeth into a fried grasshopper.

Okay, are you ready? Assuming you went ahead and ate the grasshopper,
under which circumstance do you think you would enjoy the taste of it
more—when asked to eat it by a good friend or by someone you didn't
like? Common sense might suggest that the grasshopper would taste
better when recommended by a friend. After all, a friend is someone
you can trust and, hence, would be a far more credible source of
information than someone you didn't like. But think about it for a
moment: Which condition involves less external justification? Common
sense notwithstanding, the theory of cognitive dissonance would
predict that you would come to like eating grasshoppers more if you
ate one at the request of someone you didn't like.

Here's how it works: Your cognition that eating a grasshopper is
repulsive would be at odds with the fact that you just ate one. But if
it was your friend who made the request, you would have a great deal
of external justification for having eaten it—you did it as a favor
for a good friend. On the other hand, you would not have adequate
external justification for munching on a grasshopper if you did it at
the request of someone you didn't like. In this case, how could you
justify your contradictory behavior to yourself? Simple. The way to
reduce dissonance would be to change your attitude toward grasshoppers
in the direction of liking them better—"Gee, they're pretty tasty
critters after all."

While this may seem a rather bizarre example of dissonance-reducing
behavior, it's not as farfetched as you might think. Philip Zimbardo
and his colleagues conducted an analogous experiment in which army
reservists were asked to try fried grasshoppers as part of a study
allegedly about "survival" foods. For half of the subjects, the
request was made by a warm, friendly officer; for the other half, it
was made by a cold, unfriendly officer. The reservists' attitudes
toward eating grasshoppers were measured before and after they ate
them. The results were exactly as predicted above: Reservists who ate
grasshoppers at the request of the unpleasant officer increased their
liking for them far more than those who ate grasshoppers at the
request of the pleasant officer. Thus, when sufficient external
justification was present—when reservists complied with the friendly
officer's request—they experienced little need to change their
attitudes toward grasshoppers. They already had a convincing
explanation for why they ate them—they did it to help a "nice guy."
But reservists who complied with the unfriendly officer's request had
little external justification for their action. As a result, they
adopted a more positive attitude toward eating grasshoppers in order
to rationalize their discrepant behavior.

What Is Inadequate Justification? Throughout this section, I have made
reference to situations where there is inadequate external
justification and to those with an abundance of external
justification. These terms require some additional clarification. In
the Festinger-Carlsmith experiment, all of the subjects did, in fact,
agree to tell the lie—including all of those paid only $1. In a sense,
then, $1 was adequate—that is, adequate to induce the subjects to tell
the lie; but as it turns out, it wasn't sufficient to keep them from
feeling foolish. In order to reduce their feeling of foolishness, they
had to reduce the dissonance that resulted from telling a lie for so
paltry a sum. This entailed additional bolstering in the form of
convincing themselves that it wasn't completely a lie and the task
wasn't quite as dull as it seemed at first; as a matter of fact, when
looked at in a certain way, it was actually quite interesting.

The Social Animal - Elliot Aronson - 8th Edition 1999
http://www.amazon.com/exec/obidos/ASIN/0716733129/