In many ways, education indicators are like a satellite navigation system that can help us reach our destination: Sustainable Development Goal 4. Based on the data, policymakers in the driving seat can select the best route and adjust their direction and speed as they progress or hit a roadblock. Ultimately, the success of the education system relies to some degree on the accuracy of the data used to inform decisions about how to best use resources to achieve goals.
3 papers to understand the out-of-school data
With this in mind, we want to draw attention to three technical information papers from the UNESCO Institute for Statistics that aim to help governments find their way through the multitude of data sources needed to produce, interpret and use accurate indicators on out-of-school children.
By exploring specific methodological issues, all three papers are designed to help national statisticians who want to understand why indicator values can vary depending on the data source, collection method, or technique used to calculate indicators.
This initiative involved more than 70 countries worldwide to develop and apply an innovative methodology to assess and use data from administrative records and household surveys to better identify how many children are excluded from education, who they are, and the barriers they face to being in school. Taken together, the papers make a strong case for the harmonization of data sources.
Why different estimates of out-of-school children can vary greatly
The first paper, on Estimation of the Numbers and Rates of Out-of-School Children and Adolescents Using Administrative and Household Survey Data, illustrates the significant variation that exists between different types of estimates of the number of out-of-school children – differing by millions of children in some cases – depending on the data source.
Take India, where precise measurement of school participation has been a challenge for years. Out-of-school rates from different surveys carried out in the country can vary by more than 10 percentage points, which translates to millions of children who are counted as either in our out of school.
The situation has not improved in recent years because there are still large differences in the attendance rates calculated from different data sources, as summarized in a study of data from India by the UIS and UNICEF.
The picture is further complicated by different starting ages for primary schools in different Indian States, making it difficult to estimate the precise age ranges for the numbers of children in or out of school across the whole country.
As well as setting out data issues related to the combined use of administrative and household survey data on participation in education, the paper outlines ways to improve the accuracy of out-of-school estimates by using compatible definitions of school enrollment or attendance and by reviewing the reliability of population estimates and students’ ages in enrollment records and survey data.
Such methodologies add significantly to global public goods for education – a key objective of the more than US$30 million that the GPE has invested in such activities through its Global and Regional Activities grants.
How age factors in calculating out-of-school numbers
The second paper, on Age Adjustment Techniques in the Use of Household Survey Data, focuses on the potential problems related to data by single year of age. It points out that estimates of the rates of children in and out of school are sensitive to the timing of data collection and the accuracy of age data.
In household surveys, household members’ ages are recorded at the time of data collection and the paper shows how statisticians can make the necessary age adjustments to better compare results of survey results based on different timing and duration.
One recommendation is to avoid age-specific attendance rates that are more likely to be affected by errors in age data than level-specific attendance rates.
How population estimates impact out-of-school rate
The third paper, on The Effect of Varying Population Estimates on the Calculation of Enrollment Rates and Out-of-School Rates, shows the extent to which estimates of participation in education are only as good as the underlying data on enrollment and population on which they are based.
In Brazil, for example, assumptions about under-coverage and fertility affect the accuracy of population estimates and, in turn, estimates of the number of out-of-school children and adolescents.
The findings from Brazil can also inform work by statisticians in other countries to improve understanding of the differences in primary and secondary enrollment rates according to varying population estimates and of the reliability of out-of-school estimates based on data from different sources.
UIS and GPE collaborate to get better data
The three technical papers show how all countries can benefit from the methodological work arising from the Global Initiative on Out-of-School Children.
This same collaborative approach is being taken by the UIS and UNESCO through the Capacity Development for Education (CapED) Program, which aims to bridge the gap between national education policies, data collection and use by helping countries develop and apply a range of tools to assess and improve their national statistical systems.
The UIS is also developing the standards, guidelines and methodologies to help countries report the data needed to produce internationally comparable indicators for SDG 4. To do this, the Institute works with countries and partners like the GPE through key initiatives such as the Inter-Agency Group on Education Inequality Indicators (IAG-EII), the Technical Coordination Group on SDG 4 (TCG) and the Global Alliance to Monitor Learning (GAML).
Together, the UIS and GPE are helping countries make the most of these new methodological tools in wider capacity-building initiatives to help ensure that every dollar invested in education has the desired impact and that no child is ever left behind.