What the research says

Do Live Polls and Clickers Actually Help?

Classroom response systems have been studied for decades. What a meta-analysis found about learning and engagement, why hands up hides so much, and how to ask questions that count.

By the Giga Quizzes team4 min read
Round push buttons in red, green, yellow, purple and blue on a blue panel
Photo: Moises de Paula, Unsplash, Unsplash License, cropped.
In this post
  1. 1What the meta-analysis compared
  2. 2What it found
  3. 3Hands, cards or clickers?
  4. 4Why the results are modest, and what that means
  5. 5The limits of these studies
  6. 6Questions that make a poll worth it

Long before phones, lecturers handed out small remote controls called clickers so that every student could answer a question at once. Today the same idea runs on whatever device is in a student’s pocket, under names like audience response systems, live polls and classroom quizzes. With so many studies of clickers published, it is fair to ask what they add up to. This post looks at a meta-analysis that pooled them, a smaller experiment that compared clickers with raised hands and cards, and what both suggest about using polls well.

What the meta-analysis compared

Nathaniel Hunsu, Olusola Adesope and Dan Bayly started from a problem: individual studies of clickers pointed in different directions, and their findings were, in the authors’ words, largely mixed and inconclusive. So they collected the studies that compared classes which used audience response systems with classes which did not, and combined them.

They looked at two kinds of outcome separately. Cognitive outcomes are about learning: test scores, exam results, understanding. Non-cognitive outcomes are the affective side named in the paper’s title, how students felt about the class and engaged with it.

What it found

OutcomeEffect of using clickers
Learning (cognitive outcomes)Small but significant
Feelings and engagement (non-cognitive outcomes)Near medium
Hunsu, Adesope & Bayly (2016)

Two things stand out. First, the effect on learning is positive but small. Simply putting a clicker in every hand does not, on average, transform exam results. Second, the effect on how students experience the class is larger. Students in classes with response systems tend to report more positive attitudes and engagement than students in classes without them.

The more useful part of the paper is what made the effect bigger or smaller. The authors found that the subject being taught, the size of the class, and the way the clicker questions were used all significantly changed the results. In other words, the tool is not the treatment. What a teacher asks, and what happens after the answers come in, is.

Hands, cards or clickers?

A smaller experiment by Jeffrey Stowell and Jason Nelson shows where some of the benefit comes from. They ran simulated introductory psychology classes in which every group heard the same 30-minute lecture, then compared four ways of taking part: no response method, raising hands, holding up response cards, and clickers.

  • Participation was highest with clickers, followed by response cards. Both were significantly higher than hand-raising.
  • Emotion: the clicker group reported the most positive feelings during the lecture.
  • Honesty: students using clickers were more likely to answer the review questions honestly.

The honesty finding is the one most teachers recognise. When a question is answered by a show of hands, students can see which way the room is leaning before they commit, and many quietly follow the majority. A private vote shows you what the class actually believes, which is the only version of the answer worth teaching from.

Why the results are modest, and what that means

Small average effects on learning are not a reason to give up on polls. They are a reminder that a poll is a moment, not a method. A class that answers ten recall questions and moves straight on has had a pleasant few minutes. A class that answers one well-chosen question, sees that half the room disagrees and then talks it through has had a lesson.

This lines up with other research covered in this series. Peer Instruction uses exactly the same technology but builds a discussion and a second vote around each question, and its gains are far larger. The hypercorrection effect explains why a confidently wrong answer, revealed publicly and corrected, can be one of the best learning moments of the day.

The limits of these studies

Many of the studies used dedicated clicker hardware rather than phones, and the meta-analysis reports average effects across very different courses and subjects. Stowell and Nelson’s study used simulated classes rather than a real term of teaching, which makes the comparison clean but leaves open how the differences hold up over months. Neither paper tells you the right number of questions per lesson; that depends on what each question is for.

Questions that make a poll worth it

The moderator findings point to one practical idea: choose each question for what its answers will tell you. A few kinds earn their place.

Five kinds of question for a live poll

  • The prediction. “What will happen when…?” asked before a demonstration or an explanation. Students commit, then find out.
  • The misconception check. One question built so that the most common wrong idea is one of the options. The size of that bar tells you what to teach next.
  • The confidence check. “How sure are you?” after a hard question. Confident and wrong is a different problem from unsure and wrong.
  • The application. A familiar idea in an unfamiliar setting, to see whether understanding transfers.
  • The honest opinion. A question with no right answer, where a private vote gives a truer picture than any discussion would.

Then decide in advance what you will do with the result. If most of the room is right, confirm and move on. If the room is split, ask students to discuss and vote again. If most are wrong, stop and reteach. A poll that changes what happens next is doing the job the research describes; a poll that only takes the register is not.

Sources

  1. Hunsu, N. J., Adesope, O. & Bayly, D. J. (2016). A meta-analysis of the effects of audience response systems (clicker-based technologies) on cognition and affect. Computers & Education, 94, 102–119.
  2. Stowell, J. R. & Nelson, J. M. (2007). Benefits of electronic audience response systems on student participation, learning, and emotion. Teaching of Psychology, 34(4), 253–258.

More from the blog

All posts →