Discussion
WEAKNESSES
Due to the facts that each person is directed to fill out the survey independently and the wording is straightforward, it is my contention that a confounding variable is not a threat to my study. However, if students in classes decide to fill out the survey based on what is socially “cool” or acceptable instead of making an accurate participation evaluation, than this extraneous factor could influence the results. Also, the fact that there is a large amount of response bias might have had a significant effect on the outcome of the x2 test for homogeneity.
EXTRAPOLATION
This study would be more accurate if the population was not confined to specifically NOHS students. Because participation is not effected greatly by a change of location, climate, or other geographical nuances, the sampling population could be extended so as to include all students in northern Ohio or even the entire United States. However, if the population of interest was any greater than the United States, a test on this data would be less accurate due to conflicting customs.
FURTHER WORK AND SUGGESTIONS
In order to get more accurate results, there are several things that will reduce response bias in my sample. One possible tactic would be to individually select 150 people and call them or interview them as opposed to sending them surveys. This method would assure that they would respond more accurately and would eliminate some non-response bias in the sampling. Also, sampling from a list of non-Polaris students or purely those students with a math and English course at NOHS would help to reduce non-response bias. These changes might improve the study even if they do not alter the outcome of the x2-test.