Week One: Part Two

Hello everyone, and welcome back. Sorry about the delay in putting out this blog post, I meant to release it a few days ago, but it turns out that this post was harder for me to write than I thought. 

My original plan for this blog post was to go into excruciating detail on how my project came to this point. It took me a few drafts and false starts to realize that there was no way for me to make that interesting or anything more than a slog for me to write and for all of you to read. 

In lieu of that in-depth discussion, I will summarize things here. As I did more research and worked more with the data, the project kept expanding, with more and more variables being added in order to more precisely measure the impact of certain factors. 

That brings me to the main topic of this blog post, some of the major discoveries I have made this week. Now, just to be clear, the stuff I am about to discuss is by no means proven or certain — it’s unlikely, but I might have made some minor mistake somewhere in my analysis — but if I’m right, these observations could be incredibly important for my project. 

The first real discovery of the week, the one that really kickstarted a lot of the other important observations was the discovery of a model explaining over 50% of the variation in English scores with all the variables involved in the model being statistically significant. Just to drive home how surprising this discovery is, a model which can explain even 30% of the variation in a sample is considered extremely good, and in all the regressions and models I have created I have never had a model with explanatory values this high have all its components have statistically significant impacts. 

I found that English Pass Rates were impacted by the following factors:

1. Percentage Of The District Which Is African American
2. “Local Ability To Pay”
3. Median Household Income
4. Whether The District Was Urban Or Not

This discovery led me to wonder if these same factors impacted the average scores of districts on the English exam. The answer, surprisingly enough, was no. While some of these factors still had an impact on English scores, not all of them did.  

Instead of being impacted by the four above factors, English Scores were only really impacted by two. Those factors were as follows:

1. Percentage Of The District Which Is African American
2. Median Household Income

These two factors by themselves explain the majority of the variation in English Average SOL scores across Virginia.  

These findings are fascinating, and will likely form the backbone of my analysis of English Test Scores. These findings seem to indicate that the level of state funding for local education in a district has an impact on the percentage of students that pass the SOLs but has little impact on how well the students do above that level. The same is also true of whether a school is urban or not. 

I will continue analyzing these, as well as other, results in order to figure out the exact reasons as to why this is, but I just thought that these discoveries were weird enough that I should share them. 

If any of you have any ideas as to what might be causing this, feel free to write them in the comments. 

Without further ado, I’ll talk to you all again next week.