Practice makes (nearly) perfect: Solving "students & professors"-type problems
Christianson, K., Mestre, J.P., & Luke, S.G. (2012). Applied Cognitive Psychology, 26, 810-822.
The strongest predictor of students producing correct equations was practice: In all experiments, participants spontaneously improved in equation accuracy almost to ceiling levels as they progressed, despite receiving no feedback. Tentative support is provided for the pedagogical value of repetition in solving problems, along with varying the wording of the problems.