Retrieval Practice: The Most Effective Study Strategy

Studying by Remembering Beats Studying by Reviewing

In 2006, Henry Roediger III and Jeffrey Karpicke published a study in Psychological Science that became one of the most cited papers in educational psychology. Participants read a prose passage and then either restudied it three additional times or took three recall tests (writing down everything they could remember). Five minutes later, the restudy group outperformed the testing group slightly. One week later, the testing group remembered 50% more than the restudy group. The act of retrieving information from memory did not just assess learning -- it caused learning.

This finding, known as the testing effect or retrieval practice effect, has been replicated hundreds of times across materials, populations, and settings. A 2013 review by Dunlosky and colleagues rated practice testing as one of only two study strategies (alongside distributed practice) with "high utility" -- the highest rating in their evaluation of ten common techniques.

Why Retrieval Strengthens Memory

The cognitive mechanisms behind the testing effect involve several complementary processes. No single explanation fully accounts for the phenomenon, but together they provide a coherent picture.

Key Experimental Findings

The research base for retrieval practice spans laboratory studies, classroom experiments, and meta-analyses. Several findings stand out for their practical implications.

• Karpicke and Blunt (2011) found that retrieval practice outperformed concept mapping for learning science texts, even when the concept mapping condition spent more time on the material
• Retrieval practice benefits extend beyond factual recall to inference, application, and transfer. A 2012 study by Butler found that students who practiced retrieval on one set of facts performed better on questions requiring integration of those facts with new information
• The benefit is largest when retrieval is successful but effortful -- getting it right after struggling produces more learning than either failing to retrieve or retrieving too easily
• Feedback after retrieval attempts enhances the effect, particularly for incorrect responses. Immediate feedback is generally better than delayed feedback for correcting errors
• A 2014 meta-analysis by Rowland, covering 159 research articles, found an average effect size of 0.50 for retrieval practice compared to restudy -- a medium-to-large effect

Retrieval Practice in Classroom Settings

Laboratory findings translate to real classrooms. Multiple field studies have demonstrated the testing effect in educational settings with diverse student populations.

Roediger, Agarwal, McDaniel, and McDermott (2011) conducted a multi-year study in a Columbia, Illinois, middle school. Teachers embedded low-stakes quizzes into regular instruction for some content but not others. On unit exams, students scored a full letter grade higher on quizzed material compared to non-quizzed material from the same course. The effect held across subjects (science, social studies) and across student ability levels.

McDaniel, Agarwal, Huelser, McDermott, and Roediger (2011) found similar results in a college-level brain and behavior course. Students who answered clicker questions during lectures performed significantly better on exam items covering that content than on items covering non-quizzed lecture content.

Optimal Conditions for Retrieval Practice

Not all retrieval practice is equally effective. Research identifies several factors that modulate the size of the benefit.

Spacing matters. Retrieval practice combined with spaced intervals produces synergistic effects. Cepeda and colleagues' 2006 meta-analysis of 254 studies found that spacing study sessions across time roughly doubled retention compared to massing the same amount of study into a single session. When retrieval practice is spaced -- test today, test again in three days, test again in ten days -- the cumulative benefit exceeds either technique alone.

Retrieval format. Free recall (writing everything you remember about a topic) produces larger effects than recognition (multiple choice), likely because free recall requires more effortful search. However, even multiple-choice testing produces significant benefits over restudy.

• Short-answer and free recall formats generate larger testing effects than recognition formats
• Mixing question types within a practice session (interleaving) enhances discrimination between similar concepts
• Pretesting -- attempting to answer questions before studying the material -- enhances subsequent learning, even when initial answers are wrong
• Cumulative retrieval (including older material in each practice session) prevents forgetting of previously learned content

Common Misconceptions

Several misunderstandings about retrieval practice persist among students and educators. The most damaging is the belief that if you can't remember something, you haven't learned it and should go back to restudying. In fact, the struggle to retrieve -- and sometimes failing -- is itself valuable. Unsuccessful retrieval attempts, when followed by feedback, can enhance learning more than successful retrieval of easy material.

Another misconception is that retrieval practice only works for rote memorization. The evidence shows benefits for higher-order thinking, including application, analysis, and transfer to new contexts. A third misconception equates retrieval practice with high-stakes testing. The research specifically supports low-stakes or no-stakes practice, where the goal is learning, not evaluation. Test anxiety, a real phenomenon, is a response to high-stakes assessment, not to the act of retrieval itself.

Building a Retrieval Practice Habit

The simplest form of retrieval practice requires no technology: close the book, look away from the screen, and write down everything you can remember. Then check against the source material. The gap between what you thought you knew and what you could actually retrieve is the most honest measure of learning available. Flashcards (physical or digital through apps like Anki), practice tests, brain dumps, and teaching material to someone else all qualify as retrieval practice. The common thread is pulling information out of memory rather than pushing it back in through rereading. That reversal of direction is what makes the difference.