To minimize drug name confusion errors, regulators, drug companies, and clinicians need tools that help them predict which names are most likely to be involved in confusions. Two experiments, carried out in the United States, examined the effects of stimulus frequency (i.e., how frequently a target name is prescribed), neighborhood frequency (i.e., how frequently prescribed are the "neighbors" of the target name), and neighborhood density (how many names are within a fixed distance of the target name) on the probability of pharmacists making an error in a visual perceptual identification task. In both experiments, the task was to correctly identify a series of blurry drug names after a 3s presentation on a computer monitor. In the first experiment, 45 pharmacists viewed 160 typewritten names, incorrectly identifying 60.6% of them. Random effects regression revealed a significant beneficial effect of stimulus frequency and a detrimental effect of neighborhood density. Significant two-way interactions were observed between stimulus frequency and neighborhood density and neighborhood frequency and neighborhood density. In the second experiment, 37 pharmacists viewed 156 handwritten drug names, incorrectly identifying 45.7%. Random effects regression revealed significant main effects of stimulus frequency and neighborhood density. These were contained within a significant three-way interaction: The interaction between stimulus frequency and neighborhood density was present at high but not low neighborhood frequency. Objectively measurable frequency and neighborhood characteristics have predictable effects on errors in pharmacists' visual perception. Organizations that coin and evaluate drug names, as well as hospitals, pharmacies, and health systems, should consider these characteristics when assessing visually confusing names.