Discussion
Hmm...let's grab a cup of coffee and discuss this...
Weaknesses! Ugh!
The study contained many weaknesses…unfortunately. Here they are:
- Out of the seventy two people sampled for the survey, fifty five responded. Even though this is a good amount, the seventeen students who did not respond constitute nonresponse bias. The net effect of this nonresponse bias is that the students who did not respond could have made a significant difference in the paired t-test. Perhaps the null hypothesis might not have been rejected if all of the students surveyed had responded.
- Furthermore, there are many confounding variables in this study. The effect of these confounding variables cannot be distinguished from that of the explanatory variable. Some of the factors that had a potential to confound include the grade of the student. Students in lower grades might be less inclined to read books than those in higher grades. The English class taken by the student is another factor. A student in higher level English, for example an honors class, as opposed to a student in a lower level English class, such as a general class, might be more inclined to read books than watch movies. Another factor is gender. Girls maybe more inclined to read books than boys or vice versa. And the list goes on.
- The directions on the survey were not clearly phrased. The direction read, “Please circle the appropriate number according to the index above for each item. Circle only one number for each letter choice.” Many students misunderstood this and circled only one number for each number choice as opposed to circling one number for each letter choice. The poor word choice could have interfered with the results of the survey.
- There was also undercoverage involved in the study. When the sample of students was chosen randomly, they were not selected from the complete list of students at North Olmsted High School. All the PSEOP students and students who are not full-time students at the high school were excluded from the list. Hence, this group of students was not represented by the chosen sample at all.
However many weakness the study might have had, it is still a valid and fairly reasonable one because of the large sample size. I feel comfortable extrapolating the results of the study to all the students at North Olmsted High School, except for those students who are not full-time students here. This is because the random sample of students was obtained from a list of all the students who are attending NOHS full-time. The random number generator in Minitab was used to select the students and hence we can say that the sample is fairly representative of most of the students at NOHS.
Suggestions! :)
This study can be improved if the following are done:
- Instead of a simple random sample, a stratified random sample should have been used. The strata of course would be the students’ grades. By doing this, one of the confounding variables, the student’s grade, can be avoided.
- Instead of just one sample, I should have taken four to five samples from the population. This replication will guarantee that the results produced by the samples are not the result of a fluke but instead they are directly affected by the explanatory variable (book or movie).
- The students excluded from the list should have been included in the list. They were excluded because there was no way to contact them. Perhaps they could have been contacted via mail. This would eliminate undercoverage.
- Furthermore, the instructions on the survey should have been phrased more clearly and succinctly so that there would be no confusion for the student being surveyed.
Further Study...
Many other studies can be done on this topic. For example, one of the confounding variables was gender. Another study comparing the results between male and female students can be conducted to see of the average approval score for books is higher in one of the groups.
Or another venue can be explored by looking at the relationship between the level of English class the student is taking and the student’s average approval score for books.