In recent years, virtual learning has emerged as an alternative to traditional classroom teaching. The engagement of students in virtual learning can have a major impact on meeting learning objectives and program dropout risks. There exist many measurement instruments specifically geared to Student Engagement (SE) in virtual learning environments. In this critical review, we analyze these works and highlight inconsistencies in terms of differing engagement definitions and measurement scales. This diversity among existing researchers could potentially be problematic in comparing different annotations and building generalizable predictive models. We further discuss issues in terms of engagement annotations and design flaws. We analyze the existing SE annotation scales based on our defined seven dimensions of engagement annotation, including sources, data modalities used for annotation, the time when the annotation takes place, the timesteps in which the annotation takes place, level of abstraction, combination, and quantification. One of the surprising findings was that very few of the reviewed datasets on SE measurement used existing psychometrically validated engagement scales in their annotation. Lastly, we discuss some other scales in settings other than virtual learning that have the potential to be used in measuring SE in virtual learning.