Understanding catchability—the fraction of a stock caught by a defined unit of effort—is crucial to using fisheries assessment data to index abundance. We conducted mark–recapture experiments to estimate catchability and evaluate standard boat electrofishing methods for assessing populations of Largemouth Bass Micropterus salmoides. We then used a resampling analysis to test for differences in bass CPUE (fish/h and fish/km) between two high‐density reservoirs and one low‐density reservoir and among surveys within each reservoir. We compared scenarios using surveys conducted only during (1) the standard time period (mid‐April to mid‐May) and (2) the entire assessment period (early April to mid‐June). We considered the percentage of significant differences in CPUE between the high‐density and low‐density reservoirs to represent statistical power (i.e., the ability to detect a difference in CPUE when a difference actually exists) and the percentage of significant differences in CPUE between surveys in the same reservoir to represent the false‐positive rate (i.e., the detection of a difference in CPUE when no difference in density exists). Catchability and CPUE were greatest and least variable during recapture events conducted during the standard period. The mean catchability of sub–stock length Largemouth Bass (150–200 mm) and memorable‐length bass (≥510 mm) was significantly less than those for other length categories. Statistical power exceeded 80%, and the false‐positive rate was generally less than 10% for sampling during the standard period at as few as six electrofishing sites. When including samples from outside the standard period, power was lower and the false‐positive rate was as high as 60%. Power and false‐positive rate were similar whether effort was measured in time or distance. Our results emphasize that standardized springtime boat electrofishing assessments validly index Largemouth Bass density and size structure. Received October 7, 2016; accepted February 2, 2017 Published online April 24, 2017

Electrofishing CPUE data are commonly used to index temporal trends in abundance in fish monitoring programs, but the reliability of this index requires the assumption that the fraction of fish stock caught per unit effort (catchability, q) is relatively precise and constant through time. We evaluated how fish species, season, and lake affected electrofishing catchability in Florida lakes using a field study. We used the field study results to simulate how variable electrofishing q affects statistical power and type I error rate (i.e., the probability of detecting a difference when in fact no difference occurred). Model selection showed that electrofishing catchability varied by species, season, and lake, and submodels showed that catchability varied by lake for Largemouth Bass Micropterus salmoides and by season for Lake Chubsucker Erimyzon sucetta, and was constant across season and lake for Bluegill Lepomis macrochirus. Our results revealed that statistical power decreased and the type I error rate increased substantially if q varied through time as we observed for Largemouth Bass and Lake Chubsucker. Type I error rates were well above the expected value of 0.05, reaching as high as 0.7 for Largemouth Bass and 0.55 for Lake Chubsucker at high sample sizes. This resulted because increasing sample size improves the ability to detect real changes, but also increases the probability of detecting spurious changes due to variable q (i.e., type I error). Thus, variable catchability hinders our ability to use CPUE data to index trends in fish abundance, and monitoring programs should consider how q varies before relying on CPUE data to index fish abundance.