[81alum]

Yes, the evidence has been fiercely debated for decades.  Here is a link to a 2014 PBS story that links a very good BATEs study:

http://www.pbs.org/newshour/rundown/nail...-students/

If you look at the BATEs study you will find a pretty good bibliography to other information as well.  The BATES study was a little different--since SAT scores were optional in its admissions, it compared college success among those who did and those who did not submit SAT scores.  For those who did not submit SAT scores, GPA was the predominant academic measure. 

“The evidence of the study clearly shows that high school GPA matters. Four-year, long-term evidence of self-discipline, intellectual curiosity and hard work; that’s what matters the most. After that, I would say evidence that someone has interests that they have brought to a higher level, from a soccer goalie to a debater to a servant in a community to a linguist. We need to see evidence that the student can bring something to a high level of skill,” Hiss said.According to the data, if high school grades are not high, good testing does not promise college success. Students with good grades and modest testing did better in college than students with higher testing and lower high school grades.
“The human mind is simply so complex and so multifaceted and fluid, that trying to find a single measurement tool that will be reliable across the enormous populations of American students is simply a trip up a blind alley. I would never say the SATs and ACTs have no predictive value for anybody; they have predictive value for some people. We just don’t find them reliable cross populations,” says Hiss.

This part of the write-up somewhat supports your point about the history and intentions behind the SAT, but notes that times have changed:

But the dreaded test was born of good intentions.The SAT started in the 1930s as a scholarship test for Ivy League schools. Based off of an Army IQ test, it was meant to help those who came from more humble backgrounds to be noticed by prestigious schools. Many other universities followed suit.
But standardized testing may now be hurting rather than helping disenfranchised students.
The study found that non-submitting students were more likely to be minorities, women, students with Learning Differences, Pell Grant recipients and first-generation college-goers. 

By the way, there are now 850 test-optional colleges.  There are just over 3000 4 year degree granting institutions, so this is a significant minority and is growing.

[81alum]

Yes, the evidence has been fiercely debated for decades.  Here is a link to a 2014 PBS story that links a very good BATEs study:

http://www.pbs.org/newshour/rundown/nail...-students/

If you look at the BATEs study you will find a pretty good bibliography to other information as well.  The BATES study was a little different--since SAT scores were optional in its admissions, it compared college success among those who did and those who did not submit SAT scores.  For those who did not submit SAT scores, GPA was the predominant academic measure. 

“The evidence of the study clearly shows that high school GPA matters. Four-year, long-term evidence of self-discipline, intellectual curiosity and hard work; that’s what matters the most. After that, I would say evidence that someone has interests that they have brought to a higher level, from a soccer goalie to a debater to a servant in a community to a linguist. We need to see evidence that the student can bring something to a high level of skill,” Hiss said.According to the data, if high school grades are not high, good testing does not promise college success. Students with good grades and modest testing did better in college than students with higher testing and lower high school grades.
“The human mind is simply so complex and so multifaceted and fluid, that trying to find a single measurement tool that will be reliable across the enormous populations of American students is simply a trip up a blind alley. I would never say the SATs and ACTs have no predictive value for anybody; they have predictive value for some people. We just don’t find them reliable cross populations,” says Hiss.

This part of the write-up somewhat supports your point about the history and intentions behind the SAT, but notes that times have changed:

But the dreaded test was born of good intentions.The SAT started in the 1930s as a scholarship test for Ivy League schools. Based off of an Army IQ test, it was meant to help those who came from more humble backgrounds to be noticed by prestigious schools. Many other universities followed suit.
But standardized testing may now be hurting rather than helping disenfranchised students.
The study found that non-submitting students were more likely to be minorities, women, students with Learning Differences, Pell Grant recipients and first-generation college-goers. 

By the way, there are now 850 test-optional colleges.  There are just over 3000 4 year degree granting institutions, so this is a significant minority and is growing.