The official rationale is that it reduces bias. A single 1-hour interview has a low signal to noise ratio. Two interviews is better. At four interviews a fairly

The official rationale is that it reduces bias. A single 1-hour interview has a low signal to noise ratio. Two interviews is better. At four interviews a fairly reliable signal emerges. Add a fifth interview so that you can train/calibrate new interviewers, take an hour for lunch, and you're at 6 hours.