Look at the data they’re pulling this from. It’s not numbers of staff members that left but calculated based on the share of salaries that turned over. Someone could could be overturning their junior staffers far more frequently whereas she’s overturned a few CoS, bringing up her turnover rate.