In Butler's (2010) Experiment 3, 20 undergraduates studied prose passages. Each participant completed both practice conditions on different passages: three cued-recall tests with correct-answer feedback, or repeated restudy. One week later, they answered new inference questions applying studied concepts to different knowledge domains. Mean accuracy was higher for tested material (68%) than restudied material (44%). Each final question mentioned the relevant earlier concept, and instructions told learners to use their previous learning. The finding therefore concerns cued transfer; the experiment did not test whether learners would independently recognize when knowledge was relevant. It also did not separate improved recall from improved ability to apply the concepts.

Source: Butler (2010)