responses required

I think while the statistical correlation between the two is described as "significant," I am hesitant to say that it is the direct cause of performance on the SOL.  The quote mentions that .75 is statistically high in the testing world, but even a high correlation doesn't account for all factors impacting students' SOL performance.  I think this high correlation helps to make an educated assumption but I think it's impossible to say it is the direct cause...to me that would mean the significance is not just high, but absolutely certain. 

Do you think SOL performance is causing a change in SGA performance?

It would seem, based on the study, that SGA performance is helping to predict SOL performance.  I would not agree that SOL performance has the ability to change SGA performance. To say that it is implementing change in SGA performance, I think, is a leap. 

Do you think performance on the SGA is contributing but not the sole cause of performance on the SOL?

I think this question is a better way of looking at the statistical significance between the tests than question 1....as I mentioned before, even though the correlation is high...it's impossible to say with certainty that a high score on one test will equal a high score on the other. It's impossible to say (in my opinion) that one is the direct cause of the other. 

What confounding variables could be responsible for the observed relationship between SGA and SOL?

Confounding variables impacting the relationship between the two tests could be infinite in this situation.  In my opinion, the state a student is in when taking each exam has the biggest impact on their scores.  From more abstract factors like is a student more comfortable taking an SOL after already experiencing taking an SGA, to more concrete like, how much they slept the night before - can all impact a student's score from one test to another. 

Could performance on both SGA and SOLs result from a common cause?

I think this is certainly possible.  I think this is easier to answer for smaller sections of reported scores, say per region.  There could be more funding in certain geographical locations and scores on each exam from those locations could be similar.  When taking all scores for both exams into consideration, I think it's much harder to say whether or not there is a common cause for scores between the two.  Also, discussing cause for performance is much different than discussing the probability that performance on one exam is an accurate predictor of performance on the other. 

Since both SGA and SOL performances change over time, what does that say about their relationship?

I think as performance on the two exams changes over time, the statistical significance of their relationship needs to continue to be reevaluated. Clearly at this point in time, the correlation between scores is high, but it's difficult to say how performance changes on either with effect their future relationship without having the data itself. 

Could the association between SGA and SOL be nothing more than coincidence?

I would say that the high statistical significance between the two is not something to ignore.  To say that it is a coincidence would be turning a blind eye to the data.  While it's important to take confounding factors into consideration, and to account for the fact that there a few outcomes we can predict with total certainty (in regard to testing), I do not think the association is a coincidence.