Research on seating arrangements

Embed from Getty Images

Back in 2014, I wrote a piece for the Times Educational Supplement newspaper in the U.K. about seating arrangements. This article now sits behind a paywall. There has been some discussion recently on seating arrangements on Twitter so I thought it worth republishing the article below.

THE NOTION of “rearranging the deckchairs on the Titanic” has become a popular metaphor for pointlessness. What could be more trivial than discussing seating arrangements when we have so many other important matters to attend to? It extends to education, too: why bother worrying where a child is sitting when you should be educating them’!

The trouble with this view is that evidence strongly suggests that seating arrangements have a considerable impact on behaviour, and thus on learning.

For many years, there was little debate; nearly all schools put children in rows. That all changed in 1967 with the Plowden report, which took the emphasis away from whole-class teaching and presented plans for primary schools in which the desks were arranged in groups, with children facing each other.

Discussing the notional features of a good classroom, the report states that: “Teachers would be among the children, taking part in their activities, helping, advising and discussing more frequently than standing before the class.”

The majority of schools took up the suggestion. Writing in 1987, education professor Maurice Galton noted that, by the 1970s, “the traditional classroom layout with its rows of desks and static children had been replaced by sitting in groups “.

However, it is not at all clear that such a development has been beneficial to children’s education. Mainly, this is because group arrangements create fertile ground for misbehaviour.

Cracking concentration

Considerable research has been conducted into different arrangements in the classroom, notably by Australian professor of education Kevin Wheldall. His work shows a significant advantage for rows over groups when students are completing individual assignments.

In a typical study, children stayed on task for about 70 per cent of the time when seated in groups and 88 per cent of the time when seated in rows. This pattern persisted whether the children normally sat in rows or in groups, showing that the results were unlikely to have been caused by the children’s response to something new.

It is easy enough to understand the reasons for this. When children sit in groups, some of them will always have their backs to the teacher and be face-to-face with potentially distracting peers. This also explains why Wheldall found that the children with the lowest levels of on-task behaviour made the greatest gains in moving from groups to rows.

Get it together

But what about collaborative work? Clearly, the idea of grouped seating is better for collaboration between students. Studies show that activities which require communication between students, such as generating lists of ideas, are best served by group seating.

However, groups can still be problematic, even when the goal is collaborative work. There is a well-researched problem with group work known as “social loafing”. This is the tendency for some participants within a group to sit back and let others do the work. For this reason, in his extensive review of group work, renowned researcher Robert Slavin urges the establishment of two essential conditions in order for cooperative learning to have a positive effect.

The first of these is that each group must have a shared goal such as earning a certificate or some extra break time. The second is that the success of the group must depend on the individual learning of all group members.

This could be achieved, for instance, by not letting a group decide which of their members is to report back to the class. Instead, a teacher should make this decision at the point of feedback. That way, all the children in every group would need to have a clear understanding of the brief.

With these measures in place, groups can be successful, but elsewhere alternatives should be considered. For example, horseshoes and circle arrangements are ideal for tasks that require children to pose questions to the teacher or contribute to whole-class discussion.

So, perhaps the best model is one of flexibility. Different classrooms could have different arrangements and children could move rooms according to need. Or perhaps the choice of tables could be made with ease or rearrangement in mind.

Certainly, seating arrangements should not be viewed as irrelevant. The behaviour of students has been shown to be directly influenced by how their desks are arranged so teachers need to consider layout as part of a wider behaviour and learning strategy. We already spend time over seating plans, maximising learning and minimising misbehaviour with pupil placement; it is worth arranging our rooms to meet the same aim.


Plowden, B (1967) Children and their primary schools: a report, Her Majesty’s Stationery Office.

Galton, M (1987) “Change and continuity in the primary school: the research evidence”, Oxford Review of Education, 13/1: 81-93.

Wheldall. K and Bradd, L (2013) “Classroom seating arrangements and classroom behaviour”, Developments in Educational Psychology, 181-195.

Slavin, R (1988) “Cooperative learning and student achievement”, Educational Leadership 41>/2: 31-33.


9 thoughts on “Research on seating arrangements

  1. The “data guru” referenced by the Economist article referenced below runs his school’s data platform on Excel. It involves twice termly assessments, which automatically alter not only setting but also seating arrangements for every class. Taking seating arrangements seriously is not just about general principles – rows, groups or horseshoes – but actually grappling with the detail.

    Managing this sort of problem with data information systems not only means giving teachers the right tools to do the job without destroying their social lives, but also that we start collecting the evidence for what seating principles (among and alongside other aspects of pedagogy) work (or “tend to work in a variety of different situations”, for any anti-data pedants who might be reading).

  2. If flexibility is key, then I think the next key element is teachers that can develop routines with their class to quickly shift the seating arrangement to match the instructional activity. Which requires furniture that allows this (not the two-seat tables in my university classrooms, makes it impossible for me to model these things). And of course the capacity to develop genuine cooperative tasks appropriate for the learning goal at hand. My student teachers need to learn both. It’s a common mistake for them to seat the class in groups ‘hoping’ that students will work together, only to find out that the task doesn’t really demand cooperation, so the students don’t. The Hastings & Wood paper from Karin Litzcke gets it right I think: group seating the whole time through would do damage rather than benefit learning. Working as a teacher the school director at one school decided that to stimulate cooperative learning, all rooms were to be arranged in group seating. Needless to say, a pointless measure. I’m not so keen on Slavin’s suggestion that the group goal be extrinsc such as break time/certificate. Group goals may be more content oriented (coming up with a shared definition, standpoint etc), and using clear cooperative structures helps a lot too.

  3. Stan says:

    Seen any research on how large a group is optimum. If it were 2 then the seating problem is solved. Even with up to 4 it is workable with rows.

    To illustrate that tweet about Ellenberg in this context if you did a study on the benefits of group work you might come up with a result such as groups of size N showed the benefit with a p-value < 0.05.

    If you used intervals you might come up with groups of size 1 to 3 had a 95% confidence that the benefit would be seen. This would indicate that it is likely the group work had no benefit as groups of 1 did equally well.

    If you had an result that groups of size 3 to 4 had a 95% confidence that the benefit would be see then you could conclude there was likely a benefit to the group work for the outcome measured.

    You could do the same thing with p-values indicating all the N with a p-value < 0.05 but this is easier to hack by simply omitting some N in the report.

    See the section on page 149 "Doctor it hurts when I P"

  4. Pingback: Blogs that tell you how to teach better – Thinking about teaching

  5. Pingback: Blogs that tell you how to teach better – Elephant lessons

  6. Pingback: Can meta-analysis be saved? – Filling the pail

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.