Researchers struggle with faulty study designs, flyspecking each other’s work, re-arguing decades of debate about jobs and income