influenced much by the general views of the society. However, it is important to note that even
though the data show that mass media does not have a direct influence on bottled water drinking
behavior, the respondents who acquired beliefs from mass media but does not drink bottled water
are likely to have also acquired their beliefs from other sources, especially environmental
education.
Physical location only has influence on bottled water drinking behavior if the behavior is
“do not drink bottled water”, because the respondents that do not drink bottled water tend to have
a high score for the water quality of the city where they lived. There is no direct link between
physical location and the behavior if the respondent drinks bottled water. Although about 60% of
respondents who drink bottled water have relatively low score for the water quality, the
remaining 40% have relatively high score for the tap water quality. The findings support the
study of Doria (2010) that people tend to prefer what they are used to, so the respondents who
live in places that have good tap water quality tend to drink more tap water. However, the tap
water quality of the place where the respondents live is only weakly associated with their bottled
water drinking behavior, because factors such as social and economic status could potentially
influence personal preference.
On the other hand, the link between education and interpersonal source of information
and bottled water drinking behavior is much stronger. The data (figure 2) show that 75% of the
respondents who have acquired beliefs from environmental education do not drink bottled water
and 70% of those who did not acquire their beliefs from environmental education do drink
bottled water, which indicates a strong positive correlation between environmental education and
bottled water drinking behavior. Nevertheless, there are still cases where the influence of
environmental education on behavior is not very clear due to several reasons. Firstly, there are 15
respondents who have not received any environmental education but still choose not to drink
bottled water. For these respondents, the score for tap water quality of the primary city (the city
where they lived the longest) is generally high, and their primary caretaker is more likely to
prefer tap water over bottled water. There are also respondents who have received environmental
education but still choose to drink bottled water. These respondents tend to own a re-usable
water bottle and use it several times per week if not every day, and therefore, bottled water is not
their main source of drinking water.
The positive correlation is the strongest between interpersonal source of information and
bottled water drinking behavior. The data show a strong consistency between the beliefs of the
respondents and that of their primary caretakers. When being asked if bottled water is safer, there
is 76% consistency between the respondents and their primary caretakers. The consistency is
slightly lower, which is 60%, when they are asked if tap water produces less waste. Nearly all of
the respondents agree with the statement that tap water produces less waste, with only 17%
reported to be unsure, and no one disagreed. However, 24% of the primary caretakers are
reported to disagree with that statement. The inconsistency between the beliefs of the
respondents and their caretakers could be caused by environmental education or the lack of,
because none of the parents acquired beliefs from environmental education (class/seminar/etc.),
whereas the respondents have learned that tap water produces less waste than bottled water from
various sources.
Based on the data, education and interpersonal source of information have much more
influence than impersonal source of information and physical location over bottled water
drinking behavior, and the results supports the findings of Dori (2010) and Park, Scherer &
Glynn (2001). However, the factors that have influence over behavior are often interrelated, 16
therefore it is rather difficult to study them as individual factors. In this study and also previous
studies, it is shown that the influence of impersonal source of information can be affected by
education, because education can influence the selection of environmental communication
channels, and a person with more knowledge has more control over his or her own health risks
(Doria, Factors influencing public perception of drinking water quality, 2010). Also,
interpersonal source of information can influence the respondents differently depending on their
educational level (whether received environmental education or not), because the respondents
and their caretakers do not have the amount of environmental knowledge – the respondents tend
to have learned more about the environment, especially students from Cornell University, where
the issue of sustainable future is very salient. Aside from the four factors mentioned in this study,
there are many other factors such as gender, cultural background that could potentially influence
one’s bottled water drinking behavior. In future studies on factors influencing bottled water
drinking behavior, it is important to test the influence of each factor independently and also
combined. Also important to incorporate in future studies are the style of environmental
education and best age to give such education, for example, whether an environmental lecture or
regular outdoor education is better, and whether it is better to give the lecture or course to a
certain grade or continue throughout four years of high school or middle school. Finally, this
study assumed that the behavior and attitude of the respondents are consistent, however, this may
be always be true. A future study could investigate the possible inconsistency between one’s
attitude and behavior, for example, a person is aware of the potential pollution caused by plastic
water bottles, but he or she still chooses to drink bottled water knowing the fact that such action
is harmful to the environment. 17
There are a number of improvements to this study. The most important improvement
would be an expansion on the sample size to eliminate coverage error. In this study, the majority
of the respondents are students of Cornell University, but Cornel University is one the most
“green” universities of the country, therefore to increase the validity of the study, respondents
should be chosen randomly from universities or colleges, also chosen randomly across the
United States. In addition, the survey questions are needed to be revised; the questions regarding
the beliefs of the respondents and that of their caretakers should be the same so the analysis of
the influence of interpersonal source of information would be more accurate. Finally, it would
improve the results of the study if the beliefs of the primary caretakers are not obtained from the
respondents, but through direct interview with the caretakers. Although in this study, it is
assumed that the respondents were familiar with the beliefs of their caretakers, it is also possible
that they were unsure, which could bias the results.