Because low rate, covert responses are hard to observe and measure (e.g., Azrin & Wesolowski, 1974; Henderson, 1981; Jeffery, 1969; Reid & Patterson, 1976; Seymour & Epston, 1989), well-controlled behavior analytic investigations of stealing have been rare. In fact, systematic investigations to experimentally determine stealing functions have been limited to two studies targeting food (Lambert et al., 2019; Simmons, Akers, & Fisher, 2019). The dearth of studies examining stealing functions, partly attributable to low rate and covertness, may be forestalling additional intervention studies. Given the likely futility of unsystematic attempts to observe naturally occurring instances of an infrequent, clandestine response, a possible role for indirect assessment emerges (Iwata & Dozier, 2008). This two-part study concerned an investigation into the reliability and predictive validity of the Functional Analysis Screening Tool (Iwata, DeLeon, & Roscoe, 2013) and a similarly constructed tool (The Stealing Inventory or TSI) with the latter having questions oriented towards likely stealing functions. In doing so, the comparative viability of two trial-based functional analysis (FA) models (Bloom, Iwata, Fritz, Roscoe, & Carreau, 2011; Lambert, Bloom, & Irvin, 2012) was also examined. Across 42 respondent pairs, overall tool reliability and outcome reliability for suggested functions favored the TSI (85% and 92.9%, respectively) over the FAST (80% and 73.8%, respectively). Three out of 6 participants stole during one of their two respective FAs, and the identified function matched the respective TSI outcomes for each case. FA model superiority was unclear.