Reform recently published two blogs introducing experimental research undertaken aiming to create a new way of comparing schools' performance across measures of economy, efficiency and effectiveness. The first blog focused on the importance of exploring these three areas of performance, it highlighted the rationale for chosen indicators and explained why school type comparisons were made. The second blog explored how so-called 'clusters’ of similar schools were created to try to enable fair comparisons. It also described the limitations of the method and the reasons why the school clusters might not allow us to truly compare like-for-like schools.
Even though the primary and secondary school clusters were formed using only two variables for each (meaning that they can vary more widely on other non-clustering variables), we hope the tentative results are still valuable to those seeking to better understand value for money in schools. To identify the overall performance of a school type (i.e. community schools, academy converters, etc.) the following simple weighted averages were calculated:
This gives an overall picture of how different school types perform compared to their respective cluster means (i.e compared to the average of similar schools). For example, it would allow us to say that all the community schools in our sample spend on average y per cent less than other similar schools on their teaching workforce. These weighted averages act as measure of aggregate performance of different school types. This provides an indication of how school types vary in priorities and performance.
This blog will focus primarily on schools’ teaching workforce. It will look at how schools perform across three indicators, described in the first blog of this series as Economy 1, Economy 3 and Efficiency 1 – that is, spending per teacher, spending on training and development and pupil-teacher ratios.
Spending on teachers
Staff pay is the single biggest area of spending in a school's budget, generally representing over 70 per cent of expenditure. Consequently, spending on teachers is an area of interest in the search for improved school productivity.
Evidence on the effect of teacher salaries on pupil outcomes is limited. However, two different trends have been identified. In 2013, Nicoletti and Rabe found that spending more on teachers in secondary schools had a small positive effect on test scores. This was mostly the case for high-achievers, native English speakers, pupils not on free school meals and white British pupils. Other research suggests that the impact of teacher salaries depends on differentials within the local labour market – if the market pay for people with equivalent qualifications in the surrounding area is significantly higher, pupil outcomes are affected negatively.
This seems to indicate that schools’ power to set teacher pay and academies’ power to set employment conditions could have a positive impact on pupil outcomes, as it allows them more flexibility to adjust to their local area. When Reform conducted a survey of academies in 2014, 59 per cent had introduced changes to pay policy since becoming an academy.
Figure 1 shows how both primary and secondary school types compare to their cluster averages. In both primary and secondary clusters, voluntary aided schools spend above average per teacher. Voluntary aided schools are faith schools where recruitment is determined by the religious governing body (as opposed to voluntary controlled schools, that tend to be faith schools run by the local government). One possible reason for this higher pay might be that voluntary aided schools are willing to increase their pay package to attract teachers of a specific faith, though further research would be needed to understand the cause.
Secondary academy converters are also spending above their cluster averages per teacher. Again, we can speculate as why this may be the case, but not draw any firm conclusions. In this instance, it could be connected to their locations, as academies are more likely to be in urban areas, and management could have decided to increase pay to compete with the local labour markets.
Pupil-teacher ratio
Much research has been done on the effects of class size, for which pupil-teacher ratio can be used as proxy, on student achievement. Evidence of its impact is mixed. In their seminal article, Angrist and Lavy, find that class size reduction fosters "a significant and substantial increase in test scores" for certain student cohorts. Yet, Hoxby finds that there is no significant effect of class size on student achievement. An overview of the evidence shows that smaller class-sizes have a positive effect on the most disadvantaged and lowest achieving students. Nevertheless, effects are moderate unless the class size is reduced to less than 20 pupils, according to the Education Endowment Foundation (EEF) Teaching and Learning Toolkit. This would obviously require an extraordinary level of expenditure, which is why it is not deemed to be a cost-effective measure by the EEF.
In the primary school sample, voluntary controlled schools have pupil-teacher ratios that are on average three per cent less than the average of other similar schools, meaning that they have less students per teachers. Academy converters and foundation schools have slightly higher ratios, but overall, differences are not stark across primary school types with similar pupil intakes.
Among secondary schools, foundation schools are displaying remarkably high pupil-teacher ratios compared to all other school types. They also appear to spend the least per teacher, as shown in Figure 2, potentially indicating that foundation schools in general try to create savings through their teaching workforce.
Training and development
The last indicator, directly relating to the teaching workforce, is school spending on training and development of staff. This indicator is the ratio of schools' spending per pupil, as pupil numbers reflect the size of the school.
The effectiveness of continuing professional development for teachers appears to be heavily dependent upon the form it takes. While some studies appear to suggest that it has no to minimal impact on student outcomes, other research indicates that long-term and sustained courses, with intensive mentoring, has greater effect than one-off workshops.
The approach of primary academy converters and foundation schools and secondary sponsor-led academies may therefore be better, if their increased investment means longer, more intensive courses (see Figure 3).
Given the financial strain, combined with the emphasis on improving the prospects of all pupils, schools must begin to take an evidence-based approach to their teaching workforce. The effects of having had good teacher are drawn out time and time again, and getting it right has the potential to transform the quality of outcomes more than any other interventions. Whilst this Reform research is highly experimental, it aims to contribute to the understanding of how different types of school governance may foster different types of decision-making. The results are far from conclusive, but it appears that school types overall show relatively similar approaches to their teaching workforce. This could either mean that the optimal approach has been identified, or perhaps more likely, that there is plenty of room for disruptive, innovative methods of teaching.
