Recently we have witnessed a large growth in research on written corrective feedback (WCF). However, the question posed here is: are researchers and L2 writing teachers now any wiser about the efficacy of WCF? I begin with a summary of early studies and some of their major shortcomings. I then examine more recent studies and conclude that, although many of the shortcomings of earlier research have been largely addressed, research findings are still inconclusive. I argue that currently, in the desire to conduct more robust research, the pendulum has swung too far towards experimental studies. Such studies tend to employ "one off" treatments, often provided on a very restricted range of errors, and ignore the learners' goals and attitudes to the feedback provided and to improvement in accuracy. I conclude by suggesting directions for a more meaningful and ecological valid research agenda on written corrective feedback. (Contains 4 tables and 1 note.)