Discussion
Weaknesses
Our
study had many unforeseen issues and problems. Many of the people in
our sample were unable to respond, either because they were special
education students, they had moved, or they didn’t go to school here
(homeschooled or educated somewhere else). Others simply did not
respond for no apparent reason. Many students who did respond either
lied about their GPA or did not know their GPA. It is possible that
even more students lied or misremembered than we were aware of,
since we gave everyone who provided a plausible GPA the benefit of
the doubt. Some people also gave GPA ranges, which we averaged.
Moreover, students provided GPAs of varying precisions, ranging from
one significant figure to four significant figures. Since we used a
chi-square test, we were forced to use very wide GPA ranges (1-<3
and 3-<5) in order to have expected values that were all over 5. Due
to the nature of our study, we believe that response bias,
nonresponse bias, and insufficient sample size were our greatest
weaknesses.
Extrapolation
Extrapolation of our results to other
schools is somewhat questionable, since other schools may have a
different political climate or different GPA scales. Still, it is
fairly plausible to suggest that there is not significant
correlation between party affiliation and high school GPA among all
American high school students. However, this is merely a hypothesis
and further study would be needed before that claim could be made
with any sort of certainty. As it is, we can confidently say that we
found no significant difference among the GPA distributions of North
Olmsted High School students who identified as Democrats,
Republicans, and neither.
Suggestions
A study building upon the results of our
work could survey many more students than we were able to in our
very limited time frame (perhaps even a census of an entire school).
Barring logistical difficulties involving anonymity, response bias
could be eliminated by looking up GPAs directly from school records.
Obviously this would require significant cooperation from the school
administration to be feasible. A larger study could also include
multiple schools from a larger area.