What Is Bloom's Taxonomy?
Bloom's taxonomy is a way of sorting the kinds of thinking a learning goal or a test question asks for. In its current form, revised in 2001, it has six levels: Remember, Understand, Apply, Analyze, Evaluate and Create, ordered roughly from simpler to more complex. Teachers use it to write course objectives and exams. You can use it from the other side: to work out what a question really demands, and to practice at that level before the exam asks.
The original 1956 version, by Benjamin Bloom and four colleagues, used nouns: Knowledge, Comprehension, Application, Analysis, Synthesis and Evaluation. The revision renamed three categories, swapped the top two and turned the names into verbs.
Here are all six levels on a single topic, so you can see what changes from row to row:
| Revised level (2001) | Original name (1956) | What you do | A question at this level |
|---|---|---|---|
| Remember | Knowledge | Recognize or recall facts, terms and steps | "What are the inputs and outputs of photosynthesis?" |
| Understand | Comprehension | Explain, summarize, classify, compare, give examples | "Explain in your own words why a plant needs light to make sugar." |
| Apply | Application | Carry out or use a procedure in a given situation | "A plant made three molecules of glucose. Use the balanced equation to work out how many molecules of carbon dioxide it took in." |
| Analyze | Analysis | Break material into parts and work out how they relate | "This graph shows growth under rising light. Why does the rate stop climbing, and what's limiting it now?" |
| Evaluate | Evaluation (was sixth) | Judge against criteria by checking and critiquing | "Two classmates designed experiments on light color and photosynthesis. Which design is stronger, and why?" |
| Create | Synthesis (was fifth) | Generate, plan and produce something new | "Design your own experiment to test whether light color changes the rate of photosynthesis." |
The topic never changes. The work the question makes you do does. That's the whole idea, and it's why the taxonomy is so handy for checking your own studying: if every question you practice sits in the first row, you've practiced one kind of thinking many times over.
Where Bloom's Taxonomy Came From
The taxonomy began as an examiner's problem, not a theory of the mind. Benjamin S. Bloom was associate director of the Board of Examinations at the University of Chicago, and he hoped a shared framework would cut the labor of preparing the university's annual comprehensive exams. According to the 1956 handbook, the idea took shape at an informal meeting of college examiners attending the 1948 American Psychological Association convention in Boston. From 1949, a group of measurement specialists from across the United States met about twice a year to build it.
David Krathwohl, one of the five authors, later described the goal as "facilitating the exchange of test items among faculty at various universities in order to create banks of items, each measuring the same educational objective." If two universities agreed on what "analysis" meant, they could swap analysis questions.
The result was Taxonomy of Educational Objectives: The Classification of Educational Goals. Handbook I: Cognitive Domain (1956), by Bloom, Max Engelhart, Edward Furst, Walker Hill and Krathwohl. Its categories ran from simple to complex and from concrete to abstract, and the authors assumed a cumulative hierarchy, in which mastering each simpler category was a prerequisite for the next. It was slow to catch on, Krathwohl recalled, because "taxonomy" wasn't yet a word educators used. It went on to be translated into 22 languages.
A second handbook, on the affective domain of interests, attitudes and values, followed in 1964. The original group never wrote one for physical skills, so the psychomotor taxonomies you may come across were built by others.
Forty-five years after the first handbook, a team led by Lorin Anderson and Krathwohl published A Taxonomy for Learning, Teaching, and Assessing (2001). The main changes:
- Verbs instead of nouns. Knowledge became Remember, Comprehension became Understand, and the rest became Apply, Analyze, Evaluate and Create.
- Create at the top. Synthesis was renamed Create and swapped places with Evaluation.
- Nineteen named thinking processes. Each level lists specific processes, such as recalling, summarizing, differentiating, critiquing and planning, and these get most of the emphasis.
- A second dimension. The type of knowledge (factual, conceptual, procedural or metacognitive) became its own axis.
- A looser hierarchy. In Krathwohl's words, "the requirement of a strict hierarchy has been relaxed to allow the categories to overlap one another."
The Six Levels of Bloom's Taxonomy, With Examples
The 19 processes are what make the levels usable, because they tell you what to actually do. Here's each level with the move that trains it and the trap that only looks like it does.
1. Remember
Retrieve facts, terms, dates and steps from memory, by recognizing them (as in multiple choice) or recalling them (on a blank page).
- Study move: close the book, write down everything you can, then check. That's active recall.
- Trap: rereading and highlighting feel like work at this level, but they're closer to looking at the answers than to retrieving them.
2. Understand
Build meaning from what you read or hear. The revision lists seven processes here, more than at any other level: interpreting, exemplifying, classifying, summarizing, inferring, comparing and explaining.
- Study move: explain the idea in your own words, then come up with an example your textbook didn't use.
- Trap: tidying up a summary somebody else wrote. If an AI wrote the summary, the AI did the understanding.
3. Apply
Carry out a procedure you know on a familiar task (executing), or pick one and use it on an unfamiliar problem (implementing).
- Study move: work mixed problems you haven't seen solved, so you have to pick the method before you use it.
- Trap: redoing the worked examples from class. If you remember the answer, you're practicing Remember.
4. Analyze
Break material into parts and work out how they fit: what's relevant (differentiating), how it's structured (organizing), and what point of view or purpose sits behind it (attributing).
- Study move: sketch a chapter's structure from memory, or ask what a source assumes and what its author wants you to conclude.
- Trap: a list isn't an analysis. If your notes are bullet points with no relationships between them, you haven't analyzed anything yet.
5. Evaluate
Make judgments against criteria, by checking (does the conclusion follow from the data?) and critiquing (how does it measure up to outside standards?).
- Study move: compare two explanations, solutions or sources, and decide which is better, naming your criteria.
- Trap: opinions without criteria. "The second one seems better" isn't an evaluation until you can say what makes it better.
6. Create
Put elements together into something new: propose a hypothesis (generating), design a way to do something (planning) and build it (producing).
- Study move: design an experiment, write an argument, or write the exam question yourself.
- Trap: decoration. A colorful mind map of the chapter headings is a copy, not a creation.
The Knowledge Dimension Most Guides Leave Out
Most posters of Bloom's taxonomy show only the six levels. The 2001 revision has a second axis, and for planning your studying it's arguably the more useful half. It separates what you're thinking about from how you're thinking about it:
- Factual knowledge: terminology and specific details, the basic elements of a subject.
- Conceptual knowledge: classifications, principles, theories and models, the relationships that make the facts hang together.
- Procedural knowledge: how to do something, including skills, techniques and methods, and knowing when each one is the right choice.
- Metacognitive knowledge: knowledge about thinking itself, including strategies, what different tasks demand and what you know about your own learning. This category was new in 2001.
Cross the four kinds of knowledge with the six levels and you get what the authors called the Taxonomy Table, a grid of 24 cells. Any learning goal can be placed in one or more of them.
For a student, the grid works as a quick audit. A few cells, using this article as the subject:
| Cell | What it looks like | Example question |
|---|---|---|
| Remember + factual | Recall a definition or a date | "When was the original taxonomy published?" |
| Understand + conceptual | Explain how ideas relate | "Why did the revision split knowledge from thinking?" |
| Apply + procedural | Carry out a method on a new case | "Use the question stems below to write one practice question per level for tonight's chapter." |
| Evaluate + procedural | Judge which method suits a problem | "Which of two ways to solve this problem works better here, and why?" |
| Create + metacognitive | Design your own learning plan | "Plan next week's study so it covers every level the exam will test." |
The metacognitive row is where studying gets smarter. Knowing what kind of question you'll face, and which strategy fits it, is itself a kind of knowledge in the taxonomy. The guide to metacognition covers how to build it.
Four Myths About Bloom's Taxonomy
Myth 1: Bloom's taxonomy is a pyramid
The colored triangle with Remember at its wide base is everywhere, but it isn't Bloom's. Researchers who studied how the taxonomy has been drawn found that the pyramid "was not part of" the 1956 handbook or the 2001 revision, and that "its origin remains a mystery." Pyramid versions were already in circulation by 1972, when one appeared in a University of Georgia report.
The shape suggests that the bottom layer is the least valuable, which is a stronger claim than the taxonomy makes. It also makes the ladder look more rigid than the authors left it (see Myth 3), and even the original handbook pictured the levels combining rather than stacking: "a particular simple behavior may become integrated with other equally simple behaviors to form a more complex behavior." In a 2017 blog post, Doug Lemov, the author of Teach Like a Champion, argued that the image is one reason teachers "talk with disdain about 'lower level' questions" and avoid recall practice. As he put it, "No one wants to be the rube at the bottom of the pyramid."
Myth 2: The lower levels matter less
The authors did see the higher levels as the bigger prize: Krathwohl wrote that objectives from Understand through Create "are usually considered the most important outcomes of education." But more important isn't the same as independent. You can't judge an argument about photosynthesis if you don't know what photosynthesis is. Facts are what higher-order thinking works on, even though, as the research below shows, drilling facts alone doesn't build it.
Myth 3: You have to climb the levels in order
The original authors assumed each level was a prerequisite for the next. The revision relaxed that, and Krathwohl gave an example: explaining, which sits in Understand, is more demanding than executing, which sits in Apply. Evidence for the strict ladder was thin from the start: reviewing the early studies in 1978, G. M. Seddon concluded that "the picture is uncertain." And in Agarwal's experiments below, students who practiced only with higher-order questions did well on higher-order tests without drilling the facts first. You don't have to finish the bottom layer before you're allowed to think.
Myth 4: The verb tells you the level
Lists of "Bloom's verbs" are handy for writing objectives, but they're a weak guide when you're studying. "Explain" can mean Understand or Evaluate, and "describe" can sit almost anywhere. The level depends on the thinking the question demands of you, and that depends partly on what you've already seen. If you've already seen the analysis worked out in class, reproducing it is recall, whatever the question's verb says. That's why the persuasive editorial in one of Krathwohl's classroom examples had to include "at least one supporting reason not specifically taught or covered in the class." Raters find the sorting hard too: in a 2004 study of short essay questions in veterinary education, pairs of raters agreed on a question's level between 34% and 77% of the time.
What Research Says About Studying at Higher Levels
Many courses test low, and some major exams test high
In 2010, Momsen and colleagues rated 9,713 assessment items from 50 instructors teaching introductory biology at 44 institutions. They rated 93% of the items at the two lowest levels, knowledge and comprehension. The instructors had all attended teaching workshops, so it isn't a random sample, but the pattern will be familiar to anyone who's taken a fact-heavy intro exam.
High-stakes professional exams often pull the other way. The 2026 test plan for the NCLEX-RN, the licensing exam for registered nurses in the United States, says "the majority of items are written at the application or higher levels of cognitive ability." And a 2008 study in Science by Zheng and colleagues rated biology questions from several exams and found that the MCAT asked for more higher-order thinking than the first-year medical school exams in their sample: 45% of points against 21%.
That gap matters because students prepare for the test they expect. In word-learning experiments by Finley and Benjamin (2012), students who had experienced a test format adjusted how they studied for it, and "participants who had expected that test format outperformed those who had not." Expect a fact test and you'll study for facts, which is a problem when the exam turns out to ask something harder.
Higher-order exams raised scores on both kinds of question
Jensen, McDaniel and colleagues (2014) compared two sections of the same introductory biology course for nonmajors, taught by the same instructor with the same materials. In one section, the quizzes and unit exams asked mostly for remembering. In the other, they asked students to apply, analyze and evaluate. On a common final exam, after adjusting for students' scores on a scientific reasoning test, the high-level section did better on the high-level questions (46% versus 41%) and also on the low-level ones (63% versus 59%). These were two intact classes rather than randomly assigned students, so it's a strong hint rather than proof.
Fact quizzes didn't transfer up, but higher-order quizzes did
Pooja Agarwal's 2019 paper in the Journal of Educational Psychology is the most direct test. Across three experiments, with college students and with sixth graders studying world history, students practiced with fact questions, higher-order questions or a mix, and took a final test two days later. In her summary: "higher order and mixed quizzes improved higher order test performance, but fact quizzes did not."
The first experiment, with 48 college students, shows the pattern clearly (proportion correct on the final test):
| Practice condition | Final fact questions | Final higher-order questions |
|---|---|---|
| Studied once | 54% | 44% |
| Studied twice | 54% | 49% |
| Studied once, then a fact quiz | 78% | 46% |
| Studied once, then a higher-order quiz | 53% | 72% |
Each kind of quiz helped the matching kind of final question and did little for the other, and a second read added little to either (the paper found no reliable benefit from restudying). In the later experiments, mixed quizzes that combined both kinds of question raised both kinds of score, although in the second experiment two quizzes of the same kind still beat one of each on the matching test.
What this means for you
These studies point to three plain rules:
- Find out the level of your test, and practice at that level or above it.
- Don't stop at fact flashcards for an exam that asks you to apply and analyze.
- Mix your practice. Facts still matter, so keep recall in the rotation alongside higher-order questions.
None of this contradicts the broader research on study techniques. Of the ten techniques John Dunlosky and colleagues reviewed in 2013, only practice testing and spaced practice earned their top utility rating, while rereading got their lowest. Bloom's taxonomy tells you what kind of practice questions to write. Spacing and retrieval tell you how to use them.
How to Use Bloom's Taxonomy to Study
Here's a routine for one chapter, article, lecture or video. It takes about an hour, spread over two or three sessions.
- Find the level of your test. Sort five past exam questions, practice questions or syllabus objectives by level. If most ask you to apply and analyze, your practice should too.
- Remember: highlight, then retrieve. Mark the definitions, claims and steps you'll need, sparingly. Glasp's web highlighter works on web pages and PDFs and saves each passage with its source. Then put the source away and recall what you can on a blank page.
- Understand: add a note in your own words. Attach a short note to each key highlight that explains it and gives an example the source didn't. Glasp pins the note to the exact sentence.
- Apply: work a fresh problem. If you learn from lectures or tutorials on YouTube, YouTube Summary puts the transcript beside the video, and its summary links key points to timestamps you can click. Find the worked example, highlight it in the transcript, then solve a different version without looking.
- Analyze: connect your sources. Put highlights from two or three sources side by side and map how they relate: what causes what, and where they disagree.
- Evaluate: judge, with criteria. Pick the weakest claim in your notes and decide whether it holds up. When other readers have highlighted the same page, Glasp can show you the passages they marked, which gives you a second view to argue with. You'll find more to compare in Glasp's community.
- Create: make something. Write the exam question you'd set, a one-page explanation for a classmate or a short argument of your own.
- Space it. Come back two or three days later and answer your questions again from memory, higher-order ones included.
Bloom's Taxonomy Question Stems
If you're stuck for questions, these stems work for almost any subject:
| Level | Question stems to reuse |
|---|---|
| Remember | What is...? List the... When did...? |
| Understand | Explain in your own words... What's a new example of...? How is X different from Y? |
| Apply | How would you use X to solve...? What happens to Y if X changes? |
| Analyze | What's the evidence for...? What does this argument assume? How do these parts fit together? |
| Evaluate | Which is better, and by what criteria? Is this conclusion justified? What's the weakest step? |
| Create | Design a way to... How would you test...? Write a question that would catch someone who only memorized this. |
Using AI to Write Practice Questions at Every Level
Writing a good higher-order question about material you've only just learned is hard, which is why students skip it. AI chat tools draft them in seconds, but they're less reliable than they look. In a 2023 study by Elkins and colleagues, teachers reviewing questions on biology and machine learning texts rated most model-written questions as useful, but questions requested at the Create level actually matched that level only 36% to 40% of the time, against 83% to 92% at Remember. A 2024 study by Scaria and colleagues tested five models: GPT-4 wrote high-quality questions 89% of the time, yet only about 70% of those sat at the level it had been asked for. Whether quizzing yourself with AI-written questions makes you learn more is still unsettled. Some early classroom studies weren't randomized, and a randomized 2026 preprint tested AI-supported practice as a whole rather than the questions themselves.
A prompt that works well:
Here are my highlights and notes from [chapter or article]. Write six practice questions, one at each level of the revised Bloom's taxonomy (Remember, Understand, Apply, Analyze, Evaluate, Create). Base every question on these notes. Don't include answers. After I reply, tell me which answers are weak and which level they actually reached.
Four rules keep it useful:
- The AI writes the questions. You write the answers. If it answers for you, it's doing the thinking you came to practice. The AI thinking trap explains why that habit costs more than it seems.
- Check the labels. As those studies found, a question requested at a higher level often doesn't reach it. Ask yourself whether each question needs anything the source didn't already say. If it doesn't, it's a Remember question in disguise.
- Keep it grounded in your sources. Glasp's AI chat answers from the highlights you've saved across articles, PDFs, videos and Kindle books, and points back to the passages it used, so you can check a question against the original. Ask it for questions at the level you want, then check each one against the highlight it came from.
- Use it as a sparring partner at Evaluate. Ask it to argue the other side of a claim in your notes, then decide which argument holds up and why.
Bloom's Taxonomy vs SOLO, Webb's DOK and Fink
Bloom's isn't the only framework for levels of thinking, and depending on your school you may meet one of these instead. They describe different things, so their levels don't line up one to one.
| Framework | Year | Levels or kinds | What it describes | Most useful to a student for |
|---|---|---|---|---|
| Bloom's taxonomy (revised) | 1956, revised 2001 | Remember, understand, apply, analyze, evaluate, create | The kind of thinking a goal or question asks for | Decoding exam questions and writing practice questions |
| SOLO taxonomy (Biggs and Collis) | 1982 | Prestructural, unistructural, multistructural, relational, extended abstract | How well structured a learner's answer is | Checking whether your answer connects ideas or only lists them |
| Webb's Depth of Knowledge | 1997 | Recall, skill/concept, strategic thinking, extended thinking | How complex a task or test item is | Reading tests and assignments labeled with DOK levels, as in the GED |
| Fink's Taxonomy of Significant Learning | 2003 | Foundational knowledge, application, integration, human dimension, caring, learning how to learn | Kinds of learning that interact, with no ranking | Thinking about what a course should change in you beyond exam skills |
For studying, SOLO is the best companion: Bloom's tells you what a question asks for, and SOLO tells you how good your answer is. A multistructural answer lists several relevant points. A relational answer shows how they fit together, and an extended abstract answer carries the idea into a new context. If your essays keep coming back marked "describes but doesn't explain," you're being asked to make the jump from multistructural to relational.
Frequently Asked Questions
What are the 6 levels of Bloom's taxonomy?
In the revised version from 2001, the six levels are Remember, Understand, Apply, Analyze, Evaluate and Create. The original 1956 version used the nouns Knowledge, Comprehension, Application, Analysis, Synthesis and Evaluation, with Evaluation at the top.
What is the difference between the original and revised Bloom's taxonomy?
The 2001 revision turned the categories into verbs, renamed Synthesis as Create and moved it above Evaluate, listed 19 specific cognitive processes and added a separate knowledge dimension (factual, conceptual, procedural and metacognitive). It also relaxed the strict hierarchy, so the levels can overlap.
Is Bloom's taxonomy a pyramid?
No. Neither the 1956 handbook nor the 2001 revision draws a pyramid, and researchers who have traced the taxonomy's images say its origin is unknown. The revision describes a rough order of complexity whose categories can overlap, not a stack you complete layer by layer. And remembering facts isn't a lesser kind of learning: it's the material the higher levels work with.
Who created Bloom's taxonomy?
Benjamin S. Bloom, an educational psychologist at the University of Chicago, led a group of college examiners who began meeting in 1949. The 1956 handbook was written by Bloom with Max Engelhart, Edward Furst, Walker Hill and David Krathwohl. The 2001 revision was edited by Lorin Anderson and Krathwohl.
How can students use Bloom's taxonomy to study?
Work out the level your test asks for, then practice at that level or above. For each topic, write one question per level, answer from memory and check your notes. Agarwal (2019) found that higher-order practice questions improved higher-order test scores, while fact-only quizzes did not.
What are higher-order thinking skills in Bloom's taxonomy?
The term usually means the top levels, Analyze, Evaluate and Create, and sometimes Apply as well, since studies don't all draw the line in the same place. Lower-order skills are Remember and Understand. Higher-order questions ask you to do something with knowledge rather than reproduce it.
Are Bloom's taxonomy verbs enough to tell the level of a question?
No. Verb lists help teachers write objectives, but the same verb can point to different levels. A question only tests higher-order thinking if you can't answer it by recalling an answer you've already seen. Reproducing an analysis that was worked out in class is recall, even if the question says "analyze."
Is Bloom's taxonomy outdated?
Not as a way of sorting what a question asks you to do, which is what it was built for. Universities still use the revised version to write course objectives, and Oregon State University's "Bloom's Taxonomy Revisited" (2024) adapts it for generative AI. The fair criticisms are about how it's used. Writers such as Doug Lemov and Daisy Christodoulou argue that the way it's usually presented, with knowledge at the bottom, has led schools to undervalue knowledge, and Lemov notes it was developed "pretty much before there was such a thing as cognitive science." Treat it as a map of question types, not a model of how memory works.
Is Bloom's taxonomy the same as Bloom's 2 sigma problem?
No, though both come from Benjamin Bloom. The taxonomy (1956) classifies learning goals and test questions. "The 2 Sigma Problem" (1984) is a later paper reporting that one-to-one tutoring, combined with frequent tests and corrective feedback, lifted students about two standard deviations above a conventional class. Later reviews of tutoring have found much smaller average effects.
Study at the Level You'll Be Tested
Bloom's taxonomy was built to help examiners write better questions, and that's still the best way for a student to use it. Before you study, find out what kind of thinking your test asks for. While you study, practice that kind of thinking on purpose, with questions you haven't seen answered. And keep the facts in the mix, because the higher levels are built out of them.
Glasp gives that routine a home. Highlight articles and PDFs, import your Kindle highlights, highlight lecture transcripts in YouTube Summary, attach your own notes, and quiz yourself on what you've saved at every level.
