494 



Fishery Bulletin 101(3) 



Bycatch simulations 



For the normally distributed data, the CPUE-mean-per- 

 unit estimator was the only estimator whose estimated 

 bycatch was not significantly different than the actual sim- 

 ulated bycatch {% bias=0.006, P=0.94). All four of the ratio 

 estimators significantly overestimated bycatch (Table 6), 

 although the average bias was less than a 1% overestimate 

 for the grand F:S and grand CPE ratio estimators. The 

 basic F:S ratio estimator and the basic CPE ratio estimator 

 both overestimated bycatch by 300-4007^ (Table 6). Using 

 a model that included all main effects and all 2-way and 3- 

 way interactions in the ANOVA, I found that the CV of the 

 auxiliary variable (either shrimp catch or hours fished) was 



a significant main effect for all four of the ratio estimators, 

 but there were no significant main effects for the CPUE 

 mean-per-unit estimator (Table 7). Observer coverage was 

 also a significant main effect for the F:S and CPE grand 

 ratio estimators, but was not significant for the basic F: 

 S or CPE ratio methods. The grand F:S ratio estimator, 

 the grand CPE ratio estimator, and the basic F:S ratio 

 estimator all showed several significant 2-way and 3-way 

 interactions (Fig. 2), whereas the basic CPE ratio estimator 

 had no significant 2-way or 3-way interactions. The CPUE- 

 mean-per-unit estimator showed only two significant 3- 

 way interactions among variables, and observer coverage 

 occurred in both. The correlation between fish catches and 

 shrimp catches was a significant main effect for the basic 



