30th July 2010, 12:24 PM
Limiting it to RAOs might well insert a whole new level of skewing into the data, a disproportionately larger number of smaller and very small outfits are not RAOs (and are quite likely to have been missed from any of the original data anyway), so you'd be reflecting much more the employment performance of the larger (and, dare I say it, southern) units. Any statistician/market researcher types out there?