Many instructional strategies that appear to improve learners' performance during training may not realize adequate posttest performance or transfer to a job. The converse has been found to be true as well: Instructional strategies that appear to slow the learner's progress during training often lead to better posttraining or transfer performance. For example, many studies have shown beneficial effects of random over blocked practice on transfer of learning, even though blocked practice often leads to better performance during the training session. In a 2 x 3 factorial experiment (N = 120), with the factors practice schedule (random, blocked) and critical thinking prompts (before task, after task, none), this study investigates whether this also applies to complex judgment tasks and whether critical thinking prompts can enhance the effectiveness of particular practice schedules. It is hypothesized that prompts provided after task execution yield best transfer in a random practice schedule, whereas prompts provided before task execution yield best transfer in a blocked schedule. In line with this hypothesis, a blocked schedule led to better performance than random practice during training but not on the transfer test, where a random schedule was beneficial. The hypothesized interaction effect was also found: Critical thinking prompts after task execution significantly benefit transfer performance of participants following a random schedule, and transfer performance following a blocked schedule can be a little enhanced through providing critical thinking prompts before task execution. These results warrant instruction in critical thinking processes to teach complex judgment tasks, using random practice schedules combined with critical thinking prompts provided after task execution. (Contains 7 tables, 3 figures and 2 footnotes.)