Monday, November 23, 2015

Yes, Virginia, there is a Correlation.

There has been some contention recently about a fairly straightforward question, namely “Is there a correlation between school funding and student outcomes?”

The answer is yes, and in this blog post, I am going to provide evidence to support this answer.

But before that, let’s make sure we all understand the question.  A correlation, according to Merriam-Webster, is “a relation existing between phenomena or things or between mathematical or statistical variables which tend to vary, be associated, or occur together in a way not expected on the basis of chance alone.”  

Therefore, to say two things are correlated is to say they tend to vary together.  When one is higher or lower, the other tends to be higher or lower. The correlation coefficient tells us the degree to which these things are correlated. A 100% correlation means that the two variables move perfectly in sync.  

Back to our question, what we are really asking is: “Do student outcomes tend to be higher/lower when funding is higher/lower?”  

Again, the answer is yes.  And I will show you by listing correlation coefficients for a variety of funding measures compared to a variety of outcome measures:

  • Total Revenue per Pupil
  • Current Spending per Pupil
  • Spending on Instruction per Pupil
  • Spending on Instruction as Percent of Current Spending
  • NAEP assessment results
  • ACT assessment results
  • SAT assessment results

This is all publicly available state-level data, and anyone can take a look for themselves.  The funding amounts come from the U.S. Census Bureau.  The adjustments for state cost of living were calculated using the Regional Price Parity statistics from the Bureau of Economic Analysis.  The NAEP data comes from the Institute for Education Sciences, and the ACT and SAT results come directly from the tests’ websites.  I used data across several years for all fifty states.  

Here are the results I got:

So, with the exception of the SAT, we can say that across all these student outcome measures, more money is correlated with better student outcomes.  We can also say that a higher percent of current spending going towards instruction is correlated with better student outcomes, but these correlations are not as high as those for the dollar amounts.

Does this mean increasing funding for education will lead to better student outcomes?  

No, it does not.  

The mantra that I have made everyone here at KASB who works with this data memorize is:  “Correlation is not causation.”  The fact that two things tend to move together does not mean one causes the other.  

With school funding and student outcomes, it is virtually impossible to prove via statistics that increasing funding will improve student outcomes.  But then again, it is virtually impossible to prove ANYTHING via statistics unless you are working in an environment where you have complete control over all your variables.  

School funding and student outcomes are also correlated with a lot of other things, like median household income, education level of adults, population per square mile, percent of students with disabilities, percent of students eligible for free and reduced-price lunch, percent of students eligible for ELL services, and so forth.  So to try and separate out the impact of funding on outcomes by itself is not something that can be done in a clear and concise manner.  Using different statistics for different time periods will yield different results, which is why people on both sides of the school funding argument continue to find new ways to present the data and argue their cases.

But let’s take a step back.  Do we really need statistics to tell us that if you increase funding to schools, schools are going to have a better chance of enabling students to succeed?  

You tell me.