Once all the respective surveys are paired, a match paired t-test will be used to test whether students, as a whole, showed significant growth in any one area. For example, for Teen AA Survey question 5, all of the students' initial and post scores for that respective question will be compared, and then analyzed as a group to see if there was significant growth relating to that question. In this case, the research will tell whether or not students "feel more confident in knowing there is a 'safe place' for dealing with issues" as a result of attending Teen AA meetings. Similar tests and analyses will be made from the remaining nine survey questions.
Conditions of the Research
Since it is the assumption that Teen AA produces positive behaviors amongst its participants, the matched pairs t-test will have a null hypothesis where the average difference in scores (Initial Score - Post Score) is zero, and an alternative hypothesis suggesting that the average difference in scores is greater than zero, or less than zero, depending on the question. Significance levels for each question will be with 95% confidence. It will be also assumed that the sample of scores for each question will be no less than 40. Tests for outlier and graphs for skewedness will be made for each set of questions as well.
Limitations of the Research
While it is the hope that the Teen AA program provides a strong intervention for helping teenagers with their dependency with addiction, the results cannot be recognized as 100% generalizeable for the following reasons. First, the participants are not from a random sample; they are volunteers. Ideally, it would more substantial to pool random subjects from a collection of all students who are entering the Teen AA program. However, the program is limited by the number of participants within his population. Therefore, all participating students will be asked to fill-out a survey in order to obtain some judgment toward statistical inference.
Second, the Manchester scale is a finite scale where differences in scores can only range from -10 to 10. It's possible that the distribution of score differences may appear abnormal. Finally, other outside variables may have an influence on the positive behaviors from the participating students. For example, the motivation for attending such a meeting suggests that one is seeking help and is ready to work on personal change. One may think it's the drive of the student that is most influential with conquering the disease as opposed to the intervention alone. Regardless, with a large group of subjects, the results of the survey will provide evidence as to whether students are benefitting from the Teen AA intervention. And since addictive behavior is a universal disease, it is believed the research results found within the initial group of students can be generalized to most students under similar circumstances (demographics, age, race, economic status). Future research can include comparisons of students of differing subpopulations.