We examined younger and older adults' item selection behaviors to assess heuristics for self-regulating learning of English meanings of Chinese characters varying widely in figural complexity. Two study-test trials were used to assess whether (a) item selection behaviors on the first study opportunity would show evidence for a difficulty-based heuristic as posited by Metcalfe's (2002) region of proximal learning (RPL) theory, or alternatively, influences of habitual English-language reading order (i.e. left-to-right, top-to-bottom); (b) whether second-trial selection behaviors were better predicted by RPL or by the discrepancy reduction model (DRM; Dunlosky & Hertzog, 1998); and (c) whether Trial 1 test performance would alter Trial 2 study in a manner predicted by RPL. DRM stipulates people select any item for study at Trial 2 that was not previously recalled. RPL states that people study only items they believe they can learn, avoiding complex items above their subjective RPL. Stimuli were 36 Chinese-English vocabulary pairs randomly presented in six 2 × 3-element grids. Both habitual reading order and stimulus complexity at Trial 1 affected order of study, with participants of both age groups manifesting a preference to study less complex characters. However, older adults showed larger effects of stimulus complexity whereas younger adults had larger effects of habitual reading order. At Trial 2, older adults showed a greater tendency to avoid studying the most complex Chinese characters, consistent with RPL, which contributed to their lower rates of vocabulary acquisition. Older adults' more conservative RPLs appeared to constrain their degree of self-regulated learning. (PsycInfo Database Record (c) 2020 APA, all rights reserved).