Earlier this month, obscured by a barrage of tweets about funding allocations, the Department for Education claimed the first bit of good news for teachers’ workloads in a long while.

The department’s latest teacher workload survey (TWS) found that secondary teachers’ average, term-time working hours fell from 53.5 each week in 2016 to 49.1 in 2019. Primary teachers’ hours were found to have fallen even further, from 55.5 hours per week to 50 hours.

To anyone working 9-to-5 it will be obvious that an entire profession averaging 50 hours a week in term time is still under enormous pressure and now is not the time to declare the problem solved. Nonetheless, it will have been a relief to the department that, five years since they launched the ‘workload challenge’, working hours appear be falling.

The slight puzzle is that, only a few months ago, the authoritative OECD TALIS survey indicated that teachers’ working hours in England had risen over the past five years. TALIS and the department’s workload survey do examine slightly different teacher groups with slightly different techniques, but it would be very surprising if key stage 3 teachers were working more hours whilst all secondary teachers were working 10 per cent fewer hours.

How many hours do teachers really work?

There are several surveys that ask about teachers’ working hours, all of which ask a slightly different group a slightly different question in a slightly different way. The chart below includes five such surveys and, though the numbers are not directly comparable across surveys, each survey should be reasonably consistent over time.



The picture they paint is not entirely coherent, either in the number of hours or in the trends over time. However, most surveys tend to agree that teachers have been working 49 to 50 hours per week, in term time, throughout the past decade.

The outlier is the TWS, which estimated far higher workloads in 2013 and 2016. There are several possible reasons for the outliers. First, the differences in levels and trends could be related to the selection of respondents: the TWS has had a relatively low response rate of between 15 and 57 per cent over the past decade, while TALIS 2018 achieved a response rate of 83 per cent. Understanding society (about 80 per cent) and the labour force survey (about 60 per cent) tend to sample fewer teachers than the TWS but are not primarily targeted at measuring teachers’ workload and are less likely to select for teachers with a high workload.

Secondly, measurement of working hours is very sensitive to differences in the questions, sample selection, and survey method. In this particular case, the authors of the 2016 TWS cautioned that there was a major change in methodology between the 2010 and 2013 surveys.

Finally, measuring teachers’ working hours is an inherently hard problem. Not only do teachers work most of their hours outside the classroom, and often outside the school, but their tasks can be difficult to classify reliably. How many hours should they allocate for marking books while watching Poldark? Does a lunch club they started count as a break or work? The blurring of work and non-work time, along with most teachers working far more than their contracted hours, makes reliable measurement difficult.

We should therefore be very cautious about claims, based on a single survey, that teachers’ workloads have changed dramatically over the past decade. Most survey evidence suggests that the present, high workloads have been a feature of the job since at least the mid-90s.

Nonetheless, comparisons between the 2016 and 2019 TWS surveys are interesting because they show the differences between a group of teachers who were working an average of 55 hours per week and a group who are working fewer than 50 hours per week.

The difference in hours is due to less marking and planning

EPI’s previous workload report found that the main reason English teachers work longer hours than their OECD counterparts is the additional marking, planning, and administration they do. The same pattern shows up in the difference between the 2016 and 2019 TWS: the 2019 sample, working fewer hours, does so largely because of reductions in marking and planning. The reduction in time spent on pupil supervision is also notable.



Those areas were also highlighted in the government’s 2014 workload challenge and toolkits for reducing unnecessary workload in these areas are part of the workload reduction toolkit it later published. The latest TWS supplies further evidence that these areas may account for much of the difference between teachers’ working hours.

Teachers find lower workloads more manageable

One of the key findings of the research that followed the 2016 TWS was that high workloads only became a problem when they were perceived to be unmanageable. The two things are correlated but long hours do not necessarily lead to an unmanageable workload if a teacher is well supported and their work feels meaningful.

Reassuringly, the teachers in the 2019 TWS do report that workload is a less severe problem than the teachers of the 2016 survey reported. However, with most of the difference being between those who feel it is a fairly serious problem and those who feel it is a very serious problem, there is no room for complacency. Nearly 90 per cent of secondary teachers still consider workload a serious problem at their school, as do over 70 per cent of primary teachers.



This is probably the most important message of the 2019 TWS. It is easy to quibble about whether the headline number has really changed but, whatever the number, many teachers still believe that their workload is too high, and that too much time is spent on marking, lesson planning, and administration.

Unfortunately, it is unclear that the government’s efforts over recent years have made much impression on average working hours. One thing we do know, at least, is that if working hours did not rise from the mid-2010s, then workload is unlikely to be the only reason for the declining retention of early-career teachers over the same period. A silver lining, of sorts – though one that leaves us with more questions than answers.