Conclusion

Our first test, the chi-squared test, was statistically inconclusive since there were multiple expected cell counts that were less than five, and thus the assumptions for this test were not met. However, the results of the test indicated that the population proportions were not truly homogeneous for each genre and we decided to perform confidence interval tests for each genre. With the exception of Fantasy and Science Fiction, none of the 95% confidence intervals overlapped. That is, when we generated with 95% confidence intervals within which each population mean should be, it was very clear that the means fell into a definite order because, for the most part, the intervals did not overlap. We performed a two-sample T-test to determine whether Fantasy had a higher mean page length than Science Fiction. Our test did not provide enough evidence to reach a conclusion and our final order of page length, from increasing to decreasing, was determined to be Science Fiction/Fantasy (tied), followed by Mystery, followed by Romance, followed by Historical Fiction. Thus, the original question in our study—how exactly page length was related to genre—has been answered by the creation of a hierarchy of genres based on page length. Sadly, British people did not figure into our conclusion. This was disappointing.