Deep Processing

Deep processing is the quality of the thinking you do when you first meet new material. That thinking means connecting the material to what you already hold, comparing it against what sits near it, judging which parts carry weight, and assembling the parts into something like a model. This thinking determines what you can remember later. An hour spent rereading, highlighting, or memorising the exact wording gives you an hour of contact with the material and leaves very little you can recall afterwards.

Each technique listed further down can carry that thinking. Each can also be performed with none of that thinking happening. The thinking is what produces the memory. A technique produces only the thinking that happens while it is used.

Dimensions of Learning is the five-part division that deep processing belongs to.

Core takeaways

• Deep processing is the thinking you do when you first meet material: connecting it to what you know, comparing it, judging what matters, and building a model. That thinking decides what you remember later.
• Rereading, highlighting, and rote memorisation give contact with the material and leave little to recall. Extra hours on those methods rarely close the gap, and they stop closing it once volume and difficulty rise.
• The same technique can be run deep or shallow. The result comes from the thinking, and the thinking cannot be seen from outside.
• Depth alone is not enough. Retrieval practice that matches how the knowledge will be asked for is also required, and retrieval practice can beat elaborative study on later tests.
• Deep processing feels slow and effortful, and that effort is the encoding. Methods that feel smooth are usually shallow.
• Changing this means changing a habit, which takes months of consistent, targeted practice. The test is whether you can teach the topic back with every source closed and produce a hard example from memory.

What the thinking changes

Two people can read the same page at the same speed. The first is occupied with holding the words: repeating them, trying to keep them, getting to the end. The second thinks about the material while the words arrive: what it connects to, where it could be used, how it would simplify, where it sits in the whole. Both read the same page. The second reader is left with more afterwards.

The difference between the two is in how meaningfully the material is encoded into long-term memory. Higher depth produces stronger retention and understanding that stays usable. Shallow processing produces thinner memory and understanding that stays on the surface. The shallow forms are rote memorisation, passive reading, rereading and highlighting, and isolated facts with no bigger picture.

The thinking that produces depth is a small set of questions running while the material arrives: what it relates to that is already known, how it could be applied, how it would simplify, and where it sits in the larger picture. Those same questions are how new information connects to prior knowledge, and they are how analysis and critique run. Higher-order work is two operations. The first is comparing information against other information to find similarities and differences. The second is judging which of those similarities and differences matter. Lower-order work is the same act with the comparison left out.

Building transferable models and using knowledge under pressure need this depth and also retrieval practice that matches the way the knowledge will be asked for. Depth alone delivers neither.

Material connected to what is already known lasts longer than isolated facts. Knowing why something is true and how it works produces more memory than knowing what it is. Explanations, comparisons, and implications do encoding work that a second highlighting pass does not. Extra time on a shallow method rarely makes up the difference. A shallow first pass has to be made up later with more hours and more strain. Making it up with hours works while the material stays easy. Once the volume or the difficulty rises, extra hours no longer make it up.

How knowledge gets held together and rebuilt

One structural mechanism is the way related ideas are held together so that a new fact has something to attach to. A schema is that structure. It changes in three ways. Construction builds a new model when arriving information does not fit any existing one. Assimilation adds information that fits reasonably well into an existing model. Reorganization, called accommodation under the standard name, restructures or replaces a model when contradictions or gaps force a rebuild. Running through this cycle is what makes knowledge more accurate, more connected, and more usable over time. The cycle runs on its own during learning. Schema Construction, Assimilation, and Reorganization is the same cycle run deliberately.

Why it works like a habit

A procedure can be applied mechanically. Deep processing cannot, because it is a quality of thinking during the hour of study. The same technique can be run at more than one depth, from holding the wording up to judging what matters. A technique produces only the thinking that happens while it is used. The Technique Is Only as Good as the Thinking It Produces is the check for this. It tests whether the technique produced the thinking. It adds no method of its own.

The difference between learners is in the thinking, and the thinking cannot be seen from outside. The visible method is the part that gets copied. So two learners running the same method get different results, and either result gets attributed to the method. A gap that looks like a difference in talent is often a difference in how thoroughly the material was thought about.

Most of how information gets processed is not under conscious control while a session is running. So the thing that has to change is a habit of thinking. Deciding to think harder does not change it. Habits change through repeated, targeted practice. Knowing about a habit does not change it. Self-Regulation decides whether the deeper habit is the one that actually runs under pressure.

Where it gets trained

• Bear Hunter System: questions written before the source is opened, the source worked into a map, the map pruned to what survives without it
• Prestudy: a first pass before the real one, so new material has something already known to connect to
• Mindmaps: puts the relationships on paper, where they can be judged and rearranged
• Thinking on Paper: writing the reasoning down, so the thinking has to finish. Reasoning kept in the head can stall half-formed.
• Syntopical Reading: several dense sources read against each other, in place of one after another
• Best-attempt Encoding: encoding as deeply as current skill allows, and treating the first structure as something to revise
• Layers of Learning: the depths a topic can be held at, and which one a session is aiming for
• Higher-Order Learning: comparing and judging as the operations that turn facts into a network
• Importance-Based Chunking: grouping by what matters, in place of the order the source happened to use
• Knowledge Mastery: the progression from recognising something to being able to use it
• Interleaving for Complex Problem Solving: mixing problem types so that choosing the approach is part of the practice. This is the deep-processing version of a technique that more often belongs to Retrieval.
• Deep Processing Practice: practice aimed at the processing itself instead of at a subject
• Deep Processing for Research: the applied version for research reading
• Deep Processing Tanking: the failure where strong processing compensates for a weak method until the goal gets hard enough
• Dimension Practice Tracks: the four-step track for training the dimension, in place of collecting techniques
• Deep Processing Challenge: a month of structured practice with the sessions laid out

The other dimensions it depends on

Self-Regulation decides whether the deeper approach actually gets used under pressure or fatigue, when control shifts toward habit. Metacognition shows, during a session, whether the thinking is going deep or idling. Self-Management creates the time, energy, and environment the work needs before any of it is possible. Mindset lets the learner keep going through sustained discomfort. Without it the discomfort becomes a reason to stop.

Retrieval is the other half of the pair. Depth builds richer structures. Retrieval strengthens access to those structures. Both are required. Retrieval practice can outperform elaborative study on later tests. This is why deeper encoding on its own is never the whole answer. Depth is also not automatically transferable. A session that has to survive an exam also needs encoding that matches the way the knowledge will be used.

What it costs and how to check it

Strong depth gives less time lost to shallow repetition, durable retention, and knowledge that stays usable. It does not by itself produce a career. The price is that it feels bad. Deep processing is slow, effortful, and uncomfortable by design. The mental load is the encoding itself. That load is what Cognitive Load is trying to cue. Methods that feel smooth are usually shallow.

Strong natural processing compensates for a weak method until complexity or volume shows the weakness. Tanking is the name for this. The failure shows when the stakes rise. The method was weak before that. What fails at that moment is the belief that the method was working.

The test of understanding is whether you can produce the material. Recognising it when you see it does not count. Two signs show that the structure holds. The first is an explanation that survives WPW, which means teaching the topic back with every source closed. The second is a hard example that you can generate from memory. Recognising an example someone else gives is not the same sign. If time on the material keeps growing while transfer does not, the gap is in depth. It is not in the hours spent. The fix is to change how the material is processed. Adding hours does not fix it. The change takes months of consistent practice. It does not show up after a few sessions.

How to practice this

1. While reading new material, ask what it relates to that you already know. Ask how it could be applied, how it would simplify, and where it sits in the larger picture. Notice whether you are answering those questions or only holding the wording.
2. Take one new piece of information and compare it with something nearby to find similarities and differences. Then judge which of those similarities and differences matter. Notice that the comparison is the step that makes the work higher-order.
3. When a new fact arrives, check whether it fits a model you already hold. Add it if it fits, build a new model if nothing fits, and restructure if it contradicts the model. Notice the model getting more accurate and more connected over time.
4. Close every source and teach the topic back, which is the WPW check. Then try to produce a hard example from memory. If both work, the structure holds.
5. Track how much time you spend on the material and whether you can use it on new problems. If the time keeps growing while the transfer does not, the gap is depth. Change how you process the material, and expect that change to take months of consistent practice.
6. Notice how a study session feels. Slow, effortful, and uncomfortable is how deep processing is meant to feel. A method that feels smooth is usually shallow.

Related pages

• Dimensions of Learning: the five-part division that deep processing belongs to
• Retrieval: access to what depth built; the other half of the pair
• Self-Regulation: whether the deeper approach is used when pressure and fatigue make the cheaper one more likely
• Self-Management: the time, energy and environment the work needs
• Mindset: what keeps sustained discomfort from becoming a reason to stop

Sources

• Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior. Depth of the encoding operation, not store duration, determines the trace.
• Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General. Semantic encoding beats structural and phonemic encoding on later recall.
• Anderson, J. R., & Reder, L. M. (1979). An elaborative processing explanation of depth of processing. In Cermak & Craik (Eds.), Levels of Processing in Human Memory. Connected material is retained better than isolated material.
• Ausubel, D. P. (1968). Educational Psychology: A Cognitive View. Prior knowledge is what arriving material attaches to.
• Chi, M. T. H., de Leeuw, N., Chiu, M.-H., & LaVancher, C. (1994). Eliciting self-explanations improves understanding. Cognitive Science. Actively constructing an explanation is itself encoding.
• Slamecka, N. J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory.
• 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. Elaborative interrogation rated moderate; rereading and highlighting rated low.
• Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior. Depth is not automatically transferable; encoding that matches the use is the other half.
• Karpicke, J. D., & Blunt, J. R. (2011). Retrieval practice produces more learning than elaborative studying with concept mapping. Science. Retrieval can beat elaborative mapping on later tests; both remain required.
• Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In Metcalfe & Shimamura (Eds.), Metacognition. Desirable difficulties: the harder method is routinely judged the worse one.
• Kornell, N., & Bjork, R. A. (2008). Learning concepts and categories: Is spacing the “enemy of induction”? Psychological Science; Kirk-Johnson, A., Galla, B. M., & Fraundorf, S. H. (2019). Perceiving effort as poor learning. Cognitive Psychology. The fluency illusion: learners take ease as a sign of effectiveness.
• Schwabe, L., & Wolf, O. T. (2009). Stress prompts habit behavior in humans. Journal of Neuroscience. Stress shifts control toward habit: the Self-Regulation bound.