Either the effect exists or it doesn't, and you need to show that it doesn't if you want to claim that discriminating based on overqualification is not effective. And that study doesn't show that - it only shows that overqualified candidates that hiring managers thought wouldn't leave soon don't leave soon.