Low-Stakes Quiz

Contents

Diagnosis

What kinds of tests do you give your students?
What challenges do you face?

Evidence

Every maths department I visit has regular paper-based assessments built into its curriculum. These are typically end-of-topic assessments focusing on the most recent topic covered or end-of-term assessments that include questions on several topics studied over a given period.

The stakes of these tests are high. The teacher marks the tests and enters students’ scores into a mark book. These scores contribute to students’ end-of-year grades, GCSE predictions, and set changes.

Such tests are tools of assessment. Sure, the teacher may learn something formative from them – gaps in knowledge that they may address in future lessons – but assessment is the primary purpose of the tests.

This is a missed opportunity. As we discussed, when considering retrieval opportunities and desirable difficulties, shifting our (and our students’) mindsets away from viewing tests as tools of assessment and toward viewing them as tools for learning can have positive results.

In this section, will refer to such tests as Low-Stakes Quizzes

Solution steps

Over the last ten years, I have been experimenting with Low-Stakes Quizzes in mathematics in many different settings and contexts. When departments get them right, they improve students’ motivation and retention, becoming a key part of their mathematical diet.

While no single format will work for everyone, certain features raise the probability that a Low-Stakes Quiz will be effective and sustainable.

Part 1: Purpose

1. Tell students why you are doing Low-Stakes Quizzes

I first sprung Low-Stakes Quizzes on my students without warning. They did not know what had hit them. My students assumed this was some high-stakes test and, as such, became anxious. Some students didn’t want to answer for fear of being wrong. Others looked for opportunities to gain an unfair advantage. The learning opportunity was lost.

Instead, we need to communicate the purpose of Low-Stakes Quizzes to our students. They need to know that the quizzes are not typical tests designed to give grades and rankings, they are tools of learning. We can use the ideas here to help students see the importance of retrieval for learning.

2. Call them Low-Stakes Quizzes, not tests

A rebrand can help communicate the purpose. For many students, the word test has negative connotations. They may associate it with feelings of anxiety, embarrassment, or failure, or worry that it will lead to set changes or parents being informed. Rebranding what we are doing as a Low-Stakes Quiz helps break this association.

It also provides an opportunity to develop a shared language.

However we choose to run our Low-Stakes Quizzes, if we call them Low-Stakes Quizzes, everybody—colleagues and students—knows what we are referring to.

3. Tell colleagues why we are doing Low-Stakes Quizzes

Here is another mistake I made. Before my colleagues knew it, it was a departmental policy that every class would have a weekly Low-Stakes Quiz. Because I had not explained the rationale, my colleagues did not take them as seriously as I did. As a result, they would happen later in the week and towards the end of the lesson, when there was not enough time left to do the quizzes properly and students’ attention was dipping. Students complained, and my colleagues complained, so we had to stop the quizzes for a whole term. Once again, the learning opportunity was lost.

A far better approach I see in schools is for one or two colleagues to try out different Low-Stakes Quiz formats with one or two of their classes until they find a way that works well. They then present this to the staff, sharing the purpose of the quizzes, what works well, their mistakes, and the lessons learned along the way. From there, they build a departmental policy that colleagues are more likely to buy into.

Part 2: Logistics

In Motivated Teaching, Peps Mcrea explains that routines discard unnecessary decision costs and decrease the amount of information we have to process. We can harness this by building a routine around our Low-Stakes Quizzes.

1. Run the quizzes on the same day and at the same time each week

If students expect a Low-Stakes Quiz, they will arrive at the lesson better prepared, settle down more quickly, and be better positioned to switch their full attention to the task.

I recommend choosing a lesson each week (or fortnight) to dedicate to your Low-Stakes Quiz and sticking to it. Likewise, always start the Low-Stakes Quiz at the same time in that lesson to firm up the routine. I find the start of the lesson works best so we don’t run out of time and can respond appropriately if difficulties arise.

2. Standardise the format of the quiz

If the format of the Low-Stakes Quiz remains consistent, students will pay less attention to its structure, freeing up more attention for the content.

Here is a Low-Stakes Quiz from a school I supported. Each week, the questions changed, but the quiz format (one side of A4, 10 questions, space for working, space for confidence score, etc) remained the same.

3. Where possible, print the quizzes out

In maths, there is always the temptation to project questions on the board. It is quick and saves time and money in the photocopying queue. But for Low-Stakes quizzes, printing out is a good idea for these reasons:

Students are more focussed as they don’t have to keep looking up to see the next question
Students at the back of the class are not disadvantaged
You can ask a wider variety of questions, especially those which involve diagrams that the students need to write upon

Part 3: Content

Make the quizzes mixed-topic

Research suggests teachers should give exams and quizzes that are cumulative. This differs from the themed end-of-unit tests I dished out for most of my early career. Here are four reasons why having mixed-topic Low-Stakes Quizzes is a good idea:

They allow students to benefit from the spacing effect. Instead of being challenged to retrieve a topic once, students retrieve it several times over a prolonged period, boosting storage and retrieval strength each time.

Topic-specific quizzes may lead students to believe cramming is an effective revision strategy. If a quiz will be on one specific topic, and students hammer that topic the night before, they will probably do pretty well on it. If students repeat this strategy on all topic-specific tests throughout the year, they will become unstuck when encountering a less predictable test – such as a high-stakes GCSE exam.

Mixed-topic quizzes allow students to practice switching. All high-stakes tests in maths are mixed-topic. Answering a question on percentages, switching to a question on sequences, and then on angle relationships can be challenging. Having regular opportunities to practice such switching will be beneficial.

Students and teachers are at less risk of being followed by performance. If students are taught ideas in nice tidy blocks, and then assessed on those ideas at the end of those nice tidy blocks, they will probably do pretty well on them. If this continues throughout the year, we may conclude that students understand our subject pretty well – after all, that is what the data shows. However, we know learning happens over the long term, and the only way to see if something has stuck is to provide several opportunities for students to retrieve it. Mixed-topic quizzes allow for this.

My Low-Stakes Quizzes are 10 questions long, with each question on a different topic.

2. Schedule or track content

In the previous section, we discussed why it is a good idea to track the content you include in each of the four retrieval opportunities, together with an indication of how successful students were in retrieving the information.

Here is a snapshot of the system I use:

3. Quiz the things that are hard to quiz

Low-Stakes Quizzes are better suited to those harder-to-quiz topics that require students to have equipment such as measuring angles (protractor), constructions (compass), straight-line graphs (grids and rulers) than a Do Now, which usually lasts less time, or a Homework where students might not have the equipment to hand. You could bit the bullet and choose one week to get all your tricky-to-quiz questions out of the way, instead of spreading the pain.

4. Vary the types of questions you ask

Most of the Low-Stakes Quizzes I see – and have administered – are dominated by factual recall or procedural-type questions. These are, of course, important. But they are certainly not the apex of understanding. If that is the only type of retrieval practice we give our students, their understanding will likely remain shallow, and we convey the message that this type of knowledge is all we require.

Here is how we could construct a Low-Stakes Quiz:

Questions 1 to 7: Procedural style questions on content that is relatively new to students, that they have struggled with before, or that they have not encountered for a long time.

Questions 8 to 10: More challenging question types on content that students have retrieved successfully recently. The Pointon and Sangwin taxonomy is good for helping think of alternate question types to ask.

5. Make the first question the easiest one

Research shows that an easy first question in an assessment leads to a lower dropout rate and higher average scores.

Make Q1 of your Low-Stakes Quiz super-accessible to reduce the barrier to a student starting.

An easily accessible Q1 also means that if a student does not start, it can only be because they choose not to, instead of being stuck. This means you can hold them to account.

6. Aim for 80%

We want our Low-Stakes Quizzes to be accessible enough to students that they are motivated to put the effort in, but not too easy so that they can answer them without thinking. Aiming for a mean score of 80% is a sensible approach. This means using our knowledge of our students to select both the topics and the types of questions you ask accordingly.

7. Aim for 15 minutes

As we will discuss below, I want students working for around 15 minutes on the Low-Stakes Quiz, so I will design the questions accordingly. Once again, this means using our knowledge of our students.

Part 4: In the classroom:Answering the questions

Plan the timings

Below are my rough timings. As each Low-Stakes Quiz plays out, I will circulate the room and adjust these timings accordingly.

Working on the questions: 15 minutes
Confidence scores: 1 minute
Marking: 1 minute
Review cards: 12 minutes
Who got 8/10: 1 minute
TOTAL: 30 minutes

We will delve into each of these sections below.

I recommend one Low-Stakes Quiz per week, per class.

2. Students should answer the quizzes on their own and in silence

I want to know what my students know. I want my students to know what they know. The best way to achieve this is to start the quizzes in silence so students can work independently and in a calm, focussed environment.

3. Carefully consider giving hints

In an ideal world, students will work through our Low-Stakes Quizzes without support. That way, we maximise the benefits of retrieval, and students have a truer sense of the depth of their understanding of a question. However, students are likely to learn very little if they stare at a question, clueless about how to start. Hence, we might want to give hints.

Research suggests that hints need to be carefully considered. If hints are too easy, students don’t have to think, and they will not learn. Hints must also be avoidable, so students can first try the question without any support.

For Low-Stakes Quizzes, a worked example on a related question, printed on the other side of the quiz page or on a separate piece of paper, works well.

Part 5: In the classroom:The review

Have students assign confidence scores before the answers (1 minute)

We discussed confidence scores in the previous section. Confidence scores work particularly well with Low-Stakes Quizzes. All of these quizzes have room for students to write their confidence scores:

You will notice in my timings that assigning confidence scores is a separate event that happens after students have completed all they can on the quiz. This elevates its importance in students’ eyes, and provides a check-your-work-via-stealth as students review their answers to assign their scores.

2. Don’t model solutions, project up full written solutions

For many years, my approach was to go through how you do Question 1, then how you do Question 2, and so on. Often I would ask the students to describe their method for each question. I no longer do this for two reasons:

It takes ages
It is not the best use of student time

Let’s delve into that second reason. Imagine Charlie has Questions 9 and 10 wrong, but everything else is correct. Is it really in Charlie’s best interests to make him sit through the answers to the first eight questions? Sure, he may benefit from seeing how you and other students approach the solution, but not as much as he would benefit from spending extra time thinking about questions 9 and 10.

Or what about Katie, who has got nine out of the ten questions wrong? I am not sure her time is best spent listening to me go through all the answers, either. As I go through each question, I cannot dedicate the time Katie needs, so Katie is likely to be rushed, overwhelmed and demoralised. Instead, I want Katie to focus her attention on one or two of these questions and ensure she really understands how to do them. These questions will be her highest confidence errors.

So, now I project up full written solutions, complete with annotations.

These remain on the board throughout the marking process and when students create review cards (see below) so they can refer to them in their subsequent discussions. Thus, students can focus their attention and spend their time on the questions they need help with instead of waiting for me to speed up or hoping I slow down.

Students may still have questions, but they can get help creating their review cards through paired discussions and me if needed. If my final assessment reveals that one question is particularly troublesome, I can discuss it in detail.

3. Include the topic and references to support and practice

Alongside each worked solution, I write down the topic of the question and reference any additional support or practice available.

The topic allows students to label areas that need further support. The references – pages of a textbook or revision guide, or clip numbers on Dr Frost, MathsWatch or Sparx – direct students to where they can go for the port. This information will then go onto the review cards students are about to create.

4. Students mark their work with ticks and crosses (1 minute)

There is an argument for teachers to mark Low-Stakes Quizzes. Some students may prefer it, and it may help them take the quizzes more seriously. If the point of your Low-Stakes Quiz is to have an in-depth knowledge of what students understand, then you probably need to mark it.

But I let my students mark their quizzes. I want them to see the quiz as a learning tool, not an assessment tool. There is no incentive to cheat as I will not take the marks in. I want students to be able to identify what they can do and what they cannot in as little time as possible, so they can then dedicate time to learning from their mistakes. Sure, some students will take advantage of this and do very little, but you should be able to pick up on this as you circulate the room… and some students will find ways to beat any system.

At this stage, I ask students to tick or cross their answers and not make any corrections. A one-minute visible timer can help keep this stage short and sharp.

5. Students create review cards, starting with the highest confidence error (12 minutes)

I think asking students to correct their work is a waste of time. Nothing magical that happens in the act of frantically scrawling down the correct answer to a question if the answer makes little sense. Moreover, if students have made many mistakes and hence have many corrections to copy down, there is a decent chance they will make some errors with copying, defeating the whole object of corrections. I wrote more about this here.

Also, corrections on the original piece of work are in a format that makes revising or learning from them in the future tricky. We know from research that re-reading is, in most cases, a far inferior learning strategy to testing. And yet, re-reading is all students can do when corrections surround their work. If we want our students to test themselves to see if they can now do something they had previously got wrong – or indeed, if our students want to test themselves – then they would have to find the piece of work that contains that mistake (easier said than done if it is a loose piece of paper or a random page in an exercise book), and then somehow try to avoid seeing the corrections and correct answer when looking at the question.

So, after they have marked their work, I give students 12 minutes to create as many Review cards as possible – with a target of three.

They start by finding their highest-confidence error. With the highest confidence error identified, students begin creating their first review card:

Front of card:

Write the question the topic is from – I can help my students with this by labelling questions with the relevant topic
Write the reference to where students can go for further help or practice – I label each question with this as well
Copy out the question – if it is a long one, or requires a diagram, I will consider printing it out for students so they can cut and stick. I choose the option that will take the least time.

Back of card:

A brief description of the mistake they made – students may spot this themselves, or it can come from a discussion with their partner
The correct solution – I always project the full-written solutions on the board, but if students struggle to understand it, they can ask me or their partner. I encourage students to annotate the solution to increase its meaning.

A5 card works best for maths (white or pastel colours if we are getting specific).

In the end, students have a card that looks like this (front of the card on the left, back of the card on the right):

After doing this for a while, students end up with a stack of cards that have the potential to be a highly effective learning resource for the following two reasons:

1. They are personalised to the student

Everyone’s stack of cards will be different, containing questions specific to that student. The labelling of the topic on each card helps the student know where they need to focus their revision when they use other resources, such as websites, revision guides or videos, and the reference can direct them to somewhere specific (see below).

2. The cards are quizzable

This is the big one. Students can get out their cards at any time – the end of a lesson, during tutor time, etc. They start with the top card, look at the question, and on a second piece of paper try to work out the answer. They can then turn over the card and check. If they get the question correct, they put the card at the bottom of the pile. If they get it wrong, they study the worked solution, seek help if it doesn’t make sense, and return the card midway through the deck so it bubbles up to the top quicker.

Once students are in the flow of this, I encourage them to reflect on their original mistake, and see if they can understand why they made it.

More support and practice?

The reference on the front of the card could be a page number of a revision guide or textbook, it could refer to an online program the school uses, such as Dr Frost Maths, Sparks or MathsWatch, or it could be a shortened URL that takes students to a YouTube clip. If students attempt the question on the card in the future and either want more support because they are stuck, or more practice to develop fluency, they’ve got it in one place.

Where are the revision cards kept?

I don’t let my students take home their revision cards in case they lose them. I keep them either in plastic wallets or envelopes. This changes nearer to an exam, when I allow students to take them home. Some students take photos of their cards’ front and back and upload the images to digital flash-card apps such as Quizzlet. That way they always have a copy of their cards on them.

What other schools do?

Peer-to-peer support

If students struggle when creating the review cards, I encourage them to speak to their partner, and if that fails they can ask me for help.

6. Promote the message: “Helping others is the hardest job in the classroom”

But what about the student with just one review card to create, or even no cards? The temptation might be to come up with some extension material, but who will support the student and mark their attempt? Instead, I revert to Sammy Kempner’s mantra that Helping others is the hardest job in the classroom.

I tell students that if they have created their review cards and don’t have any questions, then the real work begins. Their job is to help students around them understand anything they are confused with and support them in creating their review cards. If I make a point of quizzing a student struggling on a question, and either holding to account or praising their partner depending on their answer, then there is an extra incentive to do a good job here.

7. Ask Who got 8/10? (1 minute)

Whilst I will not be taking in their scores (see below), I still want to understand where my students are struggling. So, I use Tom Sherrington’s suggestion and ask:

Hands-up if you got 8/10

And then say:

Wow, that is a good score! Well done! Which questions did you get wrong?

Students who got 8/10 feel good about their scores and are happy to tell you where they are struggling. When we ask who else struggled with these questions, more students will be willing, allowing you to determine which questions require further attention.

Part 6: After the quiz

Plan when you will reteach

After gathering information from the class, we can decide what to do about it. Are we going to try to fix any problems there and then in the lesson? Or are we going to make a note to come back to this topic at a later date?

If students struggle with a question, then the chances are they will need more than a model solution to understand it. So, we can’t include a similar question on a subsequent retrieval opportunity until we have had a chance to re-teach it. This does not mean spending a whole lesson on it. But it may mean another explanation, another worked example, and another We Do. Then, include a related question on a retrieval opportunity – A Do Now, Low-Stakes Quiz or Homework – within the next few days to see if things are moving in the right direction.

2. Resist the urge to record the scores

Data, data, data! I know it is tempting to collect students’ scores, enter them into a spreadsheet, and let conditional formatting go to work. Before you know it, you have a beautifully traffic-lighted piece of statistical artwork in front of you.

But here is the big question: What will you do with it? If it will inform your actions in the classroom, then there is a case for it. But couldn’t you have got that information in the way described above? That takes seconds versus the many minutes taken to produce a Question Level Analysis each week.

If you are going to average out the scores, then what does an average score, of, say 76% actually mean? And if you are going to track student progress, then because these are mixed-topic quizzes with different topics covered each week, we should not expect students to score higher and higher each week because the content is forever changing.

No, I am convinced that the best thing to do at the end of a Low-Stakes Quiz is to walk away.

Part 7: Low-Stakes Quiz combos!

Quiz-Homework-Quiz combo

When I interviewed then the Head of Maths of Michaela Community School , Dani Quinn, for my podcast, she explained how she ran weekly Low-Stakes Quizzes. I have since adapted that approach with several schools I have supported:

Let students take home the Low-Stakes Quiz they complete on a lesson where they are usually set homework.
Their homework is to review the topics in the Low-Stakes Quiz using their review cards and supporting references.
In the next lesson, they take another Low-Stakes Quiz using the same topics but with different questions
This second quiz is taken in and marked by the teacher, scores are recorded on a central spreadsheet that all maths teachers have access to, and papers are returned to students the next day with ticks, crosses and a score.
Here comes the controversial, but important bit. Each class has a benchmark score that will be set by the teacher based on their knowledge of the class. It is always above 80%. If students fall below this benchmark on Quiz 2, then parents are contacted, and students are placed in detention.

I know this sounds ridiculously harsh, but here is the rationale. Students know exactly what is coming up in Quiz 2. If they listen and are active during the review process, seek help if they are still stuck, and then put in the effort at home, then there are no excuses. Quiz 1 is a test of ability, initial understanding and prior knowledge, so it would be unfair to impose a performance-based sanction here. Quiz 1 is a learning tool. Quiz 2 is fundamentally different. Performance in Quiz 2 is determined by effort, not ability. Why should students not be able to get 90%+ on these tests? Of course, everyone makes daft mistakes, hence my reluctance to set the benchmark at 100%. But If we have high expectations for our students and give them every opportunity to meet these expectations, they will rise to meet them.

2. LSQ-DQ combo

I saw a nice idea in a school combining two of my favourite things: Low-Stakes Quizzes and Diagnostic Questions.

Students are given 10 minutes to complete a Low-Stakes Quiz independently:

The teacher then stops them, and asks them a diagnostic question related to Question 1:

Students answer on mini-whiteboards

The teacher responds accordingly, either asking the students to continue on the quiz, stopping the whole class for support, or supporting a smaller group. Crucially, once the correct answer has been revealed, students can review their answers to Question 1 on the Low-Stakes Quiz to see if they want to make any changes.

The teacher then does the same for another question on the Low-Stakes Quiz:

And so on.

This kept the students focussed, gave the teacher real-time information, and supported struggling students.

Want to know more?

Sammy Kempner describes how he uses the same questions with different numbers in Low-Stakes Quizzes

Implementation planning

Here are the ideas we have discussed:

Part 1: Purpose

Tell students why you are doing Low-Stakes Quizzes
Call them Low-Stakes Quizzes, not tests
Tell colleagues why we are doing Low-Stakes Quizzes

Part 2: Logistics

Run the quizzes on the same day and at the same time each week
Standardise the format of the quiz
Where possible, print the quizzes out

Part 3: Content

Make the quizzes mixed-topic
Schedule or track content
Quiz the things that are hard to quiz
Vary the types of questions you ask
Make the first question the easiest one
Aim for 80%
Aim for 15 minutes

Part 4: In the classroom:Answering the questions

Plan the timings
Students should answer the quizzes on their own and in silence
Carefully consider giving hints

Part 5: In the classroom:The review

Have students assign confidence scores before the answers (1 minute)
Don’t model solutions, project up full written solutions
Include the topic and references to support and practice
Students mark their work with ticks and crosses (1 minute)
Students create review cards, starting with the highest confidence error (12 minutes)
Promote the message: “Helping others is the hardest job in the classroom”
Ask Who got 8/10? (1 minute)

Part 6: After the quiz