The role of confidence in learning

Contents

Diagnosis

What role does asking students how confident they feel currently play in your teaching?
What challenges do you face?

Evidence

One of the most significant changes to my teaching is I regularly ask students to assign confidence scores (usually out of 10) before receiving feedback to know if the answer is correct.

There are several reasons to ask students to indicate how confident they feel in an answer before they know if that answer is correct:

1. It increases the potency of the Hypercorrection Effect

The Hypercorrection Effect, first posited by Janet Metcalfe, is the finding that individuals are more likely to correct misconceptions they are highly confident about after receiving feedback than those they have low confidence in. The Hypercorrection Effect increases learning by helping us overcome our overconfidence, interrupt confirmation bias, and increase our metacognitive awareness. The hypercorrection effect has been widely studied – check out this paper and this paper for more.

I hypothesise that asking students to assign confidence scores to their answers before they know if they are correct compels students to reflect on how they feel about an answer, thus increasing the potency of the Hypercorrection Effect.

2. It shows students what they really know

If I asked the student in the example above, if he knows how to do long division, he would tell me he does. After all, he has given a confidence score of 9. But he got that question wrong.

Likewise, the student in the example below is unsure of her knowledge of adding and subtracting with negative numbers (she has given herself a confidence score of 2) but got the question correct.

Confeicnde scores help improve students’ metacognitive awareness by forcing them to compare their perception of their understanding with their actual understanding.

3. It helps students invest more in their answers

If students have had the opportunity to consider how confident they feel in an answer, they are more invested in knowing if their answer is correct. I find confidence scores stop the mindless ticking and crossing that often occurs when the teacher is going through answers, or the causal glance students give when their work is returned marked.

My favourite example of this is the student in the picture below:

She was convinced she had the triangular prism question correct, as indicated by her confidence score of 10. When the teacher went through the answers, and she found out she was wrong, she took an intake of breath, quizzes her partner to see if he agreed, and then frantically made her correction:

I don’t think she have been this motivated to make the correction if she had not first made that codeine judgement.

4. It gives students an order to make corrections

Where should this student begin making corrections? Without confidence scores, students must start with the first question and work from there.

The Hypercorrection Effect suggests that the largest learning games come from correcting high-confidence errors. Hence, when students review their work, we can tell them to start with their highest confidence error and work from there.

5. It gives useful information to the teacher

Look at the answers of these four students, who have indicated their confidence scores in the top-right-corner:

This may help the teacher in determining which students to call upon:

Board 1 – No response: We could ask this student first to determine whether the response is due to a lack of effort or understanding.
Board 2 – High-confidence error: Next, we could call upon this student to clarify a common misconception. We must also return to this student at the end of the questioning process to ensure they can identify their mistake and articulate the correct process so they may benefit from the Hypercorrection Effect.
Board 3 – Low-confidence correct answer: Calling upon this student next may prove a magical moment as they realise that, despite their doubts, they are correct and are probably better at maths than they think.
Board 4 – High-confidence correct answer: This could be the final student we ask as they may provide a robust explanation that others can learn from

6. It reduces silly mistakes

I don’t think there are three words in the English language that have less impact on a student than Check your work. However, if you ask students to assign confidence scores after they have completed their work, as opposed to doing it after each question, they are forced to look back at their answers to determine their confidence rating. This often leads to them picking up on silly mistakes they have made.

Solution steps

Here are some ideas for making using confidence scores as beneficial as possible.

1. Assign confidence scores for a variety of work

Here are some things I see teachers ask students to assign confidence scores for:

Basically, you can do this in any situation where students do not yet know whether they are correct.

The mock exams use-case deserves a little more attention. Usually, students look at their final score and their mates’ final score, and then put the paper away. But if you ask students to assign confidence scores, then when you give students their papers back, you can challenge them to find their three highest confidence errors and correct them.

2. Choose your confidence scale

There is nothing in the research to suggest one confidence scale is better than any other. I use a scale of 0 to 10 as I find it helps students better discriminate between answers than a 0 to 5 scale does.

However, schools use lots of different approaches:

1 to 10:

1 to 5:

Emojis:

3. Tell students why you are asking them to do this

This is always an important step in the launch of a new initiative. Discuss the Hypercorrection Effect if appropriate. The good thing about assigning confidence scores is it is not a load of extra effort for students, so you are unlikely to get many protests or moans.

4. Have students assign confidence scores after their work, not as they go

The obvious thing to do is to ask students to assign confidence scores after each question when it is freshest in their minds. But I ask them to do it at the end of the retrieval opportunity. Making it a separate event elevates its importance. But there is another reason. I can think of no phrase that has less of an impact on students than check your work. In one ear, out of the other. But asking them to go back and assign confidence scores when they have finished forces them to review their work as they remind themselves how they felt about that question. Often students will pick up on little mistakes they made as they do this.

5. When students review their mistakes, tell them to start with their highest confidence error

This is the big one. Starting to think hard about the highest confidence error gives students the biggest bang for their buck as it allows the hypercorrection effect to kick in.

This is why I project full written solutions on the board for Low-Stakes Quizzes, instead of modelling each question one at a time, because I want students to focus their attention on the question they need at the point when they have the most attention and time available. And this will be different for different pupils. The highest confidence error is where students need to start, followed by the next highest confidence error, and so on.

6. Be clear in your messaging to students who have made lots of errors

Confidence scores provide a structure to the reviewing and correcting process that is often lacking, which is important for struggling students. What do you say to the student who has scored 2/10 on their Low-Stakes Quiz and is feeling overwhelmed with the work they now have to do? Well, with confidence scores we can say: Look, let’s get your highest confidence error sorted first. Don’t worry about everything else for now.

7. Improve the accuracy of students’ confidence scores

Some students do not think hard about their confidence scores. I have seen many students stick down a 5/10 to hedge their bets. This makes the review process less effective as the student may not know which errors were the high-confidence ones.

One way to deal with this is to ban the use of 5/10 so students have to at least commit to being fairly confident or fairly unconfident. Another option is to introduce confidence-based marking. Here students’ score per question is determined by their confidence score. So, if a student gives a confidence score of 7/10 for a question, they score 7 points if correct but lose 7 points if wrong. This certainly provides an incentive to be more accurate when assigning confidence scores. Beware, though, this can lead to negative overall scores, which may affect morale. Also, calculating these final scores may prove more cognitively demanding than completing the retrieval opportunity itself.

8. Ensure corrective feedback is clear

This last tip is obvious but should not be underrated. For the hypercorrection effect to work, students need to know not only that they are wrong, but why they are wrong. So, however we choose to go through the solutions to a retrieval opportunity, we need to ensure our explanation of the correct method is clear, and that students are willing to ask us if they still do not understand.

Implementation planning

How will you incorporate confidence scores in your teaching?