The process of selecting attributes for inclusion in choice experiments frequently involves qualitative methods such as focus groups and interviews. In order for a choice experiment to be successful and the results to be valid, this qualitative selection process is essential. It often lacks rigour and is poorly described, particularly in environmental choice experiments. We propose a meticulous attribute and attribute-level selection process consisting of a scoring exercise and an interactive discussion. This paper provides a case study describing how attributes and attribute-levels were identified and selected for the National Park Hoge Kempen in Belgium. We carried out four focus groups and thirteen semi-structured interviews with various park stakeholders to select attributes from six categories: the four categories of ecosystem services (supporting, provisioning, regulating, cultural), infrastructure, and land use types. The top-ranked characteristics were nature conservation, natural forests, biodiversity refuge, wetlands, landscape variety, heathlands, air purification, and education. Both the scoring exercise and the interactive discussion contributed to the attributes selected for the CE. Following these, an ultimate expert consultation stage is recommended to approve both the attribute and attribute-level selection. The semi-qualitative protocol proposed in this paper can help practitioners and demonstrates how the results guide choice experiment design. ; We would like to thank study informants, focus group participants and interviewees, as well as Tom Kuppens, Silvie Daniels, Janka Vanschoenwinkel and Michele Moretti of the Environmental Economics Research group of Hasselt University. Johan Van den Bosch, project leader at Regional Landschap Kempen en Maasland (RLKM), and Rolinde Demeyer from the Research Institute for Nature and Forest (INBO) have provided this study with valuable assistance and comments. Sarah Jeanloz was funded by the INTERREG IVB NEW program (Grant no. D1941/56200), Nele Witters ...