Agreed they don’t really address threats to validity (HBR almost never does). Another objection - they claim that adjusting salary sooner rather than later has a big effect on turnover, but do they control for the general effect that getting a raise or not affects turnover rate, independent of whether someone was hired at a higher salary?