Three modes of deliberate vocabulary study were investigated to determine how well they assisted learners' recall of the meaning of target concrete nouns. Two modes of tablet-based augmented reality, one context-independent (AR1) and one context-dependent (AR2), were compared with each other and with paper-based word cards (WC) in the deliberate study of three sets of nonwords representative of concrete nouns. An orthogonal Latin square design was used to counterbalance 39 participants. We hypothesized that both AR conditions would be more beneficial than word cards in terms of participants' ability to retain the meaning of the target words as demonstrated by performance on Yes/No and meaning-recall test items, and that AR2 conditions would be more beneficial as compared with AR1 conditions. Generalized linear mixed models revealed that both AR study modes significantly outperformed word cards. The context-dependent and context-independent augmented reality study modes did not significantly differ indicating that a visuospatial bootstrapping effect (VSB) was likely at play regardless of how dependent on or independent of their respective scenes the items studied were. These findings offer pedagogical implications of mobile-based AR use in vocabulary study and language learning in general.