|
CRD 01-15 Table 1. Assumptions for three methods that
can be used to estimate the bycatch rate of coastal bottlenose dolphins
and if it was found to be valid. |
| Assumption |
Tested Valid |
|
Ratio Method |
| 1. The relationship between the numerator (number of dolphins) and denominator
(landings kept) of the ratio estimator is a straight line through the origin. |
No |
| 2. The variance of the landings kept about the above line is proportional
to the number of dolphin takes. |
Yes |
| 3. The CV's of the numerator and denominator should be less than 10%. |
No |
|
The sample size within a stratum exceeds 30. |
Yes |
|
Delta-distribution Method |
| 1. The data can be easily divided into events with a zero bycatch rate
and a positive bycatch rate and the data are dominated by a large proportion
of zeros. |
Yes |
| 2. Positive bycatch rates follow a log-normal distribution. |
Yes |
|
There are no small departures of the log-normal distribution, especially
for small positive values. |
No |
|
The sample size of positive events is greater than 15 within a stratum. |
No |
Log-linear Regression Method |
| 1. The bycatch rates follow a Poisson distribution, or |
Yes |
| 2. The variance in the bycatch rate is constant over the range of the
bycatch rate values. |
Yes |