2. The theoretical framework and research hypothesis
Akerlof and Dickens (1982) pioneered the application of cognitive
dissonance in economic theory. Their model showed that once people
choose a risky option, they tend to presume their risky choice to be
safe. Hence, after deciding on a risky choice, their evaluation of risk
should be undervalued. This research finding is exemplified by a decision
to work in nuclear power plants. Workers who choose to work in
a nuclear power plant tend to presume the risk of radioactive exposure
to be low. Therefore, the radioactive risk should not be considered for
the assessment of their preferences. This is an economic interpretation
of the effect of cognitive dissonance; that is, people change their preference
in order to justify their previous decision.
The model, which considers both pecuniary and psychological costs
of believing that the work is safe, shows that the precaution level, after
one decides to choose a risky option, is inconsequential. This theoretical
explanation is useful to shed light on the Japanese preference for nuclear
power even as they are tormented by conflicting feelings—the fear of
danger from nuclear power plants versus the fear of rising electricity
rates. In such circumstances, people are expected to shift toward an
affinity for nuclear power once they are required to declare their
preference for nuclear power plants.
In the current study, we set two scenarios: a forced choice scenario
and a non-forced choice scenario with a “no-choice option.” In the
forced choice scenario, we ask Japanese people to declare their choice
from just two alternatives that involve trade-off relationships between
dependency on nuclear power generation and electricity rates. In line
with Akerlof and Dickens's work, we expect a preference change from
a resolution of cognitive dissonance. However, in the non-forcedchoice
scenario,we ask Japanese people for their decision about nuclear
power plants, but this scenario is different in that they can postpone the
declaration of their decision. Because they do not need to resolve their
cognitive dissonance, we do not expect their preference to change.
To examine this hypothesis in the current paper, we administer a
short preliminary survey followed by two main surveys. The first main
survey is designed to assess preference changes caused by cognitive dissonance
(explained in Section 4), and the second incorporates discretechoice
experiments designed to assess the effect of the no-choice option
(explained in Section 5). In the second survey, we present a forcedchoice
scenario (without a no-choice option) and a non-forced-choice
scenario (with a no-choice option). The attributes and levels of the electricity
alternatives are common between the two scenarios. The only
difference is the availability of the no-choice option, forwhich the status