Potential biases in angler diary data: The impact of the diarist recruitment process on participation rates, catch, harvest, and effort estimates

Angler diaries are common tools to collect recreational fishing data. The diarists can be recruited either from the general population or directly from the angling population, for example, by using angling licence registries. The recruitment process can lead to specific biases whose magnitude is largely unknown. The present study compared socio-demographic data as well as catch, harvest, and release rates obtained from diaries from German Baltic Sea cod anglers who were recruited from a list of angling permit holders (non-probability-based sample) with those who were recruited simultaneously during a probability-based representative telephone survey among the general population. The results indicated that recruiting diarists from the list of permit holders may be more successful in terms of participation rates than recruiting from a general probability-based population survey. Both groups of diarists were similar regarding their socio-demographic characteristics. Nevertheless, the differences between the two samples in avidity and between recalled and reported angling days suggest that sampling from the list of angling permit holders may increase the risk for avidity and recall biases. Catch and harvest rates were influenced by avidity, age and angling platform, release rates by angling season, angler residence, and number of caught fish. Catch, harvest, and release rates differed more between land-based and sea-based angling than between the two sample selection processes, suggesting that the angling platform should be given special consideration when extrapolating diary data to the population level. The low explanatory power of the regression analyses suggests that relevant factors beyond standard socio-demographic parameters that influence estimates of CPUE, HPUE, and RPUE were not captured in the survey. Future research should therefore focus on evaluating such factors (e.g., factors related to the human dimension) that could better correct for biases in estimates of catch, harvest, releases, and angling effort in recreational fisheries surveys.