Douglas Willms gave the keynote address last night here at CIES and spoke about increasing both equity and achievement ("raising and leveling the learning bar"). For every country that took part in two international assessments (PISA and PIRLS) he constructed a "learning bar" that showed the relationship between socio-economic status (SES) and achievement. SES was a significant predictor of achievement in every single country, but the slope of the line is different across nations.
For example, Canadians, on average, outscored Americans. But the wealthiest Americans outscored the wealthiest Canadians by a bit (meaning the poorest Canadians outscored the poorest Americans by a lot). This means that the American "learning bar" has a steeper slope -- that SES matters more in the U.S. than it does in Canada. The report of most of the results he discussed can be found here.
He makes a distinction between "raising" this learning bar -- meaning that everybody scores higher -- and "leveling" the learning bar -- meaning that low-SES students improve more than high-SES students, the line becomes flatter, and there is less of a gap between rich and poor. Here's what I found most interesting: his findings dovetail nicely with mine -- countries that scored higher overall also had lower gaps between rich and poor. Additionally, higher scoring countries have distributions that are less negatively skewed -- meaning that there are fewer very low scoring students. In other words, there doesn't seem to be a tradeoff between "raising" and "leveling" the "learning bar" -- or between aiming for high achievement and high levels of equality.