Personnel development

18th November 2020 by Timo Hannay [link]

Update 18th November 2020: The Teacher Development Trust have also posted their take on these results.

There are many lessons to draw from the exceptional events of 2020. One of them is the importance of an education system that is flexible at every level, from coping with widespread national disruptions to responding to the individual circumstances of each child.

In recent weeks, we have written about a range of educational topics relating to the COVID-19 pandemic, including pupil wellbeing and edtech adoption. This post focuses instead on the people who must respond to these changes: school teachers and other staff. In particular, we will look at trends in school staff-development spend, updating our previous analyses to incorporate data from 2019. At a time when huge demands are being made of education professionals, it seems particularly important to revisit this topic. Once again, we are delighted to be collaborating on this with our friends at the Teacher Development Trust.

Summary:

• In 2019, reported staff-development spend at state schools in England amounted to just over £260m. In nominal terms, this has been flat since 2015. Allowing for inflation, it was lower in 2019 than in 2014.
• Among secondary schools, staff-development spending was higher in 2018 and 2019 than in previous years, but this has been offset by a decline in spending by primary schools.
• Spend per teacher in 2019 was £651 in primary schools and £476 in secondary schools. Though directly comparable figures are hard to come by, this appears to be lower than benchmarks across other industries. Only 14% of primary schools and 10% of secondary schools spent at least 1% of their budgets on staff development.
• There are considerable regional disparities: staff-development spending has tended to be higher in the south-east of England and lowest in the midlands and south-west.
• Large multi-academy trusts continue to show higher levels of per-teacher spend, but not to the degree previously observed in 2018.

Flatliners

Figure 1 shows staff-development spending levels across English state schools since 2012. In 2019, total spending was £261m, a decline of 8% compared to the previous year and almost the same as in 2015. The net annual growth rate since 2014 was less then 1%, so allowing for inflation, spending has actually fallen over this period.

As a percentage of total spend, secondary schools declined between 2018 and 2019, while primary and special schools have been in decline since 2016. Across all schools, spending in 2019 amounted to 0.55% of budget, well below the level in 2018 (0.61%) and about the same as the level in 2012 and 2013 (0.53%). On a per-teacher basis¹, primary schools spent £651 in 2019, while secondary schools spent £476. Across all schools, the average spent level was £585, which represents a slight reduction in nominal terms since 2015 (£587) and an average annual increase of less than 1% since 2014 (£565).

(Use the menu below to explore these different measures. Click on the figure legend to turn individual lines on or off. Hover over the graph to see corresponding data values and sample sizes.)

These average values might be unrepresentative of general trends if some schools showed very large year-to-year changes in expenditure. Figure 2 shows the proportions of schools in each year that have spent either more (green columns) or less (red columns) on staff development than in the previous year. Among all schools, every year since 2017 has seen a larger proportion of schools decrease spend than increase it. The same has been true for primary schools, while secondary schools bucked this trend only briefly in 2018, which also coincided with an increase in average spend. Thus the changes in average spend seen above are broad-based and not caused by statistical outliers.

(Use the menu below to explore the different school types, including special schools. Hover over the graph to see corresponding data values and sample sizes.)

In 2019, only 14% of primary schools and 10% of secondary schools spent at least 1% of their budgets on staff development. Across all schools, increasing staff development spend to at least 1% of budget would have almost doubled national spend, from £261m to £511m. To put it another way, in 2019, 46% of primary schools and 71% of secondary schools spent less than £500 per teacher.

Distribution of funds

Figure 3 shows the distribution of schools in terms of the proportion of budget that they devoted to staff development. Among primary schools (red columns), note how the mode (peak) shifts from 0.5% in 2018 to 0.4% in 2019. Notice also how the proportions of schools spending little or nothing (left-hand columns) go up over the same period, while those spending higher percentages (right-hand columns) go down. Essentially the same is true for secondary schools, though in this case the proportions of schools spending little or nothing are even higher.

(Use the menu below to explore these and other years. Click on the legend to turn different school types on or off. Hover over the graph to see corresponding data values.)

Regional development

Within this national picture, there are considerable regional differences, as shown in Map 1, which uses data corresponding to the whole period from 2012 to 2019. Across all schools, south-east England showed the highest levels of spend. London came top at £611 per teacher per year, 21% higher than the South West (£503). Secondary schools showed broadly the same pattern, though in this case the East Midlands were lowest (£373), with London 28% higher (£477). Among primary schools, the North West bucked this trend by showing the highest levels of spend (£707), 21% ahead of the South West, though the south-eastern regions weren't far behind.

(Use the menu below to explore the different school types, including special schools. Hover over the map to see corresponding data values.)

Trust but verify

Last time, we reported an interesting pattern by academy status and trust size, as shown in Figure 4. In 2018, schools maintained by local authorities (LAs) spent the least, single-academy trusts (SATs) and small multi-academy trusts (MATs) came next and large MATs showed the highest spending of all.

Looking at the latest data from 2019, that last part was still true, but smaller MATs spent less than SATs and LA-maintained schools. This was true across all schools, primary schools and secondary schools. It will be interesting see how it develops in 2020 and beyond.

(Use the menu below to switch between years. Click on the legend to turn different school types on or off. Hover over the graph to see corresponding data values.)

Incomparable?

How high or low are these figures relative to other industries? Directly comparable data aren't easy to come by. This 2013 report by the (now defunct) UK Commission for Employment and Skills (UKCES) suggests that employers were spending an average of £1,590 per employee per year on training². In contrast, Figure 1 showed 2013 spend of £573 per primary teacher and £369 per secondary teacher. However, we need to make some adjustments to ensure that we're comparing like with like.

For a start, the UKCES figure include the wages of those being trained, which accounted for about half of the cost. We don't believe that schools' expenditure figures include this, so need to adjust accordingly. This reduces the UKCES benchmark to £800 per employee.

On the other hand, the figures we calculated for schools are per teacher, not per member of staff. Teachers account for about 43% of state primary school FTEs and 55% of state secondary school FTEs. Adjusting for this, we get 2013 spend of £246 per primary school staff member and £203 per secondary school staff member – ie, about 25-30% of the UKCES benchmark. Since then, school staff development spend has risen by 2.1% a year in primary schools and 4.3% a year in secondary schools. Over the same period, the Retail Price Index (RPI) measure of inflation averaged about 2.5%. So if anything, primary schools have lost ground since 2013, while secondary schools have caught up a bit, but nowhere near enough to close the gap with the UKCES benchmark.

Does this UKCES figure really seem reasonable? This source puts per-head training expenditure among US companies at just under $1,300 (about £1,050) per head per year in 2019. And this source gives a very similar figure. On the (not unreasonable) assumption that US employers spend a bit more on training their staff than British ones do, this appears to corroborate the UKCES benchmark.

Unless we're missing something, then, education may be spending a lot less than other industries on educating its own. The irony is not lost.

As ever, we welcome your thoughts to [email protected]. In addition, users of our premium service, SchoolDash Insights, can access further detail on a wide variety of income and expenditure patterns at national, local and school levels – just go to the Finances section.

Footnotes:

1. The use of per-teacher values should not be taken to imply that all staff development spend is necessarily devoted to teachers alone, though presumbly the vast majority of it is. Note also that per-teacher and per-pupil figures use full-time equivalents (FTEs), not headcounts.
2. There is an alternative way of interpreting the UKCES data, which is to focus on the amount of money spent on external providers (£3.3bn in 2013) rather than the total spent on training employees (£21.3bn). Presumably the difference arises because many employers provide their own in-house training staff and resources rather than relying on third-parties or sending staff on external courses. We don't believe that this would be a fairer comparison because most schools don't have in-house trainers, and where they do (eg, large multil-academy trusts), this would presumably be captured as staff-development spend.

Why worry about white working-class boys?

16th November 2020 by Timo Hannay [link]

The UK Parliamentary Education Committee has launched an enquiry into 'Left behind white pupils from disadvantaged backgrounds' and will continue to take evidence on this tomorrow. One of the contributors at their last session was Mary Curnock Cook, a former head of UCAS (and friend of SchoolDash), who afterwards wrote an insightful post about her concern regarding low progression rates to university among white students, especially poor white boys.

At first sight, being concerned about this particular subgroup may seem perverse. Of course it is right to focus attention and interventions on poor children, who show persistently lower attainment and progression rates than their peers. But why focus on poor white boys in particular when whiteness and maleness are generally seen as indicators of privilege, not disadvantage? This post attempts to answer that question, at least in broad-brush terms.

In all the charts that follow, each dot represents a mainstream state secondary schools in England (of which there are just over 2,500). In particular, we are going to look at the relationship between poverty (as measured by the proportions of pupils who are eligible for free school meals), ethnicity (by looking at the proportions of pupils who identify as white British) and GCSE attainment (using each school's Attainment 8 score). We will also distinguish between the attainment of boys and girls. All data are the latest publicly available, which means they are from either 2019 or 2020.

First, let's look at the relationship between attainment and poverty:

No surprises here: schools with more poor kids tend to do worse. But notice just how big these differences are – schools where 0-10% of pupils are eligible for free school meals (FSM) achieved an average Attainment 8 score of 52.5 (equivalent to just over a grade 5 in each subject¹); among those with more than 40% of FSM pupils, the average Attainment 8 score was just 39.0 (ie, less than a grade 4 in each subject). In addition, girls typically do better than boys (by almost half a grade per subject), though importantly, both sexes are similarly affected by poverty.

What about the effects of ethnicity?

Interesting: in contrast to poverty, ethnicity seems to have no effect on attainment for either boys or girls.

So poverty matters but ethnicity doesn't? Not so fast. Take a look at this:

Here we can see that whiter schools tend to have fewer poor kids. Something doesn't add up: poor schools do worse, schools with fewer white pupils tend to be poor... but schools with fewer white pupils don't tend to do worse. What's going on?

Let's just look at a subset of the poorest schools – those where at least 30% of pupils are eligible for free school meals (of which there are about 350):

Here we see that there is a relationship between ethnicity and attainment: schools with more white kids tend to do worse. The difference between schools where 20% or less of pupils are white British and those where they account more than 80% comes to about half a grade per subject (ie, about the same as the difference between boys and girls). This is particularly significant among these schools because they show average levels of attainment around the grade 3/4 boundary, which is often important for progression into further or higher education.

When we looked across all schools in Figure 2, this effect was hidden because whiter schools also tend to be less poor, so the effects of ethnicity and poverty cancelled each other out.

On that note, could the trend in Figure 4 be caused by the whiter schools being even poorer than the less-white schools? No, on the contrary, even within this group of very poor schools, whiter ones tend to be a bit less poor:

This means the effects of ethnicity are once again being partially cancelled out by the effects of poverty, so they are almost certainly bigger, not smaller, than they appear in Figure 4.

Of course, this simple overview hides a lot of nuance: poverty comes in a wide range of types and degrees; ethnic minorities are not a monolith; and GCSE attainment is an imperfect indicator of life chances. Yet certain statistical rules of thumb are inescapable:

• In England, if you don't want to be poor, it's better to be born white
• But if you are poor and you want to do well at school, it's better to come from an ethnic minority
• Whatever your class or ethnicity, you're more likely to get good results if you're a girl

There are many biases here that deserve serious attention, but one implication is that if you're poor, white and a boy then the educational odds really are stacked against you.

How big a problem might this be? The DfE does not routinely release information on the intersections between pupil poverty and ethnicity, but we can estimate the rough sizes of these groups based on the data that are available. Across all mainstream primary and secondary schools in England, just under 1.4 million pupils are eligible for FSM. We also know that about 65% of all pupils identify as white British. As we saw in Figure 3, white British pupils are less likely to be poor, but even if we assume that they are only half as prevalent among FSM pupils, that still amounts to around 450,000 children, or 5.5% of the total school population, half of them boys.

As ever, we welcome your comments: [email protected]

Footnotes:

1. Attainment 8 scores can be converted into GCSE grades by dividing by 10. This is because, although eight subjects are included in the score, English and maths are double-weighted.

The impact of lockdown on children's education

11th November 2020 by Timo Hannay [link]

Over the last two years we have been fortunate to work with RS Assessment from Hodder Education in analysing their amazing MARK database of primary assessment tests. Together, we've been able to better understand the effects of factors such as deprivation, age and gender across a broad range of primary subjects and topic areas (see our previous white papers from 2018 and 2019).

Of course, this year such analyses take on increased significance and urgency. Children almost certainly suffered learning loss during lockdown, but what have been the sizes and shapes of these effects, and how should mitigation strategies be targeted?

To help answer these questions, the RS Assessment and SchoolDash teams have collaborated intensively over the last few weeks to crunch the results of attainment tests that hundreds of thousands of primary pupils across England have taken since their return to school in September. The results are worrying but informative:

• Younger year groups generally show bigger reductions in attainment than older year groups
• Children who are eligible for the Pupil Premium show larger average declines than those who are not
• Schools with higher levels of deprivation, situated in urban areas or located in the north or midlands also show greater declines
• There are often large differences between topics within the same subject

For full details, see our joint report and the accompanying interactive charts.

We hope that the insights provided will help educators, policymakers, parents and others to better understand the current state of children's learning and to more effectively target strategies for quickly bringing them up to more normal levels of attainment.

As always, we welcome your feedback and suggestions: [email protected]

Eighty years on: Planes, pandemics and provinces

7th November 2020 by Timo Hannay [link]

Update 21st November 2020: It has come to our attention that there are instances of ignorance, inattention and intemperance on the Internet. We are shocked, truly shocked. For readers who can't get through all 1,000 words below or parse the phrase “None of this is to imply a causal link”, please be aware that what follows is not an account of the direct effects of Second World War bombings on contemporary Britain. Rather, it is a critique of the political economy that has prevailed in the intervening eight decades. To anyone who was confused, we apologise for this gratuitous use of nuance. (And to anyone else who might claim that all this was really just a thinly veiled excuse for us to play around with Google Maps, we plead guilty as charged.)

Update 7th November 2020: Here is BBC News coverage of our analysis. We have also added some footnotes in response to reader questions and comments.

As England enters its second Covid-induced national lockdown, we also happen to be marking Remembrance Sunday exactly 80 years after The Blitz. Two very different airborne threats – one silent and invisible, the other all too brutally tangible – but both of them pervasive, chilling and deadly enough to represent defining moments in our national story. A suitable time, perhaps, to add some historical context to the contemporary number-crunching that provides the usual fare for this blog.

We will examine here the patterns of bombing that Britain endured during the Second World War, and its human cost. Then we'll jump forward in time to see how some of the worst-hit cities are faring today in terms of wealth, health, learning and happiness. Spoiler alert: don't expect a feel-good ending.

Battling Britain

Map 1 shows the sites of German bombings as recorded in the fascinating 'Bombing Britan' data set on the War, State and Society website. We have converted these into a 'heatmap' in which red represents the areas of most intense bombardment, while yellow and green indicate areas subjected to correspondingly lower levels of attack. (You can pan and zoom the map, but if you zoom in a long way then don't take the exact locations too seriously: bombing sites are generally specified to the level of towns, not neighbourhoods or streets.)

By default, we see data for the whole period from September 1939 to March 1945, but we can also view them by year. In 1939 (during the so-called Phoney War), there was little activity, but things picked up rapidly in 1940 and 1941, during the Battle of Britain and The Blitz, before subsiding in 1942 and then rising again in 1943 to reach a second crescendo during 1944. During the final months of the war in early 1945, with Germany already well on its way to defeat, there was once again relatively limited bombardment.

You can use the menus below to display specific months and years. Alternatively, click on the following links to automatically step through months or days, providing an animated visual history of the bombings:

Step through months | Step through days | Stop: September 1939

The data we have been looking at so far show bombings, but these do not necessarily correspond to casualties. For example, many bombs fell on Kent in part because German planes were in the habit of jettisoning leftover payload on their way home. When these untargeted bombs fell on sparsely populated areas, the human cost was inevitably lower.

Looking specifically at casualties (over all months and all years), we can see that these were much more concentrated in urban areas. London suffered by far the greatest number of injuries and deaths, but many other places were involved, including industrial cities in the midlands and north, and strategically important towns along the coast. This was especially the case in 1940 and 1941.

Use the menu above the map to switch between bombings and casualties. You can also click on the animation links provided above to see how casualty patterns evolved over time.

Figure 1 shows a timeline for these attacks. Bombings were particularly heavy in late 1940 and early 1941 (The Blitz) and then again in mid-1944 (around the time of the D-Day landings and Germany's V-1 rocket attacks), when activity was very focused on London and Kent. The casualty data also show two peaks, but their relative sizes are reversed: dead and wounded Britons were far more numerous during 1944 than during 1940-41.

(Use the menu below to switch between bombings and casualties.)

All in it together?

It is therefore not unreasonable that so many accounts of German bombing raids focus on London and the immediately surrounding areas. But this shouldn't obscure the fact that many provincial towns and cities also suffered greatly, particularly during the early part of the Second World War, when Britain felt so vulnerable and alone. Moreover, while London and south-east England have generally prospered since then, the all-in-it-together spirit that (legend has it) pervaded the country during wartime has not been so evident in the intervening eight decades.

To illustrate this, Figure 2 shows the 15 areas beyond London and the home counties that suffered the highest relative casualty rates as a result of wartime bombing¹. (Here, 'area' is defined as a present-day local authority and casualty rates take into account the size of their current populations².)

The results are striking. Today, almost all of these have child poverty rates well above the national average – and above even the average for the northern and western parts of England in which most of them are located. (The only glaring exception to this is Bath.) They also tend to show lower attainment among pre-school children, 7-year-olds and 16-year-olds, as well as lower academic progress up to the age of 16. Perhaps unsurprisingly given all this, they also show low participation in higher education. The same areas usually have very low house-price ratios (making property affordable, but also indicating low demand). Self-reported life satisfaction also tends to be low. On top of all this, COVID-19 death rates in many of these areas – specifically, those in the midlands and the north – have been much higher than in the country as a whole.

(Use the menu below to explore these and other indicators. Note that for some indicators the y-axis does not start at zero; this is to make differences easier to see.)

Levelling up or just levelled?

None of this is to imply a causal link between 1940s bombing raids and current social challenges in England's provincial towns and cities. It is merely to point out that some parts of the country – especially London and south-east England – rebounded long ago from their wartime destruction, while others continue to languish in a wide variety of ways: economic, educational and medical. Time to think about how to avoid confronting similar statistics on the centenary of The Blitz in 2040.

Footnotes:

1. Note that this omits specific targets on the coast of Kent because, in modern administrative terms, the whole of Kent constitutes one local authority.
2. This is normalised to 100 for the whole of England. It uses current population figures because we can't think of a good way of estimating wartime populations for present-day local authorities. If you can then please let us know: [email protected]