The many faces of disadvantage
24th October 2024 by Timo Hannay [link]
Update 8th November 2024: This is the first of two posts about disadvantage in England's schools. Here is Part Two.
Update 25th October 2024: Further coverage here from Schools Week.
Update 24th October 2024: See also this summary from The Gastby Foundation.
Disadvantage is one of the most analysed and discussed characteristics of England's education system. In particular, a great deal of attention is rightly paid to academic outcomes for pupils who are eligible for the Pupil Premium (or, roughly equivalently, for free school meals) compared to their less disadvantaged peers. This gap has persisted for as long as it has been measured and, after years of steady if incremental progress, has once again widened in the wake of the COVID-19 pandemic.
But Pupil Premium (PP) eligibility is not the only measure of disadvantage, and educational outcomes are not the only way in which less fortunate pupils and schools fall behind. We have made this point before, but with wide a range of analyses now at our disposal and a new centre-left government in charge, it seems like an good time to review things. To that end, this post analyses certain aspects of disadvantage that are often overlooked, but have potentially important implications for developing effective solutions to the perennial inequalities in our education system.
This is the first of two posts. Part 2 will explore new ways of looking at the national picture. Meanwhile, this post will focus on some of the myriad ways in school-level experiences differ by level of disadvantage. It revisits a few of our previous studies as well as presenting some new analyses. Our sincere thanks to the Gatsby Foundation for their support of this work and some of the earlier studies cited.
In summary, we show that:
- Poorer schools appear to have more difficulty in recruiting and retaining teachers. They tend to advertise teacher vacancies at higher rates, have greater headteacher turnover and spend more on supply teachers. Some of these gaps have grown since the pandemic.
- There are large regional differences in the supply of newly trained teachers entering state schools, with London relatively high and increasing, while other regions are low and declining. Although regional differences do not translate directly into trends by disadvantage, they point to important underlying disparities that are very likely to affect poorer schools more than others.
- Poorer schools also spend more on staff development. This might be seen as a positive trend: they are investing some of their extra funding in improving staff effectiveness. But it could also be a sign that they experience greater challenges and have less experienced staff. In support of the latter view, poorer schools tend to have younger teachers. They also lose more teacher days to sickness leave. Given this, how can more experienced teachers be incentivised to work at the most challenging schools, and how can all teachers at such schools be better supported?
- Analysis of topics arising in Ofsted reports reveals very different preoccupations at poorer schools (eg, exclusions, absences and behaviour) than at more affluent ones (eg, wellbeing and individual subject areas). This provides objective evidence of the common anecdotal observation that life is qualitatively different in ways that are not fully reflected in narrow measures of academic outcomes.
- While Education Investment Areas (EIAs) can provide a useful lens through which to view geographical disparities, they are too blunt to be used for targeting interventions because EIA-level averages hide even greater variation within in each area. In general, the most effective level at which to intervene and support is the individual school. Fortunately we now have the data and analytical capabilities to do this.
- PP as an indicator of disadvantage has significant limitations because (a) it is binary and (b) it focuses on income deprivation alone. In contrast, real-world deprivation experienced by pupils, families and schools exists on a continuum and has a variety of aspects – including housing, health, crime and the environment – that do not always correlate closely with income. A more nuanced view of disadvantage would not place all schools on a one-dimensional scale but instead characterise them by both the kinds and levels of disadvantage faced by their pupil and local populations.
- These observations have potential implications for policies and mitigation strategies. Tackling disadvantage-driven disparities in educational outcomes requires more attention on 'upstream' factors such as the difficulty of recruiting suitably experienced teachers. In addition, superficially similar schools may in reality experience very different local demographic and socioeconomic conditions. Fortunately we now have the information and analytical tools required to target interventions in ways that recognise such nuances. We will explore this further in Part 2.
Teacher turnover
Since 2017, SchoolDash and the Gatsby Foundation have collaborated to monitor, understand and interpret the recruiting activity of secondary schools and colleges in England by tracking vacancies posted on their websites. This real-time, longitudinal data complements official vacancy statistics, which provide only annual snapshots. While information gleaned from websites cannot achieve the comprehensiveness of a census, we find vacancies for well over 90% of schools and colleges, indicating high coverage and representativeness, and therefore a strong basis on which to understand national disparities as well as changes over time.
The analysis shown in Figure 1 compares vacancy advertising rates at state secondary schools in different regions of England. The blue columns show activity levels in September 2023 to June 2024, while the red columns show the equivalent period in the last full pre-pandemic academic year, 2018-2019. After controlling for their different teacher populations (hence the 'Adverts per 1,000 teachers' scale), all regions have shown an increase in activity since the pandemic. It is also evident that activity levels tend to increase as one traverses the country from north to south. This might suggest that difficulty in recruiting is a particular problem in more affluent areas, and perhaps related to the tightness of local labour markets. There is likely some truth to this, but note that London has similar proportions of pupils who are eligible for free school meals (FSM) as the North East.
When schools are split into groups with low (<20% FSM), medium and high (>35% FSM) levels of disadvantage in their pupil populations, it is clear that those with higher levels of disadvantage have been recruiting at higher rates, and this gap has grown since the pandemic.
The overall picture that emerges here is one in which schools located in more northern regions tend to recruit at lower levels, perhaps due to looser labour markets, but this masks the fact that all across the country schools with poorer pupil populations tend to recruit at higher rates, perhaps because they are seen as representing less attractive career opportunities for teachers. An obvious consequence is that policymakers should consider how to better recognise and reward teachers who take on these roles.
(Use the menus below to explore different school groups and subjects. Hover over the columns to see corresponding values.)
Figure 1: Relative teacher recruitment rates at state secondary schools by location and type
A somewhat similar picture emerges when we look at headteacher turnover, as shown in Figure 2. Changes in headteacher have tended to be higher in more southern regions (though this pattern has weakened since the pandemic) and in less deprived local areas (particularly since the pandemic). But when we look by in-school disadvantage it's clear that poorer schools have tended to have higher headteacher turnover, and that this trend has remained consistent bon either side of the pandemic.
(Use the menu below to explore different school groups. Hover over the columns to see corresponding values.)
Figure 2: Headteacher changes at mainstream state primary and secondary schools
Another way to assess difficulty in recruiting permanent teaching staff is to analyse spending on supply teachers using the financial returns provided by all state schools. The regional picture (controlled for the different teacher populations) is shown below in Figure 3. With the notable exception of London, inter-regional differences here are more modest, and if anything southern regions tend to spend less. But here too, all regions have shown increases since the pandemic.
The pattern by in-school disadvantage level shows poorer schools spending more, and this gap has grown considerably since the pandemic. Schools located in poorer areas have also spent more on average, especially since the pandemic.
(Use the menu below to explore different school groups. Hover over the columns to see corresponding values.)
Figure 3: Spend on supply teachers for mainstream state primary and secondary schools
Teacher training
The flow of newly trained teachers shows marked regional differences, as shown in Figure 4. (Note that these refer to the regions in which teachers were trained, not those in which they ended up working, but we would expect these to be correlated.) Relative to its overall teacher population, London has a high proportion of newly trained teachers and this has risen since the pandemic. Other regions have lower proportions and all have seen declines since the pandemic.
Since we do not have access to this data by school, we cannot analyse it by in-school disadvantage level.
(Hover over the columns to see corresponding values.)
Figure 4: Newly trained teachers working in state schools as a proportion of all state school teachers
The amount spent on staff development also varies by school disadvantage level, as shown in Figure 4. Both before and after the pandemic, poorer schools tended to spend more on staff development, though they also showed the greatest decline over this period.
(Use the menu below to switch betwee school types. Hover over the columns to see corresponding values.)
Figure 5: Staff development and training spend per teacher by school type
It's possible to interpret this disparity in both positive and negative ways. Poorer schools might feel the need to spend more in order to make up for having less experienced teachers (see below) and/or exposing them to more challenging situations. On the other hand, those same schools may feel that they have greater scope to invest in staff development due to the additional funding they receive through PP payments and other government grants. These two possibilities are not mutually exclusive.
Only young once
One consequence of the long history of shortfalls in newly trained teachers is the fact that average teacher age is rising. In some ways this is good: older teachers generally have more experience. In other ways less so: they are also more expensive and closer to retirement. Importantly, there are also difference in the age distributions of teachers across poorer and more affluent schools. As shown in Figure 6, around 62% of teachers in the poorest primary schools were under 40 in 2018, compared to around 55% in the most affluent schools. By 2024, these proportions had fallen to around 59% and 52%, respectively, but the gap remained roughly constant.
(Use the menu below to view different years. Hover over the columns to see corresponding values.)
Figure 6: Age bands for primary school teachers by in-school disadvantage level
Figure 7 shows the same analysis for secondary schools. Here the gap is somewhat larger. In 2024, around 59% of teachers at poorer schools were under 40 compared to about 48% at the most affluent schools. Once again, the proportions of younger teachers have fallen across the board since 2018.
(Use the menu below to view different years. Hover over the columns to see corresponding values.)
Figure 7: Age bands for secondary school teachers by in-school disadvantage level
There are similar differences when we look by local income deprivation, as shown in Figure 8 for primary schools.
(Use the menu below to view different years. Hover over the columns to see corresponding values.)
Figure 8: Age bands for primary school teachers by local deprivation level
And Figure 9 for secondary schools.
(Use the menu below to view different years. Hover over the columns to see corresponding values.)
Figure 9: Age bands for secondary schools teacher by local deprivation level
In sickness and in health
We also see differences in levels of teacher absence due to sickness. As shown in Figure 10, since the pandemic these have risen across all schools, but they remain highest in those serving the poorest pupil populations. There is a very similar pattern by local deprivation level.
(Use the menu below to switch between school types. Hover over the columns to see corresponding values.)
Figure 10: Sickness leave among teachers at mainstream state primary and secondary schools
Ofsted observations
Another indication of the fact that life in schools serving poorer pupils is different to that in schools serving more affluent ones comes from variations in the content of Ofsted reports. Figure 11 shows the relative incidence of selected topics in Ofsted reports published since 2006. Topics appearing relatively more frequently in reports about poorer schools are shown as positive scores. These include mentions of exclusions, absences and EAL. Those appearing more often in reports about more affluent schools are shown as negative scores and tend to cover personal development and wellbeing, as well as particular subject areas such as computing, technology, languages and art. Note also that there have been some changes to the ordering of the topics since the pandemic.
(Use the menu below to switch between time periods.)
Figure 11: Relative frequencies of topics in primary school Ofsted reports – high FSM versus low FSM
The differences across secondary schools are somewhat similar, as shown in Figure 12. Here, too, there are some differences between the ordering of topics across all available reports published since 2006 versus just those published since the pandemic.
(Use the menu below to switch between time periods.)
Figure 12: Relative frequencies of topics in secondary school Ofsted reports – high FSM versus low FSM
Geography lessons
The previous government's Levelling Up White Paper, published in February 2022 , declared its intention to support educationally underperforming parts of England:
"The UK Government will drive further school improvement in England through 55 new Education Investment Areas (EIAs) in places where educational attainment is currently weakest."
On the face of it, this is sensible. English education is riven by geographical disparities and on average, these underperforming areas – 55 of a total of 152 local authority (LA) areas in England – need greater support. But that doesn't mean that every school or locality within an EIA is doing badly. On the contrary, the variation in school performance within each LA is much greater than the variation between LAs. For example, Figure 13 shows the distribution of FSM percentages across all schools in EIAs (red columns) and elsewhere in England (blue columns).
While schools in EIAs tend to be less numerous at the low end and more numerous at the high end, there is huge overlap. In other words, whether a school is located inside or outside an EIA is a very weak predictor of its level of disadvantage. The same is true for other measures of disadvantage, including PP and local deprivation (specifically, IDACI, the Income Deprivation Affecting Children Index), as well as variations in educational outcomes such as Key Stage 2 attainment, Attainment 8 and Progress 8.
For this reason, EIA performance and similar area-based measures may provide useful indicators of geographical disparities, but are a very blunt way of targeting extra resources or interventions.
(Use the menu to switch between indicators. Hover over the columns to see corresponding values and sample sizes.)
Figure 13: Distribution of disadvantage and performance metrics across EIA and non-EIA schools
Definitions of disadvantage
So far, we have been talking about disadvantage as if it's binary (yes or no) and one-dimensional (related solely to family income). In reality it exists on a continuum and takes on multiple forms, from housing and health to crime, employment and the environment.
To appreciate the first point, see Figure 14, which shows data for all primary schools that reported exactly 20% of their pupils to be eligible for FSM in 2023. Despite being identical on this measure, the home postcodes of these pupils vary widely in their levels of deprivation. Red bars indicate pupil home postcodes with IDACI values in the top 2.5% of all postcodes in England (ie, the most deprived); purple bars indicate those in the lowest 62.5% (ie, the least deprived); other colours indicate intermediate values.
At some of these schools, almost all pupils live in relatively affluent neighbourhoods (even if some of the families are themselves poor), while for other schools large proportions of pupils live in the most deprived neighbourhoods in the country. In short, these schools serve very different communities, even when assessed by income deprivation alone and despite them having identical PP scores. As we have previously reported, the same principle holds for secondary schools too. Such qualitative and quantitative differences are surely important to consider when crafting policies and remedial interventions.
(Hover over the bars to see corresponding values.)
Figure 14: Deprivation bands of pupil home postcodes for selected primary schools
What about forms of disadvantage other than income deprivation? The government's Index of Multiple Deprivation (IMD) combines seven components: crime, education, employment, environment, health, housing and income. Figure 15 shows the deprivation bands for these measures in the local areas around the same primary schools we saw above in Figure 14. Looking at the bands for the composite , we see once again that even schools with identical proportions of PP pupils are often situated in communities that experience very different levels of overall deprivation. Moreover, these bands can show completely different distributions depending on which aspect of disadvantage we're considering. For example, the distribution of around each school contrasts with those for, say, or (notice the patterns and orderings changing each time). This is because income, crime and environmental deprivation are only very loosely correlated with each other. The upshot is that the experiences of each of these nominally 'similar' schools are often completely different.
(Use the menu below to select a local deprivation measure. Hover over the bars to see corresponding values.)
Figure 15: Deprivation bands for local areas around selected primary schools
The takeaway message is that schools do not exist on a one-dimensional ranking from most to least deprived. Still less do individual pupils fall into two unambiguous groups: disadvantaged or not disadvantaged. Rather, schools and pupils alike experience different levels and forms of disadvantage. In other words, disadvantage (like other important attributes such as personality type or academic ability) is something to be characterised across multiple dimensions rather than reduced to a single number.
Does this mean that we have to treat each school individually as its own special case? Not necessarily. There are ways of grouping schools that capture at least some of these nuances more effectively than simply lumping them together based on PP percentages or geographical location. How to do this will be the subject of our next post.
(If you have questions or comments, please write to: [email protected]. SchoolDash Insights subscribers can explore the characteristics for individual schools and follow national trends in real time. Non-subscribers can request a trial account or demo. To keep up to date with more analyses like this one, sign up for our free monthly-ish newsletter.)
