Research comparing the use of retrospective interviews (RIs) and think-aloud protocols (TAPs) as two data sources of writers' composing processes is almost nonexistent. Trying to address this research gap, the present study examined the data obtained from the task-specific RIs and TAPs about the composing processes of 30 second language (L2) writers. Each participant thought aloud by concurrently verbalizing thoughts while performing an argumentative writing task, and then was interviewed immediately after the think-aloud session about their composing processes. The RI and TAP data was analyzed and compared in terms of five aspects of L2 writers' composing: pre-writing stage, while-writing planning, first language (L1) use, composing problems and problem-solving, and text reviewing and changing. Though the results generally indicate that concurrent verbalizations provide much richer data than RIs about L2 writers' composing processes, RIs were found to be a valuable source for revealing the 'why' of such processes. Besides, the gap between the data provided by the two sources differed from one composing aspect to another. Recommendations for how to optimally combine and make use of the two data sources in writing process research are provided.