“I Don’t Know” Should Not Be the End of the Thought

When a Computer Science Student Says “I Don’t Know”: Why the Answer Shouldn’t Come Immediately

“I don’t know.”

It is one of the most common answers a Teacher or Tutor will hear. It can be tempting to respond by explaining the solution, particularly during a one-to-one lesson where there is plenty of opportunity to give a detailed explanation.

But “I don’t know” does not necessarily mean that a student knows nothing.

In my experience teaching and tutoring Computer Science, it can mean several different things. A student may not understand the question. They may understand the topic but be unsure how to begin. They may remember part of the answer but lack confidence. Or they may simply have become accustomed to receiving help as soon as they encounter difficulty.

Finding out which type of “I don’t know” we are dealing with is therefore much more useful than immediately supplying the answer.

The Problem With Giving the Answer Too Quickly

Imagine a student is shown a programming problem and immediately says:

“I don’t know how to do this.”

If I solve the problem for them and explain every line of code, the student may understand my explanation perfectly.

But that does not necessarily mean they could solve a similar problem independently five minutes later.

Computer Science requires more than remembering information. Students frequently need to decompose problems, recognise familiar patterns, select an appropriate technique and then apply their existing knowledge in a new context.

This means that the process of working out what to do when you do not immediately know the answer is itself something students need to learn.

This idea is supported by research into metacognition and self-regulated learning. The Education Endowment Foundation describes metacognitive strategies as helping pupils to plan, monitor and evaluate their own learning, and its Teaching and Learning Toolkit identifies metacognition and self-regulation as a high-impact approach.

Turning “I Don’t Know” Into a Diagnostic Tool

Instead of treating “I don’t know” as the end of the conversation, I use it as the beginning of another question.

For example, rather than immediately explaining a networking question, I might ask:

“What part of the question do you recognise?”

If the student identifies the word router, I can continue:

“What do you already know about a router?”

We have now discovered that the student does know something.

With a programming problem, the questions might be different:

“What information goes into the program?”

“What output are we trying to produce?”

“Have you solved anything similar before?”

“What would the first step be?”

This allows me to establish where the student’s understanding actually stops.

That distinction matters. A student who has forgotten one keyword needs very different support from a student who does not understand the underlying concept.

Give the Least Help Necessary

There is an important balance here. Leaving a student completely stuck is not productive either.

The aim is not to repeatedly ask questions until the student becomes frustrated. Instead, I gradually increase the amount of assistance.

The Education Endowment Foundation describes a similar principle as “least help first”: begin with prompting, then provide clues, modelling or correction when greater support is actually necessary.

For example, suppose a student cannot remember how a for loop works.

I might initially ask:

“What programming structure could we use when we know how many times something needs to repeat?”

If that does not help, I can give another clue:

“We have used two main types of loop. Which one normally uses a counter?”

Only after those prompts fail would I move towards modelling the solution.

The student therefore receives the support they need, but they still perform as much of the thinking as possible.

Why This Is Particularly Important in Computer Science

One reason I find this approach particularly useful in Computer Science is that decomposition is already fundamental to the subject.

A programming problem that initially appears difficult can often be divided into several much smaller problems.

A student might look at an entire program and say that they cannot write it. But when I ask them separately how they would obtain the input, perform the calculation and display the result, they may already know all three steps.

The difficulty was not necessarily their programming knowledge.

The difficulty was organising that knowledge.

The same principle applies to theoretical Computer Science. A six-mark GCSE Computer Science question may appear intimidating until the student identifies what the command word requires, which topic is being assessed and which pieces of knowledge are relevant.

Questioning exposes those smaller gaps much more effectively than simply saying whether an answer is right or wrong.

Building the Student’s Own Internal Questions

Ultimately, the tutor should not always need to ask these questions.

The longer-term objective is for students to begin asking them themselves.

Research on metacognitive learning emphasises explicitly teaching and modelling this kind of thinking. Teachers can verbalise how they approach unfamiliar problems and use questions to help students reflect on their own strategies. Over time, students can become increasingly independent in planning and evaluating their work.

That is why I would consider the progression from:

“I don’t know.”

to:

“I’m not sure, but I know this part…”

to be significant.

The second student has started analysing the problem rather than simply waiting for the answer.

Eventually, I want to hear something closer to:

“I don’t know the answer yet, but I know how I can start working it out.”

That final word - yet - represents an important change.

“I Don’t Know” Should Be the Start of the Conversation

A good tutoring session is not simply about how many answers a tutor can explain.

It is about gradually making the tutor less necessary.

When students encounter something they do not immediately understand, they need strategies for deciding what they already know, identifying what they are missing and choosing what to try next.

So when a student tells me “I don’t know,” I do not necessarily see an empty space that needs to be filled with an explanation.

I see an opportunity to find out what they do know - and then help them take the next step themselves.

References

Education Endowment Foundation, Metacognition and Self-Regulated Learning: Guidance Report (2025). EEF Metacognition and Self-Regulated Learning guidance

Education Endowment Foundation, Metacognition and self-regulation - Teaching and Learning Toolkit. EEF Teaching and Learning Toolkit