We investigated the mechanisms underlying sandwich priming, a procedure in which a brief preprime target presentation precedes the conventional mask-prime-target sequence, used to study orthographic similarity. Lupker and Davis (2009) showed the sandwich paradigm enhances orthographic priming effects: With primes moderately related to targets, sandwich priming produced significant facilitation, but conventional priming did not. They argued that unlike conventional priming, sandwich priming is not susceptible to an uncontrolled counteractive inhibitory process, lexical competition, that cancels out moderate facilitation effects. They suggest lexical competition is eliminated by preactivating the target's representation, privileging the target over similar lexical units (competitors). As such, it better measures orthographic relatedness between primes and targets, a key purpose of many priming studies. We tested whether elimination of lexical competition could indeed account for the observed orthographic priming boost with sandwich priming. In three lexical decision experiments and accompanying simulations with a competitive network model, we compared priming effects in three preprime procedures: no preprime (conventional), identity (target) preprime (sandwich), and competitor preprime (included to exacerbate lexical competition). The related prime conditions consisted of replaced-letters, shared neighbor (one-letter-different from both competitor preprime and target), and transposed-all-letter nonword primes. Contrary to the model's predictions, the competitor preprime did not attenuate (Experiment 1) or even reverse the priming effect (Experiment 2). Moreover, the competitor enabled facilitatory priming that was absent with no preprime (Experiment 3). These data suggested that the sandwich orthographic boost could not be attributed to reduced lexical competition but rather to prelexical processes in word recognition.