Individual meta-analyses were conducted for six training methods as part of a U.S. Army basic research project. The objective was to identify evidence-based guidelines for the effectiveness of each training method, under different moderating conditions, for cognitive skill transfer in adult learning. Results and implications for two of these training methods, learner control (LC) and exploratory learning (EL), are discussed. LC provides learners with active control over training variables. EL requires learners to discover relationships and interactions between variables.

There is mixed evidence on the effectiveness of both LC and EL learning methods on transfer relative to more guided training methods. Cognitive load theory (CLT) provides a basis for predicting that training strategies that manage intrinsic load of a task during training and minimize extraneous load will avail more resources that can be devoted to learning.

Meta-analyses were conducted using a Hedges’s g analysis of effect sizes. Control conditions with little to no learner freedom were contrasted with treatment conditions manipulating more learner freedom.

Overall more LC was no different from training with limited or no learner control, and more EL was less effective than limited or no exploration; however, each can be effective under certain conditions. Both strategies have been more effective for cognitive skill learning than for knowledge recall tasks. LC exhibited more benefit to very near transfer, whereas EL’s benefit was to far transfer.

Task type, transfer test, and transfer distance moderate the overall transfer cost of more learner freedom.

The findings are applicable to the development of instructional design guidelines for the use of LC and EL in adult skill training.