The role of awareness has been under investigation in second language (L2) acquisition research for many years. Williams' study (2005) has attracted particular interest and has been actively reexamined from a variety of perspectives, and the replications and extensions of the research currently form a research horizon on the intersection of artificial grammar learning and study of L2 acquisition. This paper discusses the potential biases of previously used methodologies and their effects on research outcomes. The current analysis finds that the learning outcomes depend considerably on the extent of assessment of awareness or knowledge acquired.