Weaknesses and variables of the study:

The study contains many flaws and variables, many of which were not realized prior to and at the time of conducting the study.  The data collected for the study were confined to the books available at Porter.  Therefore, we may not get the same result had we performed our study elsewhere.  The genre Historical contains more children’s books than any other genre, thus, influencing the data.  The data also contain many outliers, as evidenced by the box plots.  While collecting data, we experienced several cases of nonresponse bias, where the library’s database failed to provide the necessary book lengths.  Another variable of this study is font size; each book’s font differs from another’s, and this difference influences the result.  The invalidity of the chi-square test for homogeneity of populations is another weakness of the study.  Although the test clearly indicates that the population for each book genre are not homogenous (with P = 0.000), many of the expected cell counts were less than five.

Extrapolation:

Despite of the variables and imperfections of our data, the relatively large sample sizes—n = 86 for Fantasy, Science, and Romance; n = 90 for Historical; n = 95 for Mystery—we can confidently extrapolate the result of our study to all of the books belonging to the five genres published by the East Coast at Porter library. 

Suggestions for future studies:

For a better and more confident result for this study, larger sample sizes are suggested.  Prior to collecting data, the samples should also be stratified according to library branch and should be confined to only adult fictions.  An even better study would compare word count with genre instead of book length.  This would eliminate the confounding variable, font size.