What is retrieval practice and why is it more effective than re-reading?
Retrieval practice is any study activity that requires you to generate information from memory rather than recognize it from a source. Re-reading is recognition; flashcards and practice questions are retrieval. Each successful retrieval strengthens the memory trace that produced it.
In 2006, Henry Roediger and Jeffrey Karpicke published a study in Psychological Science that changed how learning researchers think about the most fundamental question in education: what is the best way to study? They divided students into groups: some read a passage once, some read it four times, and some read it once and then took three free-recall tests. One week later, the retrieval practice group remembered 50% more than the group that had read the material four times.
That result — replicated dozens of times across different subjects, age groups, and material types — established retrieval practice as the single most effective study technique we know of. For IT certification candidates, this research has a direct and actionable implication: the time you spend trying to remember what you have studied is more valuable than the time you spend re-exposing yourself to it.
"Retrieval practice is not just testing what you know. It is the act of learning itself. Every retrieval attempt — whether successful or not — changes the structure of memory in ways that passive review simply cannot replicate." — Henry L. Roediger III, Make It Stick: The Science of Successful Learning, Belknap Press, 2014
This article covers the specific retrieval practice techniques that work for certification study and how to integrate them into a coherent study approach.
What retrieval practice is and why it works
Retrieval practice is any study activity that requires you to actively produce information rather than recognize it. The defining feature is the generation attempt: you must try to recall, reconstruct, or explain something before checking a source.
The testing effect
The cognitive mechanism is called the testing effect. Each time a memory is successfully retrieved, the neural pathway connecting the retrieval cue to the stored information is strengthened. The act of finding the memory — even when effortful, even when partially unsuccessful — is itself a memory-encoding event.
The strength of a memory trace is determined not by how many times you have encountered the information but by how many times you have successfully retrieved it. A fact you have retrieved ten times under challenging conditions is encoded far more durably than a fact you have read fifty times.
The role of forgetting
Forgetting is also part of the mechanism. Information that is slightly difficult to retrieve — because time has passed since you last studied it, because the retrieval cue is unfamiliar, or because you are encountering it in a new context — produces the strongest encoding when successfully recovered. This is why spacing practice (retrieving information after a forgetting interval) produces better results than massed practice (retrieving it repeatedly in a short window while it is fresh).
Technique 1: Free recall
Free recall is the simplest and most powerful retrieval technique. After studying a topic, close everything and write down everything you can remember about it. No prompts, no cues, no structure — just write.
This produces two things simultaneously: retrieval practice (the act of remembering encodes more deeply) and a gap inventory (what you cannot write down reveals precisely what you have not encoded).
For a CompTIA Security+ candidate who has just studied network security, a free recall exercise might produce: "VPN types — IPSec (L3) and SSL/TLS (application layer). IPSec has AH and ESP, ESP provides encryption, AH provides integrity. IKE for key exchange. Firewall types — packet filtering (stateless), stateful, application layer, next-gen. IDS vs IPS — IDS is passive, monitors and alerts; IPS is inline, can block. NIDS vs HIDS..."
The moment the flow stops, the blank page reveals the boundary of your retrievable knowledge. That boundary is your study target.
Free recall sessions should happen at the end of each study block (immediate retrieval) and at the beginning of the next session on the same topic (spaced retrieval). The spaced version is typically more valuable because the forgetting interval means more effort is required to retrieve, which produces stronger encoding.
Technique 2: Spaced flashcard retrieval
Flashcard systems like Anki implement a specific algorithm (SM-2, or variations of it) that schedules each card for review at the optimal interval to exploit the forgetting curve. Cards you answer easily are scheduled further out. Cards you answer with difficulty are scheduled soon.
The result is an adaptive system that maximizes retrieval practice for the concepts you need most while minimizing time spent on concepts you already know well. For a candidate covering 400-500 concepts across a certification domain, manual scheduling would be impossible. Anki does it automatically.
Effective flashcard design for certification study requires generating cards that test understanding and application rather than raw recall.
Weak cards (test recall only):
Q: "What port does HTTPS use?" / A: "443"
Q: "What does OSPF stand for?" / A: "Open Shortest Path First"
Strong cards (test understanding and application):
Q: "Your OSPF neighbor relationship is stuck in Exstart/Exchange state. What is the most likely cause?" / A: "MTU mismatch between the two interfaces"
Q: "An application needs to securely store user session tokens with sub-millisecond access latency. Which AWS service fits best and why?" / A: "ElastiCache (Redis or Memcached) — in-memory key-value store, sub-ms latency; ElastiCache for Redis preferred if session persistence is needed across cache failures"
Strong cards require you to produce reasoning, not just recall a fact. They directly prepare you for scenario questions.
Technique 3: Practice questions with committed answers
Practice questions are retrieval practice when used correctly. The correct use requires committing to an answer — deciding what you think the answer is and why — before reading answer choices.
The process: read the question stem. Cover the answer choices with your hand or a blank sheet of paper. Form your answer. Write it or state it explicitly. Then uncover the choices and select the closest match to what you said.
This is significantly harder than reading question and choices together, which allows your brain to evaluate options rather than retrieve answers. The evaluation mode is recognition; the committed-answer mode is retrieval. The retrieval mode produces stronger encoding when your answer is correct, and stronger correction when your answer is wrong.
For multiple-select questions, commit to your specific set of answers before looking at the options. If a question asks for three correct answers, decide which three you think they are before uncovering the choices.
Reviewing missed questions should use the same committed-answer principle: after reading the correct answer, close the explanation and write your own explanation of why the answer is correct. Then compare your explanation to the provided one. The gap between them is your precise learning target.
Technique 4: The brain dump
A brain dump is a variant of free recall applied at the beginning of a study session rather than the end. Before opening any materials, spend five to ten minutes writing everything you currently know about the topic you are about to study.
The brain dump serves two purposes. First, it is spaced retrieval of previously studied material — the effort to retrieve produces encoding. Second, it activates the memory traces related to the topic, which research shows improves encoding of new information presented immediately after. Information is encoded more efficiently when it can be connected to existing memory traces.
For a candidate beginning a study session on Azure Active Directory after previously studying it two days ago, a brain dump might produce:
"AAD — Microsoft's cloud identity service, different from on-prem AD. Users, groups, service principals. App registrations vs enterprise apps. OAuth 2.0 and OIDC for modern auth. Conditional Access policies — conditions + controls, grant or block access. MFA — different methods: authenticator app, FIDO2, SMS (weakest). B2B and B2C for external identities. SSPR — self-service password reset. Hybrid identity — password hash sync vs pass-through auth vs ADFS. Azure AD Connect for sync. Privileged Identity Management for JIT access..."
Each item in that dump is a retrieval event. The items that come quickly are well-encoded. The items that require effort to surface — "what was the other hybrid identity option besides password hash sync?" — are candidates for targeted review.
Technique 5: Practice exam simulation under timed conditions
Full-length timed practice exams are the highest-fidelity retrieval practice available for certification study. They replicate the exam conditions — time pressure, randomized question order, no external resources — and force retrieval across the entire exam domain in a single session.
The retrieval benefit of a full practice exam is maximized when you:
Take the exam without any interruptions or resource checking.
Flag questions you are uncertain about but commit to an answer for each one.
Review results immediately after completion, focusing on missed and flagged questions.
Apply free recall to each missed question before reading the explanation.
The common mistake is taking practice exams in a relaxed, open-book manner — checking your notes when you encounter a hard question. This eliminates the retrieval challenge that produces encoding. An open-book practice exam is useful for calibrating what you do not know; it is not useful as retrieval practice.
Start taking full practice exams earlier than feels comfortable — when you have covered roughly 60% of the exam content. The questions on domains you have not yet studied force retrieval on partial knowledge, which is more productive than waiting until you have covered everything and taking exams from a position of apparent completeness.
Building a retrieval-dominant study week
Weekly schedule structure
A study week for a certification candidate six weeks from exam day:
| Day | Primary activity | Retrieval technique |
|---|---|---|
| Monday | New content: chapters 9-10 | End-of-session free recall (15 min) |
| Tuesday | Active review: chapters 7-10 | Brain dump (10 min) + Anki due cards (25 min) |
| Wednesday | Practice questions: 60 questions | Committed answers, then gap review |
| Thursday | New content: chapter 11 | End-of-session free recall (15 min) |
| Friday | Mixed review: all domains | Brain dump + Anki due cards |
| Saturday | Full practice exam (timed) | Full simulation, no resources |
| Sunday | Practice exam review | Missed question free recall + Feynman explanations |
In this week, retrieval practice accounts for roughly 60% of total study time. Content acquisition accounts for 40%. This ratio reflects what research on effective learning supports: most of your study time should be retrieving, not re-reading.
Time allocation across a study campaign
The appropriate ratio of retrieval to content study shifts over the course of a campaign. Early weeks need more content acquisition to build the knowledge base that retrieval practice draws from. Later weeks should be heavily retrieval-dominant.
| Campaign phase | Weeks | Content acquisition | Retrieval practice |
|---|---|---|---|
| Foundation | Weeks 1-2 | 60% | 40% |
| Deep study | Weeks 3-5 | 40% | 60% |
| Integration | Weeks 6-7 | 20% | 80% |
| Final consolidation | Last week | 5% | 95% |
See also: Active Recall vs Passive Review: Why Re-Reading Your Notes Fails
Retrieval techniques matched to certification types
The retrieval method with the highest return varies by exam format. Multiple-choice-dominant exams reward one pattern; performance-based and lab exams reward another.
| Certification | Current exam code | Fee | Highest-ROI retrieval method |
|---|---|---|---|
| CompTIA Security+ | SY0-701 | $404 | Anki flashcards + weekly brain dumps |
| CompTIA A+ | 220-1101/1102 | $506 | Anki + performance-based simulation |
| AWS SAA-C03 | SAA-C03 | $150 | Mixed practice tests + architecture sketching |
| AWS SAP-C02 | SAP-C02 | $300 | Full-length practice tests weekly |
| Azure AZ-104 | AZ-104 | $165 | CLI/PowerShell blank-page retrieval + labs |
| Cisco CCNA | 200-301 | $300 | Daily config-from-memory + Packet Tracer |
| CKA | CKA | $395 | Timed kubectl from memory exercises |
| CISSP | CISSP | $749 | Domain-by-domain mind map reconstruction |
| PMP | PMP | $555 | Situational question drills |
| Google PCA | PCA | $200 | Architecture decision drills from case studies |
The common thread is that effective retrieval matches the cognitive demand of the exam. CCNA candidates should retrieve CLI commands from memory because that is what the exam requires. CISSP candidates should retrieve cross-domain conceptual connections because that is what the exam requires. Wasting retrieval effort on content the exam does not test produces study hours that do not translate to exam performance.
Practice test ROI and when to deploy them
Full practice tests are the single highest-value retrieval activity for most certification exams, but timing and sequencing matter significantly.
| Campaign phase | Practice test frequency | Test type | Primary purpose |
|---|---|---|---|
| Foundation (weeks 1-3) | 0-1 per week | Short topic tests (25 questions) | Identify knowledge gaps |
| Deep study (weeks 4-7) | 1 per week | Domain-focused (50 questions) | Domain mastery |
| Integration (weeks 8-10) | 2 per week | Full-length mixed (all domains) | Build exam stamina |
| Final consolidation (week 11+) | 3 per week | Full-length timed | Score calibration |
"Testing is learning. The moment you retrieve information from memory under testing conditions, you strengthen that memory more than you would by studying the material again. This has been replicated across hundreds of experiments, multiple age groups, and diverse subject matter. The literature converges on the same conclusion: if you are not testing yourself, you are not learning as efficiently as you could be." - Henry L. Roediger III, PhD, Washington University in St. Louis, in Annual Review of Psychology [1].
Brain dump variations
The basic brain dump - close all materials and write everything you remember about a topic - has variants suited to different study contexts.
Timed brain dump - five minutes, one topic, write everything retrievable. Forces speed and exposes shallow knowledge.
Structured brain dump - with a topic heading, attempt to produce a hierarchical outline from memory. Better for conceptual domains.
Diagram-first brain dump - draw a diagram (network topology, VPC layout, process flow) from memory, then label. Best for visual/spatial content.
Compare-to-yesterday brain dump - write today's brain dump, then compare to yesterday's on the same topic. The delta shows what stuck and what slipped.
Peer brain dump - two candidates dump the same topic simultaneously, then compare. Reveals gaps one candidate has and the other does not.
Brain dumps produce maximum value when they are reviewed against source material within 24 hours. A brain dump never compared to the source is a self-assessment with no feedback - candidates reliably overestimate the accuracy and completeness of what they wrote.
The "lag effect" in retrieval scheduling
Retrieval practice is more effective when spaced than when massed. Three retrieval attempts over three weeks produce significantly better retention than three attempts in the same session.
| Retrieval schedule | Total retrieval attempts | Expected retention at exam day (6 weeks later) |
|---|---|---|
| Massed (same session) | 3 in 30 min | ~40% |
| Spaced daily | 3 in 3 days | ~60% |
| Spaced weekly | 3 in 3 weeks | ~75% |
| Expanding intervals | 3 at days 1, 3, 10 | ~80% |
Anki's default scheduling algorithm approximates expanding intervals. Candidates who use Anki for the majority of their retrieval practice benefit from this automatically. Candidates relying on manual review schedules should explicitly build in expanding intervals: review a new concept the same day, two days later, one week later, and three weeks later.
References
[1] Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255.
Brown, P. C., Roediger, H. L., & McDaniel, M. A. (2014). Make It Stick: The Science of Successful Learning. Belknap Press. ISBN: 978-0674729018.
Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for learning. Science, 319(5865), 966-968.
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4-58.
Adesope, O. O., Trevisan, D. A., & Sundararajan, N. (2017). Rethinking the use of tests: A meta-analysis of practice testing. Review of Educational Research, 87(3), 659-701.
Kornell, N., Hays, M. J., & Bjork, R. A. (2009). Unsuccessful retrieval attempts enhance subsequent learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(4), 989-998.
Carey, B. (2014). How We Learn: The Surprising Truth About When, Where, and Why It Happens. Random House. ISBN: 978-0812993882.
Frequently Asked Questions
What is retrieval practice and why is it more effective than re-reading?
Retrieval practice is any study activity that requires you to generate information from memory rather than recognize it from a source. Re-reading is recognition; flashcards and practice questions are retrieval. Each successful retrieval strengthens the memory trace that produced it. Research by Roediger and Karpicke found that three retrieval practice sessions produced 50% better retention one week later than four re-reading sessions using the same total study time.
How do I use practice questions as retrieval practice rather than passive review?
Commit to an answer before looking at the choices. Read the question stem, cover the answer options, decide what you think the answer is and why, then uncover the choices. This forces retrieval rather than evaluation of recognition. For missed questions, write your own explanation of the correct answer before reading the provided explanation. The effort of generation is the encoding event; bypassing it by reading choices first eliminates most of the learning value.
When should I start taking full practice exams?
Earlier than feels comfortable — when you have covered roughly 60% of the exam content. Early practice exams expose gaps across the full domain before you have studied everything, which creates focused motivation for the remaining content. Waiting until you have covered everything before taking any practice exams means you have no diagnostic data about your weakest areas until late in your study campaign, when there is less time to address them.
How is a brain dump different from a free recall exercise?
Both involve writing everything you can recall about a topic without consulting any materials. The practical difference is timing: a brain dump is done at the beginning of a study session to activate prior knowledge and identify current gaps before new learning. Free recall is done at the end of a study session to consolidate what was just learned. Both are retrieval practice; they serve slightly different functions in the session structure.
How many Anki flashcards should I create for an average certification exam?
Quality matters more than quantity. A deck of 200 well-designed application cards — testing decision-making, scenario reasoning, and concept relationships — is more valuable than 800 definition cards. For an AWS SAA-C03 with roughly 100 distinct concepts across five domains, 300-400 strong cards is a reasonable target. For CompTIA exams with significant acronym and port number content, you may legitimately need 500+ cards, since those details appear directly in exam questions.