Initial Research into Financial, Geographical, and Population Effects on Academic Achievement Part 1

I haven’t even done a lit review yet, so don’t look for much here other than some raw data and some preliminary tests on it. I am only using data from Tennessee.

Someone once asked me if money made an impact on education and how. The how is pretty easy. You can pay more, get your choice of employees, hire more assistants and support staff, fund software and other programming, and have better facilities to do it all in. In sum, C.R.E.A.M.

I didn’t know (and still don’t know) if anyone had done a study on that, but I thought it would be easy enough to look at that in Tennessee. I took the 2019 school report card data file and the TACIR Fiscal Capacity Index for the following data points: District Success Rate (percentage of all students who were proficient on TN Ready tests), ACT Composite Average, 2017-2019 Taxable Sales, and SY2021 Fiscal Capacity.

I chose sales tax base to have an idea of the county’s wealth, and I took the Fiscal Capacity since that’s how the state ranks county’s in terms of being able to fund education. I looked at Success Rate from 2018-2019 since that’s the last year we’ve had reliable data.

To see if there is anything going on there, I ran simple Pearson correlations for the following:

  • TN Ready Success and ACT Composite Average (r=0.82)

  • TN Ready Success and Sales Tax Earnings (r=0.36)

  • ACT Composite Average and Sales Tax Earnings (r=0.34)

  • TN Ready Success and Fiscal Capacity (r=0.39)

  • ACT Composite Average and Fiscal Capacity (r=0.36)

  • Sales Tax Earnings and Fiscal Capacity (r=0.99)

All of the above showed positive correlations. The highest correlations were between TN Ready and ACT (r=.82) and Sales Tax and Fiscal Capacity (r=.99).

Counties with municipal or special districts

There was some replication of data because the state data shows this information for counties, and in TN, there are several municipal systems within counties. For example, Gibson and Carroll Counties both have multiple municipal systems rather than have one consolidated county system. Blount County has a county system and two municipal systems (Alcoa and Maryville). When looking at the data, it is clear that the small municipal systems outperform the counties they are in. Out of curiosity, I did initially break down Blount County’s.

Blount.png

The disparity between Blount County Schools and the two municipal districts in terms of TN Ready Success and ACT Average could be somewhat shocking (to a lay person) considering all three of these districts are sharing the same funding base that is prorated based on Average Daily Membership (ADM, or enrollment) and Average Daily Attendance (ADA, or attendance rate). So looking at that led me to wonder how other county/municipal districts looked like.

In a word, they looked similar.

IMG_3268.jpeg

Having worked in one large county with no municipal districts and in two municipal districts in large counties, I have anecdotal evidence of what is happening here, but I am going to avoid using it if I can help it.

Square Miles and Population Density

One of the aspects of municipal and special districts that separates them from county districts is their general geography. For example, Elizabethton City Schools cover 9.6 miles while Carter County Schools cover 331.7 square miles. That is a lot of ground to cover. The population density between the two is 1,421.2 persons per square mile for Elizabethton and 129.2 persons per square mile for Carter County Schools.

This is obviously the same for every municipal/special district in comparison to the county districts in which they are located, thus I wanted to see to what extent square mileage and population density correlated to academic achievement for all school districts - not just counties versus smaller districts within them.

I did a Pearson correlation comparing square mileage and TN Ready Success Rate. It had a medium negative association (r=-0.33). The same was found when comparing square mileage and ACT Composite averages (r=-0.34). This is suggesting that the more square mileage a school system has, the lower their academic performance will be.

But what about population density? I compared population density to TN Ready Success Rate and found a large positive association between the two (r=0.46). The ACT comparison was even stronger (r=0.52).

Correlation is not causation, and there are many variables to consider when looking at achievement rates.

In Part 2 we will look at how different systems rank in terms of size, finances, and achievement.