Most articles retracted by medical journals are withdrawn because of misconduct, not research error, and the proportion of articles retracted because of fraud -- while still comparatively small -- has increased 10 times since 1975, says a study published this week in the Proceedings of the National Academy of Sciences.