Over the last three decades, there has been increased attention on the collection and interpretation of "response processing data" to inform claims of learners' knowledge and skills (e.g., see Ercikan et al., 2010; Kobrin & Young, 2003; see also, Leighton, 2004). Response processing data are perhaps most consequential in the development of validation arguments to support claims or inferences about test-takers (see AERA, APA & NCME, 2014). However, these data may also serve to inform test development and encourage dialogue on fair student assessment of 21st century skills. One challenge is that methods for collecting response processing data vary in non-trivial ways and must be employed judiciously depending on the cognition of interest being measured. Thus, the objective of the paper is to highlight how two methods -- cognitive labs and think alouds -- can and should be used adaptively to inform distinct goals in test development, validity arguments, and to further the credibility of claims made about test-takers.