Research on Processing Instruction (PI) has yielded consistently positive results across a variety of measures. In addition, in comparison to other instructional interventions, PI tends to yield superior results. The one difference is comparative research with Meaning-based Output Instruction (MOI) in which in a number of studies, MOI is seen to be as good as PI in terms of group performance on certain tasks. The present study is a partial replication of one such study: Keating and Farley (2008). We compared three groups--PI, MOI, and Control--on two measures using a pretest/posttest/delayed posttest design. Our results do not replicate Keating and Farley's and instead resemble more those of original research on PI and traditional instruction (e.g., Van-Patten and Cadiemo, 1993). We discuss our findings in relation to the nature of MOI.