Combating the disconnect between actual and paid participation: a model to predict license purchase based on angler behavior and characteristics

An angler may purchase a fishing license but not actually participate in that activity; this is a discrepancy between paid and actual participation. For example, previous research indicated 56% of people who purchased a... [ view full abstract ]

An angler may purchase a fishing license but not actually participate in that activity; this is a discrepancy between paid and actual participation. For example, previous research indicated 56% of people who purchased a Kansas floatline fishing permit actually went floatline fishing.  The ideal situation would be for people to continue purchasing licenses and participating to avoid attrition or relapsing.  The objective of this study was to develop a model to predict the likelihood of an angler to purchase a license based on his or her confirmed fishing participation, demographic variables, and other characteristics.  The model was developed from a group of people actively fishing who were interviewed by KDWPT staff at Kansas State Fishing Lakes during the 2010 Memorial Holiday weekend.  Based on information anglers provided, I was able to data mine KDWPT’s license files to determine their license purchasing patterns.  To test the model’s performance, I used a sample of respondents from the 2013 Kansas Licensed Angler Survey for which I had similar predictor variables.  Explanatory variables that were significant included an angler’s sex, previous year’s fishing license purchase, and an angler’s participation in the current year.  The initial correct classification rate for the model was 81%.  The correct classification rate using the validation data set was 74%.  The model was successful because the significant predictors were those which were readily available in the license database and can serve as a useful tool for predicting paid participation (i.e., license sales) the upcoming year.