Self-regulated learning (SRL) is a fundamental skill to succeed in Massive Open Online Courses (MOOCs), but many learners do not know how to self-regulate their learning. The need to support SRL in MOOCs led to the idea of social-psychological interventions that promise to improve course performance and decrease dropout rates. However, past research provides mixed evidence of the effectiveness of SRL interventions in MOOCs. In this randomized control trial (RCT), the heterogeneous effects of SRL intervention in three MOOCs were examined. The SRL intervention was embedded in a precourse survey, where learners were randomly assigned to experimental (N = 383) and control (N = 444) conditions. Both groups answered contextual questions, and then the experimental group was guided through a writing activity to boost SRL skills. The study aimed to assess how learner demographics may affect the results of the RCT. The results yielded no significant differences overall between the experimental and control conditions. However, the results of the binary logistic regression demonstrated that the heterogeneous effect is prevalent in SRL interventions in regard to learner demographics: males and older learners received advantages from the intervention. The current study adds to the field of SRL intervention in MOOCs and presents directions for future experiments. Based on the results of the paper, a number of methodological issues of SRL interventions in MOOCs were formulated, including self-selection bias and interventions that were not a part of the learning process, that focused on academic outcomes, and that had no follow-up analysis.