A critical factor in the design of tagging experiments and choice of methods is the manner of recovery. Fisheries managers often state that returns from fishermen are critical to the success of the project, and that tags must therefore be large and colorful. However, such tags introduce an unknown and potentially serious bias on the survival, growth and behavior of the animal. Second, fishermen may not notice even large tags, they may forget or not get around to responding, or they may choose not to report from concern of being further regulated. In one study, fish in anglers' catches were secretly tagged after capture. In spite of rewards being offered for return of tags, only 29% of the tags were reported (Green, et al., 1983). Similar studies, utilizing "seeded" catches have consistently shown significant, but unpredictable, levels of non-reporting of external tags in commercial and recreational fisheries. Even if all recaptures are reported, it is difficult to establish the true size of the "sample" of which the tagged recaptures formed a part.
A more robust statistical approach is for the researcher to scan samples of fish catches for marked individuals. Such sampling can be planned and appropriately stratified to address specific questions and to obtain reliable and unbiased answers. Although this approach involves an additional phase in the project it can be highly cost-effective since a reasonable volume of robust data may be much more valuable than a larger volume of doubtful validity. The NMT fish marking systems are all particularly suited to this latter approach, although visible implants and coded wire tagged salmonid fish, "flagged" by adipose fin clipping, can be noticed and reported by fishermen.
A useful discussion of these and other aspects of fish tagging programs is provided by Bergman, et al. (1992).
For information on selecting the tagging system most suited to your needs see: Matching the Tag to the Program.