From the 2014 Annie E. Casey Foundation, KIDS COUNT Data Book. Percentage of childhood poverty vs. percentage of 8th graders listed as non-proficient in mathematics (higher percentage is worse), by US state. Source: aecf.org, 2014. Mouse over points to see data for each state in 2005 (green points) and 2013 (blue points).
While the overall correlation is for poorer performance with greater poverty there also appears to be a general trend for greater mathematical performance between 2005 to 2013...however, looking state by state this often seems to occur with a corresponding increase in poverty (many, but not all states move down (better math performance) and to the right (increase poverty) from 2005 to 2013).
What does this say about the actual relationship between poverty and performance (at least using this metric)?