Retrieval Practice and Bloom's Taxonomy
By Cindy Nebel
As researchers attempting to bridge psychological science and education, we come across several challenges. Each of us has expertise in the area of retrieval practice, where many of the classic studies take place in the laboratory with simple materials. Researchers have therefore attempted to answer several questions to address the extent to which the research conducted in the laboratory meaningfully applies to the classroom. We have discussed these issues in previous blogs, specifically with relation to the lab to classroom model, transfer to related material, and how knowledge of facts are required for application.
One criticism that we have received is that retrieval practice is primarily good for fact learning, but not for higher-order learning. In other words, retrieval practice might serve rote memorization but not the ability to critically think about material or apply it. We addressed this question here and explained that facts must be sufficiently encoded in order to use those facts in a new situation. Because retrieval supports knowledge acquisition, retrieval practice of facts should therefore support application.
However, a recently published study (1) by one of our colleagues, Dr. Pooja Agarwal, examined whether retrieval practice could do more than just support the acquisition of factual information. The study was based on a common prescription for using Bloom’s taxonomy (2): students should first focus on the lower levels of the taxonomy before higher-order thinking can be accomplished. Dr. Agarwal directly compared retrieval practice with the use of lower vs. higher-order thinking to determine if the lower levels of Bloom’s taxonomy were indeed necessary before moving to more complex thinking. Here is what she did:
First, she examined the effect of quizzing using either fact or higher-order questions on a final test two days later using laboratory materials. Participants read passages that supported two sides of a controversial issue (e.g. “Does welfare do more harm than good?”) and then were given the opportunity to study the passage again, took a fact-based quiz, or took a higher-order quiz over the passage. In order to illustrate the difference, here is an example of each type of question:
After two days, participants returned to the lab and took a quiz comprised of rephrased versions of the same fact and higher order questions. If the typical use of Bloom’s taxonomy proved correct, we would expect that retrieval practice on fact questions should promote better higher-order thinking about those same facts. However, instead, Dr. Agarwal found that performance on the final test was facilitated by quizzing of the same type of questions (see results to the right).
In a second experiment, Dr. Agarwal compared a condition in which participants received both the fact and higher order quizzes to just the fact or higher-order quizzes (now taken twice). In that experiment, what they received on the initial test directly improved performance of that same material on the final test. That is, two fact quizzes provided a big boost for fact performance and two higher-order quizzes provided a big boost for higher-order performance. A single test of each fact and higher-order quiz produced a boost for both types of final test questions but the boost was not as high as when getting two quizzes (as seen to the right).
Finally, to directly test whether these effects persist in the classroom, Dr. Agarwal performed the same experiment using two social studies textbook chapters in a 6th grade classroom. In the classroom, having received the mix of both fact and higher-order quizzes resulted in a larger boost in higher-order final test performance. So, in sum, fact quizzing did not promote boosts for higher-order final test performance, but a mix of both fact and higher-order quizzing did. Here are those classroom results:
As a reminder, Dr. Agarwal was directly testing the idea that retrieval practice for factual material should promote higher-order thinking. Across these three experiments, there was no evidence for that assumption. Instead it appears that final test performance is based on the type of processing that has previously occurred, a principle known as transfer appropriate processing (3).
There is one potential issue about this study that should be considered: the initial and final test questions were extremely similar. The effects that were found still support transfer appropriate processing, but the extent to which these big boosts would have been found with different questions is unclear. In many classrooms, the instructor does not use such similar questions on the final test and so the applicability to all classrooms is somewhat limited. In addition, we cannot know the situations in which students will be required to use the knowledge they have acquired in the classroom and so this study represents somewhat near transfer compared to the far transfer in the “real world”. (You can read more about this issue here.)
Based on this study, we can provide some advice for educators. If we want students to be able to apply, analyze, or create, we must encourage this type of processing within the classroom. Ideally, we should try to anticipate the situations in which students are likely to need to use this higher-order thinking and encourage it. But one thing is clear: you should not avoid higher-order questions for fear that students need a base of factual questions first. In order for students to process information at each level of Bloom’s taxonomy (including understanding), we need to require students to use each level within the classroom.
(1) Agarwal, P. K. (2019). Retrieval practice & Bloom’s taxonomy: Do students need fact knowledge before higher order learning? Journal of Educational Psychology, 111(2), 189-209.
(2) Bloom, B. S. (Ed.), Engelhart, M. D., Furst, E. J., Hill, W. H., & Krathwohl, D. R. (1956). The taxonomy of educational objectives: The classification of educational goals (Handbook 1: Cognitive domain). New York, NY: David McKay Company.
(3) Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer-appropriate processing. Journal of Verbal Learning & Verbal Behavior, 16, 519 –533.