Methodology and Analysis
First Round of Analysis
The methodology I employ is by looking for relationships between the three elements that I’ve highlighted above; The first round of analysis I do is by ranking and dividing the school districts’ “Language Speaking Abilities” data into four quartiles (labeled A, B, C, and D). This ranking is weighted, meaning that the first quartile will represent the districts with the highest ratio of bilingual students in the state, and the fourth quartile will represent the districts with the highest ratio of monolingual students. I use all four quartiles to conduct my next round of analysis.
Second Round of Analysis
The second round of analysis involves dividing the data once more, but this time it is based on some conglomeration of the data under the “Family Income and School District Wealth”. The second round aims to evaluate the degree to which financial wellbeing may be playing into the academic performance of students. At this point, I evaluate my findings to see if they follow a reasonable criteria: for instance, I would expect that school districts in counties around major metropolitan areas such as Houston, Dallas, El Paso, San Antonio, Austin, and Fort Worth would have a higher number of net revenue – therfore, I divide the total revenue for each quartile by the total number of students and English Learners in that quartile to get a "Net Revenue/English Learner". I realize later on that the Texas Education Agency has publically available data on this – more specifically, it has public data on % of School Revenue that goes into ESL/Bilingual Programs and the total school investment/Pupil in that school – I compared publically available data to my own findings, and found that there were negligeble differences between the two end results, which enforced my confidence in my findings. This encouraged me to be more original with my findings so I shifted my methodology: I decided to group the school districts by % English Learners in that school, and then conglomerate the total % of English Learners in a school to encompass the total % of English Learners in that district and in that county. Essentially, I broadened my calculation range to encompass the entire state as a whole rather than focusing on individual school districts alone; however, I still used a school district's specific financial data to create groupings within the entire state so that I could compare how school districts with similar levels of school funding in the entire state compare to one another.
As previously mentioned, this step is generally ignored in current academia, and is a huge proponent of language learning and academic success. Many researchers either 1) look at how investing more into school systems affects academic performance on standardized exams, or 2) only look at how English speakers – regardless of financial wellbeing – perform on standardized exams. This approach can be problematic because the current state of school funding in Texas is one which rewards already excelling schools by giving them more money, whereas unsuccessful schools do not receive more money to improve their programs. This creates a constantly recurring cycle which harms schools with higher percentages of English Learners who don't perform as well on standardized exams. Therefore, it is important to include as much data points as possible to best represent the financial wellbeing of the students in each of the districts.
Third Round of Analysis
Finally, the third round of analysis involves individually dividing the groups of data from the previous round based on a combination of data under the “Academic Performance” element. This is to evaluate how the districts with the highest number of bilinguals compare in academic performance to the districts with the highest number of monolinguals. I initially divided the schools into groupings of four, but I later realized that this created categories that were far too large to extract any meaninful information; additionally, divying up my data by four (25%) created an issue: the majority of schools were in the 25%-75% EL range, with only a few school districts falling in to the >75% range. Instead, I decided to divy up my data based on increments of 5%, which showed up me how schools with additional 5% English Learners performed on the STAAR, the SAT, and the ACT.
The file below shows my initial draft of approaching the data – as highlighted above, it was flawed for a variety of reasons; one of the biggest flaws was that creating 64 groups of students meant that some groups had overwhelmingly large sample sizes and others were far too small. Therefore, I chose to incrementalize my English Learners % data in groups of 5%, up until the 65% range, where the sample size was getting exceedingly too small, so I grouped all school districts with more than 65% EL into one category.