Interesting, but flawed. It specifically says it compared the results from grammar school areas such as Bucks, but found that the region had no higher scores than a non selective area.
So the study collated all pupils from Bucks (et al) schools results, not just the grammars?
Bucks has many, many more pupils going to secondary schools than grammars...far, far more children fail than pass the 11+, so then are placed in pretty crappy secondaries, with a lot of very poorly behaved children, who make it harder for those in the middle abilities to flourish.
If course there are some students who fly, even in a poor secondary, but a lot more who fail due to disruption.
This will negatively affect the scores overall.
It's also worth mentioning that grammars (such as in Bucks), attract not just those from the area, but those from neighbouring areas, who don't have grammars.
Grammars in some parts of Bucks are in really poor areas, and I don't believe that abolishing them will improve the chances for those who are too poor/unable to move into 'better' areas; it would simply remove the chances of brighter, but poor, children.