The Attentional Boost Effect (ABE) refers to the counter-intuitive finding that the detection of infrequent targets in a divided-attention (DA) condition enhances memory of images co-occurring with targets (as compared with images co-occurring with distractors; Swallow & Jiang Cognition, 115, 118-132, 2010). Previous studies have shown that the ABE also applies to verbal materials (words; Spataro, Mulligan, & Rossi-Arnaud Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 1223-1231 , 2013) and documented an important moderating factor, word frequency-the ABE was robust for high-frequency words, but small or non-significant for low-frequency words (Mulligan, Spataro, & Picklesimer Journal of Experimental Psychology: Learning, Memory, and Cognition, 40, 1049-1063, 2014). The present experiment tested the predictions of the early-phase-elevated-attention hypothesis of the ABE by manipulating the orthographic distinctiveness of the to-be-remembered words. Results revealed that the ABE was significant for low-frequency words with common orthographic features, but not for low-frequency words with rare orthographic features. As a consequence, the orthographic distinctiveness effect (ODE) was eliminated in the DA condition. These findings are in line with the proposal that the ABE-related attentional enhancement occurs during an early phase of stimulus perception and comprehension, as well as with the proposal that the ODE is mediated by high-level, attention demanding comparative processes.