Premium
Considerations for Conducting Field Experiments with Baited Traps
Author(s) -
Miller Robert J.
Publication year - 1983
Publication title -
fisheries
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.725
H-Index - 79
eISSN - 1548-8446
pISSN - 0363-2415
DOI - 10.1577/1548-8446(1983)008<0014:cfcfew>2.0.co;2
Subject(s) - logbook , trap (plumbing) , fishing , trapping , statistics , environmental science , data quality , computer science , fishery , mathematics , ecology , biology , operations management , engineering , environmental engineering , metric (unit)
The uncontrolled variables, the large number of variables potentially testable, and the exigencies of field operations are all distractions which can compromise the quality of field experiments. Catch records from fishermen's logbooks are not good substitutes for experimental fishing. They rarely, if ever, meet experimental assumptions on randomization of treatments over space and time, and the quality of the recorded data can be wanting. A review of studies on catch per baited trap versus soak time gave peak catches at 3 to 10 days when data were taken from logbooks, and 0.1 and 2 days when data were taken from experimental fishing, suggesting a systematic bias in the logbook data. Experiments with baited traps usually measure effects of soak time, trap design, or bait type on catch per trap. In these cases preliminary trapping is important for locating an area of approximately uniform catch rates (uniformity trial), to determine the necessary sample size for desired precision, and to solve logistical problems. Variability among replicates is minimized by choosing an area of uniform catch rates and by using uniform technique with regard to bait quality and quantity, soak time, and trap quality. Fishing a few traps carefully is preferable to fishing many uncarefully.