add bootstrapping method of raised data

[mebrooks]

1. resample hauls
2. resample individuals observed within hauls
3. raise resampled individuals according to sampling fractions or raising factors (requires new input format other than qratio)
4. refit model to raised numbers

This assumes that larger hauls should have more influence on the fit, even if all fish weren't observed. This may be based on an assumption that selectivity changes systematically with total catch; thus it may be possible to achieve similar results by including total catch in a fixed effect.