Paper investigating the Retrieval practice and transfer learning. Includes several experiments investigating progressively larger degrees of transfer.
Subjects were psychology undergrads at WUSTL. They read short passages sourced from Wikipedia on general knowledge topics. Experiments 1b and 2 asked subjects to make inferences within the same topic. Experiment 3 asked subjects to use what they’d read to make inferences in a different topic. Both experiments found that retrieval practice produced better transfer performance than re-studying, by a similar margin to verbatim questions, on both conceptual and factual questions.
Experiments 1a and 1b also explored Would spaced repetition memory systems perform better with varied question texts? with a varied-questions condition but found a null result.
One key limitation of this study: students were “explicitly told that the questions on the final test were related to the information they had learned in the previous study.” And so students did not need to recognize the applicability of information to the far-transfer task; they just needed to recall the relevant info. (Thanks to Martin Bernstorff for pointing this out).
Q. High-level experimental setup?
A. Psychology undergrads read short passages, studied or took tests repeatedly, then were tested one week later.
Q. Main finding?
A. On both near and far transfer, retrieval practice produced better transfer performance than re-studying, by a similar margin to verbatim questions.
Q. Finding w.r.t. Would spaced repetition memory systems perform better with varied question texts??
A. Null result.
Q. Key limitation w.r.t. implications for Retrieval practice and transfer learning?
A. Students were explicitly told that final test questions were related to information they’d learned in the previous study—so they didn’t need to recognize its applicability.