Within-subject designs are uniquely able to reduce threats to internal validity with repeated measurements of performances and measurement of effects of the repeated application of the independent variable. However, in some skill acquisition research with typically developing learners (e.g., college students) that use non-arbitrary stimuli (i.e., actual academic content) traditional experimental designs might be insufficient in mitigating threats to internal validity, particularly with regard to testing (i.e., practice or exposure) and history effects. An analysis of previously published research and preliminary data suggest that when identical tests are conducted on a single day, performances (measured as percent correct for a given test) for typically developing learners on computerized selection-based tests improve from the first test to the second test in the absence of experimenter/system provided reinforcement. We do not believe that any published studies have investigated this possible threat to research designs that are currently being used by experimenters (e.g., non-concurrent multiple baseline across participants designs with one or two pretests prior to intervention). In addition to the presentation of our findings, we discuss possible solutions and options available to researchers that might better control for threats to internal validity inherent in skill acquisition research.