This paper provides an overview of an extensive recreation use values database that was developed to investigate the presence of and effects of publication selection bias in this literature. The recreation use values literature has a long history of value transfer applications, which may be affected by publication selection bias. The database consists of 325 studies providing 2,594 estimates of value for over 26 recreation activities for the US and Canada. Preliminary analyses show that document type (journal, agency report, consulting report, dissertation, etc.) result in statistically different mean values. Similarly, motivations or the primary contribution of documents show that documents introducing efficiency in design or estimator provided statistically lower estimates of value than documents that present tests of biases or whose primary purpose is to present a new estimate of value.