Innovative Concept Tests in Educational Assessment

Most teachers have had this experience: you finish what feels like a clear, well-organized lesson, ask “does everyone understand?” – and the class nods. Then the exam results come back, and it’s obvious many students didn’t truly grasp the core idea at all. The disconnect between surface-level familiarity and genuine conceptual understanding is one of the oldest challenges in education. Concept tests – short, targeted assessments built around key ideas rather than rote recall – are one of the most research-backed tools for closing that gap, in real time, while learning is still happening.

What are concept tests? 🔗

ConcepTests are conceptual multiple-choice questions that zero in on a single key concept tied to a lesson’s learning goals. Unlike conventional multiple-choice questions that test whether a student can recall a fact, concept tests probe whether students can actually think with what they’ve learned – making predictions, drawing inferences, classifying new examples, or applying a principle to a novel situation. They are typically used alongside short mini-lectures and are designed to be answered quickly, in a low-stakes environment, during the flow of instruction.

The idea was pioneered in the early 1990s by Eric Mazur, a physics professor at Harvard University, who noticed that his students could solve textbook problems yet still hold fundamental misconceptions about how the physical world worked. Mazur developed peer instruction as a method to improve his students’ conceptual understanding of introductory physics, and concept tests became the central tool of that method. Since then, the approach has spread across disciplines – from biology and chemistry to the social sciences and humanities.

How concept tests differ from ordinary quiz questions 🔗

Not every multiple-choice question qualifies as a concept test. The distinction matters. A ConcepTest should not require students to simply retrieve a fact from memory – a question like “Which is an igneous rock?” tests knowledge, not understanding. A true concept test, by contrast, demands comprehension or application. It might ask a student to predict what happens to a system if one variable changes, to identify which of several definitions most accurately captures a concept, or to apply a known principle to an unfamiliar scenario.

This maps directly onto higher-order thinking as described by Bloom’s Taxonomy. Comprehension involves making predictions or recognising how variables relate, while application involves using rules or principles in new contexts. Both levels go far beyond the simple recall that dominates many traditional quizzes. When teachers design questions at these levels, they get far richer information about what students actually understand – and where their thinking breaks down.

The role of concept tests in formative assessment 🔗

Concept tests belong firmly in the category of formative assessment – assessment that happens during learning, not at the end of it. Formative assessments are employed while learning is ongoing to collect information on whether course objectives are being advanced and how teaching can be improved. They aim to surface misconceptions, identify strengths, and allow instructors to make adjustments before small misunderstandings calcify into persistent errors.

Concept tests can also serve a diagnostic function – used before instruction to reveal what students already believe about a topic. Pre-assessment, sometimes called a threshold knowledge test, is conducted before instruction to establish a baseline from which individual student growth can be measured. Running a brief concept test before a lesson lets the teacher see where misconceptions already exist and tailor the instruction accordingly. Post-instruction concept tests then reveal how much genuine understanding has developed.

The feedback loop is tight. Formative assessment is the diagnostic use of assessment to provide feedback to teachers and students over the course of instruction – and concept tests deliver that feedback almost instantaneously. When a teacher sees that 60% of the class has chosen the wrong answer, they know exactly where to focus the next five minutes.

How concept tests work in practice: the peer instruction protocol 🔗

Concept tests are most powerful when embedded in a structured classroom routine. The effective use of ConcepTests follows a simple five-step protocol that keeps the process moving without consuming large blocks of class time.

Step 1: Short lecture segment 🔗

The instructor delivers a focused mini-lecture, typically lasting between 5 and 15 minutes. This is not a full lesson – it’s a targeted explanation of one key concept that will be tested immediately afterward. Keeping the lecture brief is intentional: the goal is to give students enough context to engage with the question, not to cover everything before asking for a response.

Step 2: Individual response 🔗

The concept test question is displayed, and students consider it on their own for 30 seconds to a minute, then commit to an answer. They may respond by raising hands, holding up answer cards, writing on a slip of paper, or using an electronic response system (often called “clickers”). The instructor gives students one to two minutes to think, then asks them to record their individual responses. The key here is that every student must commit – there’s no drifting along with the class majority before forming an opinion.

Step 3: Peer discussion 🔗

After the initial vote, the instructor checks how responses are distributed. The optimal range of correct initial responses is between 35% and 70% – enough students have it right to support a productive conversation, but enough have it wrong that real discussion is necessary. Students are then asked to discuss with neighbors who gave different answers, for one to two minutes. This peer discussion phase is where much of the learning actually happens.

Unlike common informal class questions which typically engage only a few motivated students, this structured process involves every student in the class. And as Eric Mazur himself has observed, students who have only just grasped a concept are often better at explaining it to peers than an expert teacher – because they haven’t yet forgotten what it felt like to not understand it.

Step 4: Re-vote and debrief 🔗

After peer discussion, the class is polled again. Typically, the proportion of correct answers rises significantly – even after a short peer discussion, students can move from around 50% correct toward 80-90% correct. The instructor then either asks a student to explain the reasoning behind the right answer or provides a brief clarification. The cycle is complete and instruction continues.

If fewer than 35% of initial responses are correct, that’s a signal the question may be unclear, or the concept hasn’t been adequately prepared for. If more than 70% get it right on the first try, the question is likely too straightforward to generate meaningful discussion. Both are useful signals for refining future lessons.

Types of questions that make strong concept tests 🔗

Writing a good concept test question takes some deliberate thought. According to the Science Education Resource Center (SERC) at Carleton College, several question types work particularly well.

Interpreting questions ask students to shift between forms of representation – for example, reading a graph and identifying which stage of the rock cycle it depicts. Classifying questions require students to apply a classification scheme to a new example they haven’t previously seen. Summarising questions present several possible definitions or descriptions of a concept and ask students to choose the most accurate one – a deceptively demanding task that requires genuine understanding. Inferring questions ask students to find patterns or draw logical conclusions from presented information, rather than simply recalling a memorised relationship. Application questions involve using a known equation, rule, or principle to solve a problem or complete a task in a new context.

What all of these have in common is that they cannot be answered correctly by memorisation alone. A student who has passively sat through a lecture may recognise the vocabulary but will struggle to apply it – and that’s precisely what concept tests reveal.

What the research says about effectiveness 🔗

The evidence in favour of concept tests and peer instruction is substantial. A longitudinal study at Harvard University exploring the impact of peer instruction compared to traditional lecture found that normalised learning gains with peer instruction were regularly twice those observed with traditional lecture alone. Students also performed significantly better on exams compared to students taught through conventional methods.

The benefits extend beyond test scores. Research over 25 years has shown that courses incorporating peer instruction produce better conceptual understanding, more effective problem-solving skills, increased student engagement, and greater retention of students in science programmes. The method has also been tested beyond STEM – studies in the humanities and social sciences have reported similarly positive outcomes.

Crucially, a review by Vickrey and colleagues found that when used in a research-based way, peer instruction is associated with major improvements in students’ ability to solve both conceptual and quantitative problems. The gains are not simply the result of students copying the answers of stronger peers. Studies using follow-up transfer questions – new questions requiring students to apply the same concept differently – show that real understanding has developed, not just answer-sharing.

Grading and low-stakes design 🔗

One important feature of concept tests is that they work best when kept low-stakes. This assessment process works best in a non-threatening environment where students are not penalised for getting questions wrong. When students fear being marked down for incorrect answers, they hesitate to commit honestly and the diagnostic value of the exercise collapses. The goal is accurate information about understanding, not a grade.

Teachers have several options for handling this. Concept tests can be entirely ungraded, with students simply told that similar questions may appear on future exams. They can count toward a participation grade based on whether students responded – regardless of whether they answered correctly. Or they can contribute a small portion of the overall grade, perhaps 5-10%, with credit assigned for correct answers across many questions over the term. Any of these approaches preserves the low-stakes spirit while keeping students accountable for engaging with the process.

Concept tests across subjects and grade levels 🔗

While concept tests were first developed and extensively studied in physics, they have since been adopted across a wide range of disciplines and educational levels. Peer instruction is a highly engaging, active learning strategy that encourages students to create knowledge among themselves, and it scales effectively whether the class has 20 students or 200.

In a biology class, a concept test might ask students which of several scenarios correctly represents natural selection – separating those who understand the mechanism from those who have merely memorised the definition. In a history class, it might ask students to identify which of four events best exemplifies a particular political ideology. In mathematics, it could ask students to predict the shape of a graph before they calculate it. The format adapts readily to any domain where core conceptual understanding matters – which is to say, virtually every domain.

Teachers working with younger students can also apply the principles without using formal clickers or response cards. A show-of-hands vote, coloured sticky notes, or even thumbs up/thumbs down signals can serve the same function: every student commits to an answer, discussion follows, and the teacher gets immediate, actionable feedback on what the class understands.

Getting started with concept tests 🔗

For teachers new to this approach, the entry point is straightforward. Start by identifying one or two concepts in an upcoming lesson that students typically find confusing or commonly misunderstand. Write a multiple-choice question at the comprehension or application level – not a recall question. Make sure the wrong answer choices reflect actual misconceptions students are likely to hold, rather than obviously absurd options. Present the question after a short instructional segment, have students vote individually, then facilitate a brief peer discussion before revealing and explaining the correct answer.

The process doesn’t require any special technology. Low-tech tools such as flashcards produce the same learning gains as electronic systems – it is the pedagogy that drives improvement, not the platform. Over time, teachers can build a bank of strong concept test questions, refine them based on how students respond, and use them both before and after instruction as diagnostic and summative checks on conceptual understanding.

What do you think? If you were designing a concept test for a topic you teach regularly, which misconception would you most want to surface – and how would you turn it into a question that reveals genuine understanding rather than just memorisation? And how might shifting even a portion of classroom time from lecturing to structured concept testing change the way your students engage with difficult ideas?

References

1. https://serc.carleton.edu/sp/library/conceptests/index.html
2. https://mazur.harvard.edu/research-areas/peer-instruction
3. https://www.frontiersin.org/journals/education/articles/10.3389/feduc.2018.00033/full
4. https://serc.carleton.edu/sp/library/conceptests/write.html
5. https://poorvucenter.yale.edu/Formative-Summative-Assessments
6. https://en.wikipedia.org/wiki/Educational_assessment
7. https://eric.ed.gov/?id=ED470206
8. https://serc.carleton.edu/sp/library/conceptests/how.html
9. https://lsa.umich.edu/technology-services/news-events/all-news/teaching-tip-of-the-week/peer-instruction-for-active-learning.html
10. https://uwaterloo.ca/centre-for-teaching-excellence/catalogs/tip-sheets/peer-instruction-and-concept-tests
11. https://instructionalmoves.gse.harvard.edu/using-peer-instruction-improve-student-learning
12. https://pmc.ncbi.nlm.nih.gov/articles/PMC4353089/
13. https://blogs.kcl.ac.uk/activelearning/2019/05/02/peer-instruction-eric-mazurs-approach/