Independent recollection-familiarity (RF) ratings are sometimes collected to measure subjective experiences of recollection and familiarity during recognition. Although the RF ratings task purports to measure the degree to which each recognition state is experienced, the rating scale has been worded in terms of confidence rather than amount. Given prior evidence that wording influences recognition and remember/know judgments, we compared RF rating scales worded in terms of amount versus confidence across 2 groups. A robust levels-of-processing effect occurred on both recollection and familiarity ratings, and its magnitude was similar across scale wording. Scale wording did not influence recognition, and, most importantly, it had little influence on ratings of recollection and familiarity. These findings suggest that participants may use confidence to rate amount, or vice versa. Regardless, researchers should align their task instructions and scale wording, and should publish them. Such alignment and transparency is crucial for interpreting measures of the memory states that arise during recognition memory. (PsycINFO Database Record (c) 2019 APA, all rights reserved).