Looking ahead: Assessing the accuracy of Education Policy and Data Center projections

Recent education projections from the UNESCO Institute for Statistics (UIS) and the Global Education Monitoring (GEM) Report show that the world is far off track from achieving educational outcomes under Sustainable Development Goal 4 (SDG4). Education enrollment projections, which use past data to predict future trends, are frequently used to set targets and establish expectations for progress. How accurate those projections are depends on the quality of the available data and reasonableness of the assumptions around the key driving forces in education – population growth, intake rates and retention. A certain amount of fluctuation is expected, and they cannot be accurate in all cases. Still, it’s helpful to know how much one should trust and rely on projection models, particularly when they are used for large-scale monitoring against SDGs.

Our team at FHI 360 has developed and used projection models since 2009. Recently, we tested the accuracy of our Education Policy and Data Center (EPDC) projection model from 2012 against five years of actual data from UIS. While our expectations of accuracy had been modest, we find that on the whole, where country contexts are stable, the projections are close to reality. Where we have been inaccurate, and sometimes highly inaccurate, is in countries that have experienced shocks to the education system. Here we share examples from our initial findings.