Building a Survey Methodology

¶ … school district wants to study the effectiveness of the new curricula, so the research question is "what is the effectiveness of the new curricula." This requires before and after data, where the new curricula is the independent variable. The school can therefore only really work with data that it had before implementing the new curricula. Presumably, this includes ethnicities, ages and grade, which are the most important variables. In this research, the independent variable is the new curricula, and the dependent variables will be the grades.

Age and ethnicity will be lines along which the populations are drawn for this study. The null hypothesis is that the new curricula is correlated with a significant increase in grades for ESL students. The alternate hypothesis is that the new curricula is not correlated with a significant increase in grades for ESL students. The broad population is the entire set of ESL students in the district.

The research will hopefully be able to get more refined data, however, that allows the school board to understand if there are differences between performance among students of different ages, and linguistic backgrounds. This is because people from some linguistic backgrounds are likely to pick up English more quickly -- in general Indo-European languages are more similar to English than languages from other language families. Moreover, students of different ages are likely to learn at different rates -- younger students likely learn English more quickly than older ones.

So subsequent hypotheses will reflect these ideas -- whatever the research shows will be the null and alternate hypotheses. But for the main hypothesis, the entirety of ESL students is the population. Data is available for all students, and the data set is by no means onerous, so sampling is not necessary. Validity is improved when the entire population is used. Moreover, the likelihood of being able to learn facts about sub-populations is greater when the entire population is used -- the numbers of each sub-group will be greater.

For example, if the sample yields 10 speakers of African languages, it may be difficult to draw conclusions about their responses to the software as distinct from the aggregate response, but if the total population of speakers of African languages is 90, then that number is more likely to yield statistically significant data about how those students respond to the curricula. The data will be the demographic data of the students, and their grades. The grades will be in various courses, and an average grade.

What form this data takes in the study will depend on what form it has been recorded already -- remember we need to compare this with the historic data that was taken years ago prior to the implementation of the new curricula. If letter grades are used, these will need to be operationalized in some way, such as GPA. If numeric grades are used, they likely do not need to be adjusted in order to be used in this study.

If GPA is used, then conversion to this scale will make it easier to determine percentage changes after the new curricula was introduced -- those percentage changes are essential to determining the effectiveness of the new curricula. Of course there needs to be protection, in.