Roy Osherove wrote "To read a good summary of that paper go to Phil's post and read it to the end."
To read a good summary of that paper go to the end of the paper and read the "CONCLUSIONS AND FUTURE WORK" section ;-)
"Test-First programmers did not achieve better quality on average, ... Writing more tests improved the minimum quality achieved and decreased the variation, but this effect does not appear to be specific to Test-First."
Thanks for updating with my analysis. It's an interesting report with some support for TDD, but also some fascinating data points that I think are troubling as well (that I think should have made it to the abstract).