An investigative analysis of Queensland’s statewide recreational fishing surveys : Identifying bias in self reported catch and effort data

Abstract

Accurate recreational fishing information is essential to sustainable and effective management strategies. Several statewide recreational fishing surveys have been conducted in Queensland, yet their results have varied widely depending on whether the ‘RFish’ diary program or the interview-based ‘NRIFS’ approach was employed. The validity of RFish survey data has since been called into question, with concerns of possible nonresponse and recall biases. Estimates of catch and effort were compared between the two survey methodologies to search for evidence of bias in the RFish data and to explore the possibility of adjusting for any identified biases. Within each RFish survey, the high rate of dropout, gradual sample loss of particular demographics and steady decline in fisher avidity strongly indicates the presence of an attrition-derived avidity bias. The RFish surveys also exhibited an overall inflation of fisher avidity and catch success in addition to varying degrees of catch or effort overestimation for 11 of the 12 investigated taxa. Previous fisheries research would suggest that the likeliest explanation for these results is a recall bias arising from the RFish surveys’ considerably longer recall period. However, the true source of these errors is ultimately unknown and other possible explanations include prestige bias, deliberate misreporting, terminal digit preference, species misidentification and nonresponse bias. Regardless of whether the biases in the RFish data can be correctly identified, the significant variation in reporting error between taxa is largely unpredictable and may preclude any reliable adjustments. Re-expanding the raw RFish data to the statewide level using current statistical techniques will prove fruitless unless such adjustments can be made. Although adjustment may not be a viable option, the RFish data may be utilised to assess trends in relative rather than absolute terms. If other options are exhausted, the omission of RFish survey results may be the only way to ensure an accurate time series of statewide catch and effort in Queensland’s recreational fisheries.