Self Explanation: Talk Yourself Into Understanding

Two students read the same worked example in a chemistry textbook. One reads it through, nods, and moves on. The other stops at each line and asks, "Why did they do that here?" A week later, only the second student can solve a new problem of the same type. That stopping and asking is called self explanation, and it is one of the most practical study habits in the research literature.

This post shows what the evidence says, gives you a set of prompts that work, and walks through a full example. It is part of our practice testing guide, since self explanation pairs naturally with testing yourself.

The idea came from Michelene Chi's research on how strong and weak students learn from worked examples. In a 1994 study, Chi and colleagues gave 24 eighth graders a text about the human circulatory system. Fourteen were simply asked to explain each line to themselves after reading it. Ten read the text twice. The prompted group gained more from pretest to posttest, and the students who explained the most built the most accurate picture of how blood flows through the heart and body.

Kiran Bisra and colleagues published the meta-analysis in Educational Psychology Review. Learners who were prompted to self-explain outperformed those who were not, and the benefit appeared in sciences, math, social sciences and other areas, with both print and digital material. The authors called prompts to self-explain a potentially powerful intervention across a range of instructional conditions.

One honest caveat: a large 2013 review of study techniques rated self explanation as moderate utility, mainly because fewer studies had tested it in real classrooms over long periods. It is well supported, just less battle tested than practice testing.

Here is a short textbook solution: "How many moles are in 18 g of water? Molar mass of H2O = 18 g/mol. Moles = 18 g divided by 18 g/mol = 1 mol." Reading it takes five seconds. Self-explaining it sounds like this:

1. "Why 18 g/mol? Because hydrogen is about 1 and oxygen about 16, and there are two hydrogens, so 2 + 16 = 18."
2. "Why divide? Molar mass is grams per mole. To get moles from grams, I divide grams by grams per mole, and the grams cancel."
3. "What if it were 9 g? Then it would be half a mole. So the answer scales with the mass."
4. "Where else does this show up? Any time I need to go from mass to moles, such as before using a balanced equation."

Notice that the explanation added nothing new to the textbook. It made the reasoning visible, filled in the atomic masses, and tested the idea with a quick "what if". That is what you will need when the exam gives you 36 g of something else.

When self explanation goes wrong

• Explaining too fast. "That makes sense" is not an explanation. If you cannot say the reason in a full sentence, you have found a gap.
• Explaining only the easy steps. The steps you want to skip are usually the ones you do not understand. Slow down there.
• Never checking. Your explanation can be wrong. Compare it with the textbook or ask your instructor when a step still feels shaky.

Self explanation works best before and alongside testing. Explaining builds the model. Testing later checks that you can rebuild it from memory. The Feynman technique is a close cousin that explains to an imagined listener, and elaborative interrogation focuses on asking why facts are true.

A clear summary of a dense chapter can make self-explaining easier, because you can see the structure before you dig into the details. LearnKit can summarize a PDF or lecture into sections with key terms, which gives you a map to explain your way through.