The EQOP project primarily uses an exceptionally complex data set, where administrative records on e.g. income and education for the entire population of Norway are combined with a large health survey covering multiple cohorts of children and their families, as well as genetic information on survey participants. We have set up a unique data base and code library to take full advantage of these data sources.
Our main research results can be divided into three areas:
First, we have focused on educational inequalities. Most importantly, we have shown that economic inequalities in language development are evident from about two years of age and increase by entry into formal schooling. Economic inequalities in achievement have increased in Norway over the last decade, and that it increases with age throughout compulsory schooling. Yet, the economic gradient in achievement varies considerably between schools—some schools manage to erase this. Through this work, we have established a framework for using population-based big data in the field of educational research, where it has very rarely been used.
We have also identified heterogeneity in neighborhood effects (i.e. that some children’s achievement is more affected by neighborhood disadvantage than others’), and expanded models for how family socioeconomics affect child outcomes through family processes by being conditional on child characteristics, but also quite universal across societies with different welfare systems.
Using data from the US, we have also demonstrated that that socioeconomic gaps in achievement are opportunity gaps, and that a holistic approach to educational opportunities across age and institutions is the most promising venue for reducing inequality in life chances.
Second, we have found that in Norway, the scale-up of ECEC starting in toddlerhood reduced the socioeconomic gaps in 5th-grade test scores by up to 50%. Moreover, using data from the US, which includes ECEC quality measures, we have found that sustained high-quality ECEC contributes to reducing socioeconomic differences in income in early adulthood. We have also found evidence for social selection into ECEC quality in a universal system like Norway’s, and shown that scale-up of ECEC across countries in itself does not have an effect on achievement. Finally, using multiple datasets from several countries, we have shown that there is little reason to be concerned that full-day ECEC negatively affects children’s behavioral development, and that early entry into ECEC has no negative (or positive) effect on parenting quality.
Third, to take full advantage of the data sources available, we have done some methodological work. This has included approaches to causal inference in non-experimental data, and developments in a type of statistical analysis (quantile regression) which allows us to investigate heterogeneity in distributions of outcomes in complex statistical models. We have also expanded methodologies for using genetic data to identify causal effects.
Our work on social inequalities in education, and on ECEC, has been extensively disseminated to the wider public through news media (including national TV and international news outlets, a blog post for the Brookings Institution in Washington D.C and features in the Hechinger report), and to practitioners through features and interviews in specialized journals for practitioner. The EQOP PI briefed the OECDs working group on the most recent Starting Strong report, and was interviewed for the OECD Education podcast Top Class. The PI and one team member participated in a Government appointed expert group on child poverty in Norway, and the PI has been meeting with the Minister of Education (twice) as well as several groups of national and international policymakers to talk about our research. In September, the PI I will present results from the EQOP project at a seminar by the Nordic Council.
