Virtual experiments offer opportunities for students to engage in the scientific process and observe scientific phenomena. However, virtual experiments are not always effective for learning. The current study examined how two small manipulations might support better learning from virtual experiments and the extent to which this varied across learners. Undergraduates in an introductory biology class (n = 415) were randomly assigned to a 2(agency: active, passive) × 2(strategy instruction: control condition, control of variables [CVS] strategy instruction) between-subjects design and completed two prior knowledge tests (topic, general science). Comprehension tests results indicated that the students had relatively good comprehension of the experiment content, but that performance was driven by prior knowledge. Although there were no significant effects of the agency or strategy instruction manipulations, exploratory interactions suggest a need to further investigate which types of virtual experiment supports might be most beneficial for different learners. [This paper was published in "American Journal of Distance Education."]