Concept

Finding a useful explanation and being able to explain it yourself are two different things.

Issue

After a search, you can feel more sure than you can show without the source.

Solutions

Learn how to tell having an answer from understanding it, then check one explanation.

Say you’ve just read an explanation of a process you use at work. It makes sense while the page is open. Later, someone asks why one step matters. You remember the conclusion, but you can’t explain how the steps connect.

That gap gives you a clear question to look into. It doesn’t cancel out the value of what you found. The useful difference is between having an answer to hand and being able to explain why it works.

Use this module when you need to explain or judge something using your own understanding.

What you’ll get from this module

In this module

By the end of this module, you’ll have a clearer way to judge your understanding: explain something yourself, then check it. We’ll look at experiments on searching, confidence and how good people’s explanations really are. You’ll see why it helps to know what a good answer must include. Then you’ll choose a situation from your own life and check one explanation. You’ll finish knowing what your explanation supports and which gaps still need work.

When confidence and understanding part company

An online experiment: guessing a quiz score without the internet.

Adrian F. Ward’s 2021 research makes the difference easy to see. In Experiment 3, 159 people recruited online through Mechanical Turk were split at random. One group answered a first quiz with Google, and the other without it. Everyone then saw the correct answers. Next, they guessed their score on a new ten-question quiz and took it with no outside help.

People who had used Google guessed 5.95 correct answers on average. People who hadn’t guessed 4.58. Their actual averages were 3.73 and 3.17. The gap in actual scores was not statistically significant. So the experiment did not show that the earlier search led to better scores on the new quiz.

Why it matters: Both groups had to take the next quiz alone. Yet the people who had used Google expected to know more without it. Their higher expectations didn’t come with a clear gain in actual scores. That’s why the conditions of the next task matter when you judge what you can do.

Read the study comparing internet access and personal knowledge.

An online experiment: search can teach you something while confidence climbs too high.

Emmaline Drew Eliseev and Elizabeth J. Marsh’s 2023 study adds a useful comparison. Their first experiment looked at 295 US people recruited through Mechanical Turk, in three groups. One group searched for explanations. One read the same explanation text. One got no information. Everyone rated how well they could explain the topics. Then they wrote explanations with no outside sources.

Searchers rated their ability higher than people who read the matching text. But their explanations got lower scores. Even so, searchers scored higher than people who got no information.

Why it matters: The searchers did learn something: they could explain more than people given nothing. But feeling more confident didn’t mean giving the best explanations. Both findings matter. You can gain from looking something up and still have gaps to check afterward.

Read the experiments on search confidence and explanation quality.

Studies of explaining devices: familiar things can be hard to explain.

Leonid Rozenblit and Frank Keil’s 2002 studies asked people to rate how well they understood everyday devices. Then they asked them to explain some of those devices. Study 1 involved 16 Yale graduate students. Study 2 involved 33 Yale undergraduates.

People rated their understanding lower after trying to write explanations. The studies also asked questions that tested key parts of each explanation. Then people read expert descriptions and thought again about how well they had understood the devices.

Why it matters: You can recognize a device and know what it does without being able to explain how its parts work together. Trying to explain it makes you fill in those links. These small studies show how that attempt can reveal a gap in what you thought you understood.

Read the studies of the illusion of explanatory depth.

A controlled online experiment: a drop in confidence needs care to read.

Research by Ethan A. Meyers and colleagues in 2023 helps make sense of that drop. Their second experiment recruited 327 Mechanical Turk participants. It included a group who copied unrelated text instead of explaining a device.

Ratings of understanding fell after people explained the device they had rated. They also fell after people explained a different device, and even after people just copied text. The fall was bigger when people explained than when they copied.

Why it matters: Feeling less sure doesn’t tell you exactly where your explanation went wrong. In this experiment, the change in confidence spread to devices people hadn’t explained. To find a particular gap, you need to look at the explanation itself and spot what’s missing.

Read the controlled study of explanation and broader confidence changes.

Does feeling sure mean you understand it?

Two online experiments: make clear what a good answer looks like.

Björn Mattes and Stephanie Pieschl’s 2026 experiments analyzed 82 and 143 German-speaking adults. People answered questions that asked them to explain something. Some had internet access and some didn’t. Some were given clear standards for a good explanation and some weren’t.

Internet access improved both the quality of explanations and confidence. Four comparisons planned in advance did not find that internet access led to significantly more overestimation. But many participants were excluded. A mistake in planning the sample size also made the second experiment less able to detect the expected effect. Keep those limits in mind before reading much into the lack of a significant difference.

Why it matters: The standards cut overestimation more consistently when people judged their ability before answering than when they judged their answers afterward. In some conditions, people rated their answers below their actual quality. Knowing what a good explanation must contain gives you something specific to check. Simply feeling less confident still doesn’t tell you how good your answer is.

Read the experiments on explicit explanation standards.

An experiment before searching: make your starting answer visible.

Ward’s 2021 Experiment 5 tested writing an answer before looking it up. Its 134 Mechanical Turk participants were split at random into three groups: no Google, ordinary Google use, or writing their own answers before using Google.

Compared with ordinary searching, answering first lowered how people rated their own thinking ability. It also lowered their guesses about how they would score on a later quiz without help. That later quiz was only predicted. People didn’t take it in this experiment.

Why it matters: Writing a first answer leaves you a record of what you could supply before searching. You can then compare it with what the source added. The experiment found a change in how people judged their knowledge. So you still need to check that comparison against the answer itself.

Read the experiment on answering before searching.

Answer first, then search

A randomized experiment with students: a different think-first task.

Ju Ae Kim and colleagues’ study, published online in 2025 and in a 2026 issue, tested a different task before searching. It split 104 undergraduates in Seoul at random into two groups. One group brainstormed answers to three questions before searching. The other looked at a slide of keywords before searching. Both searched about modern pentathlon using Naver.

Why it matters: Afterward, the groups did not differ significantly in how they rated their thinking ability, how well they expected to do or how much study time they expected to need. Brainstorming about a topic is not the same as answering each question before looking it up. A first attempt can give you something to compare with the source. But you can’t assume every think-first activity makes your confidence more accurate.

Read the comparison of brainstorming and keyword viewing before search.

What this means for you

Write an explanation you can check. Does it include the steps that matter, why they connect and the conditions that change the answer? Comparing it with a source you can trust gives you more to go on than how sure you feel.

Decide what the task needs from you

Choose something you need to explain, an answer you need to judge or advice you are qualified to give. Ask what you need to understand yourself, and what information the real task lets you look up.

Explaining something from memory shows what you can supply yourself. Doing a task with a reference shows what you can do with that help. If the real work allows a reference, show that process too, and say what information it depends on.

If you haven’t searched yet, write your current answer and mark what you’re unsure about before you look anything up. Afterward, compare that answer with what you learned. Recognizing the answer on the page doesn’t show you could have given it beforehand.

Work out what a sound explanation must include

Use an explanation you can trust, a worked example or a qualified reviewer to work out what a sound answer needs to cover. Pick out the essential steps, why one leads to the next and the conditions that affect the answer. These are what you’ll check in your own explanation. Naming every stage isn’t enough if you leave out why the next step follows.

Say you’re getting ready to explain a work handover. You might know the order of its stages but not be able to explain why one stage has to come before the next. If a required stage isn’t finished, would the next step change? That’s a link your explanation may need to cover.

Check the real process documents or ask a qualified reviewer to settle the answer. The handover is a made-up example. Its correct procedure has to come from the process you actually use.

If two accounts leave you unsure about an essential condition, ask a qualified reviewer to settle it. Until then, record that part as unresolved. You need a trustworthy account of what the answer should contain. Being harder on yourself can’t supply the missing information.

Explain the connections, then ask what would change

Close the source and write the explanation in your own words, or draw a diagram. Show how the steps connect. Mark any link you’re unsure of rather than quietly filling it with a guess.

Then ask a new question that puts the explanation to use: what would change if one relevant condition changed? Choose a question whose answer you can check. For the handover, you could ask what happens next if a required stage isn’t finished.

To answer, you have to use the link between that condition and the next step. This helps you see whether your explanation holds more than a list of stages. Check your answer against the documents or a qualified reviewer. Making up the question doesn’t make your answer right.

Check your answer and work on what is missing

Reopen the trusted source or ask a qualified reviewer to check your explanation. Compare the essential steps, how they connect and the conditions that matter. Then check your answer to the question about what would change.

If a link is missing or an important part of your answer is wrong, name that specific gap. Learn what you need to fix it, close the source and try the explanation again. If you can’t check a critical part, keep it unresolved and look for the missing information or qualified help.

For this exercise, you’ve shown you understand the task only when all three conditions hold. Your explanation includes the essential steps and conditions. Your answer to the changed-condition question is correct. Checking finds no serious error in the explanation or the answer.

If a critical gap remains, you haven’t yet shown that you understand this task. Even a successful check only supports what you actually checked. It never gives you permission to work beyond your professional competence.

Check again when the task changes, when a gap appears or when you next need to rely on that understanding. Keep the checked explanation, and be specific about anything that still needs work.

What do you need to be able to explain?

How about your daily life?

Where do you need to explain why something works, rather than just find a conclusion? You might be preparing a handover, judging an answer or talking through a process you recently looked up.

What question would show whether you understand how the steps connect? What source or person could help you check it?

List situations where this difference would be useful. One good idea is enough. Add more if they help you choose.

A situation I want to understandWhat I need to explain or judge
________________________________________
________________________________________

My chosen situation: ____________________

Choose one situation to take into the exercise. You can change your choice if there’s no trustworthy way to check it.

Check one explanation

Start with your chosen situation. If you came straight here, choose a specific question you need to explain or judge using your own understanding. Keep the task within your competence. Never use this exercise as permission to go beyond it.

  1. Choose the question and what a good answer must cover. Write the question and why you need to understand it. If you haven’t searched yet, first write your own answer and what you’re unsure about. Then use a trusted source or a qualified reviewer to pick out the essential steps, links and conditions. If you can’t work these out, record what information or help is missing.
    Question, use and required content: ____________________
  2. Close the source and explain. Use your own words or a diagram to connect the steps. Mark gaps and doubts instead of filling them with guesses.
    My explanation and gaps: ____________________
  3. Ask what would change. Answer a new “what changes if…?” question that uses a link in your explanation. Choose one whose answer you can check against trustworthy evidence or qualified feedback. If the answer can’t be checked, keep that part unresolved.
    Changed condition, question and my answer: ____________________
  4. Check, correct and try again. Reopen the trusted explanation or ask a qualified reviewer for feedback. Check how the process works, its essential steps and conditions, and your answer about what would change. Fix specific gaps and try again without looking. If you still can’t check a critical part, record the missing evidence or help.
    What checking confirmed or changed: ____________________
  5. Record what the check shows. All three conditions must hold to show you understand this task. Your explanation includes the essential steps and conditions. Your answer to the changed-condition question is correct. Checking finds no serious error in the explanation or answer. If not, record that you haven’t yet shown understanding of this task, and name the critical gap.
    Demonstrated for this task, or unresolved gap: ____________________

If the real task allows a tool or reference, show that process separately and say what help it relies on. Keep the record with your explanation. Come back to it when the task changes, a gap appears or you next need to rely on the understanding.

Ask Jing

Become a Jingius.

Find out what you can explain without the source

Loading interactive tool…

Your next step

You’ve finished with a checked explanation, or a specific open question and the information or help it needs.

A note on the evidence

These experiments measured different things: how people rated their knowledge, the scores they expected and the answers they actually gave. Search can support learning, and how it relates to overestimation varies from task to task. Trying to explain something can lower confidence without making every judgment accurate.

The standards experiments excluded many participants. Seeing the standards earlier could also affect later judgments, and some people rated their answers below their actual quality. Use the findings to get clear on what an answer should contain, then check the explanation itself.

How we prepared this module

This module brings together experiments on search, explanations and the standards for a good answer. It uses a made-up work-handover situation to make the judgments clearer. Then it turns those ideas into a written explanation check.

Sources