Better ways to measure school performance

Joel Hernandez
New Zealand Herald
4 October, 2019

For too long many New Zealanders have viewed low-decile schools as lower quality.

And it is easy to see why. Looking at most NCEA league tables, any parent will quickly see high-decile schools at the top and low-decile schools down at the bottom.

Reading Education Review Office (ERO) reviews of your local schools will tell you a similar story. Low-decile schools are disproportionately categorised in the poor-performing category compared to middle- and high-decile schools.

At the same time, many parents have heard from others in their community that the high-decile school across town is doing spectacularly in NCEA, music and sports.

Given all the information available, it is not surprising that many parents have flocked to high-decile schools in an attempt to get the best education outcomes for their kids.

While existing measures of school performance have their merit, all of them produce measures of school performance that are a complicated mix of family background contribution and school contribution. Producing better education outcomes for students is much easier for schools dealing with kids from highly educated parents in advantaged communities.   

Last month, the New Zealand Initiative released the results of more than a year’s worth of work in Statistics New Zealand’s data lab developing a better measure of school performance.

Our results show that after accounting for the contribution of family background, we found that most schools, approximately 80%, perform very similarly.

However, we also found 42 decile 1 and 2 schools that outperformed 75% of every secondary school in the country. In contrast, we also found 9 decile 9 and 10 schools in the bottom 25% of every secondary school in the country.

Better measures

Using data from Statistics New Zealand’s Integrated Data Infrastructure (IDI) – New Zealand’s largest research database – the Initiative developed a revolutionary school performance tool that provides the Ministry of Education and ERO a better measure of school performance.

Otherwise known as a contextualised value-added model, our tool is unique in New Zealand in that it separates the contribution of each school from the contribution of each student’s family background. This gives the Ministry and ERO better and more relevant information on secondary school performance.

While other models internationally do a similar job, the breadth of data on students and their families available in the IDI allows for far better modelling – to give far more relevant results.

Over the past year, the Initiative analysed data on nearly 400,000 students from 480 secondary schools over 10 years to determine how New Zealand secondary schools are performing after adjusting for differences in family background.

The results shown in the data visualisation below are in stark contrast to what many New Zealanders believe about the quality of our low-and high-decile schools. We find high-performing schools across all deciles.

Made with Flourish

Better outcomes

The purpose of our work was not to create new league tables; it was to get a better measure of how our schools are performing so that we can lift the absolute level of education in New Zealand.

Using this method, the Ministry and ERO could identify the top-performing schools in New Zealand and learn what types of practice are the most effective in improving the outcomes of our students. In doing so, the best practices can be disseminated across all schools in New Zealand.

Insights from our tool should not be limited to learning from the best schools in New Zealand. Annual reports could be sent to every principal at every secondary school in the country, providing them objective, fair and data-driven evidence on how their school’s performance.

For some schools this might reveal spectacular performance that might have otherwise remained hidden under current measures; for others, it may reveal underperformance that may have been concealed because of the advantaged communities that those schools serve.

In addition to evaluating individual school performance, this kind of analysis could be used by the Ministry to evaluate any education policy affecting secondary schools. At the same time, it could be employed by ERO to evaluate school interventions. Currently, there is very little work done to evaluate the effectiveness of both education policy and school interventions.

We must learn what works and what doesn’t to improve the education outcomes for students. Without better evaluation, and without better measures of school performance, we cannot get better outcomes for our students.

The work presented in our report, In Fairness to Our Schools: Better measures for better outcomes, demonstrates what can already be done with the data that is available to the Ministry. It is time for the Ministry and ERO to step up. It is time for better measures, better information, and better outcomes.

Stay in the loop: Subscribe to updates