40 



Fishery Bulletin 100(1) 



between sites within the same estuary was unexpected. 

 In addition, variation between estuaries in the same 

 state made comparisons between states difficult. However, 

 "reward" tags were returned less often than "$100 reward" 

 tags from 11 of the release sites in the unpartioned data 

 set. After identifying and excluding possible sources of 

 bias, we found that there were statistically significant dif- 

 ferences between reporting level of "reward" and "$100 

 reward" tags in all areas (Table 4). The range of 19-82''i in 

 levels of reporting between sites was more variable than 

 anticipated (Table 4). Removal of the suspected biased 

 anglers from the data set resulted in a mean unbiased 

 reporting level of 67.1'~f in Charleston Harbor and 44.8'f 

 in Calibogue Sound (Table 4). Unbiased reporting in GA 

 was somewhat higher than in SC (63.4'7f vs. 56.7'"*). The 

 fact that significant differences were found only after 

 biased angler data were removed from the data set illus- 

 trates that a small number of skilled anglers can have an 

 effect on fisheries-dependent data. Their failure to report 

 tags may be due to a lack of novelty in encountering 

 tagged fish, or to insufficient reward incentives (having 

 already received a number of t-shirts, fishing caps, etc. I. 

 These data suggest that use of noncash rewards is ben- 

 eficial only for the first time an angler catches a tagged 

 fish and decreases as anglers catch additional tagged fish. 

 Further repeated exposure to tagging programs within 

 each state eventually results in angler ambivalence and 

 reduced cooperation. This indifference is of particular con- 

 cern with the use of a constant regional reporting rate as 

 described by Hocnig et al. (1998). A decreasing rate of tag 

 return could be mistaken for lower hai-vest, reduced fish- 

 ing effort, poor survival, or increased population size. 



Lack of differences in reporting levels between "reward" 

 and "$100 reward" in the single-return (one fish) parti- 

 tion of data (Table 3) confirms that anglers who capture 

 many tagged fish per trip or per season (who were omitted 

 from this data set) significantly influence reporting. Sin- 

 gle return-data also suggest that anglers who catch fewer 

 fish (tagged or not tagged) are more likely to report cap- 



tures of tagged fish regardless of reward amount. Consid- 

 ering the impact a few skilled anglers can have on tag re- 

 porting estimates, these results demonstrate the need for 

 further evaluation of the interaction between tagging pro- 

 grams and angler behavior. The 50*7^ reporting level cur- 

 rently used by managers is approximately a IT^'i under- 

 estimate (.50/60=0.83) of actual reporting recorded for the 

 red drum fishery in SC and GA. Continuing to use the 50'7( 

 reporting estimate for this fishery will be more conserva- 

 tive than using the actual reporting level (A) to calculate 

 angler recovery rate (H). Reporting was also extremely site 

 specific, and application of data from one site to a broader 

 area may not be appropriate. Ideally tag-recapture models 

 should be weighted by site-specific reporting information 

 to account for this variability which could be accomplished 

 by regular deployment of high value (>$100) reward tags 

 within each system to gauge angler reporting. Even if of- 

 fering a $100 does not result in lOO*^? reporting, as Nichols 

 et al. ( 1991) suggested, it may yield the highest reporting 

 possible with monetary incentives, meaning that our unbi- 

 ased reporting may have been slightly overestimated. Re- 

 gardless, this approach is still more accurate than that of 

 adopting a regional average. Our results emphasize that 

 researchers need to conduct controlled tag reward studies 

 regularly and also to offer sufficient rewards in order to 

 avoid under reporting. Furthermore, tag reports must be 

 followed up with angler interviews to determine attitudes 

 and give managers an opportunity to remove bias from the 

 data (Reinecke et al., 1992. Zale and Bain, 1994, Pegg et 

 al.,1996). 



Acknowledgments 



We would like to thank the staff of the Inshore Fisheries 

 Sections of the SCDNR and GADNR for tagging, distri- 

 bution of fish and tag collection and processing. We espe- 

 cially thank John Fortuna and Carolyn Belcher for their 

 assistance with statistical design and data analysis. We 



