Free child care in Oslo
Report 1: Following the children in third grade
In the period 2011 to 2014 the research department at Statistics Norway and Fafo undertook an evaluation of a policy that provided free child care for four- and five-year olds in the city districts of Alna, Stovner, Grorud, Bjerke and Søndre Nordstrand. The evaluation report concluded that child care use was about 15 percent higher among children from immigrant families in city districts with free child care compared to city district without such offer when the children turned four and became eligible. Moreover, the report concluded that children with immigrant background in city districts with free child care performed better on tests in reading and mathematics in first and second grade (Bråten et al. 2014).
This first follow-up report explores how the children fare on assessment tests in third grade. We find that results for the third grade tests are in line what we found for first and second grade tests. Children from immigrant families in city district with free child care for four and five year-olds perform better on third grade tests compared to children with such background in city districts without an offer. For example, children with immigrant background in city districts that offer free child care are about four percentage points more likely to score above the limit for concerningly low performance in both literacy and numeracy on third grade tests. There is no such difference for children without immigrant background in these city districts.
When we compare children from immigrant families across backgrounds in city districts with and without free child care, we find that it is children from families with low income and low maternal labor market participation who perform better. We also find stronger results for girls than boys.
In the analyses we take into account a comprehensive set of observable differences between the children and their families. However, we cannot rule out that children in city districts offering free child care are different from children in comparison city districts in ways we cannot observe. To get closer to a causal approach we conduct a difference-in-difference analysis where we compare test results of children from immigrant families with results of children without immigrant background in city district with and without free child care. This method will remove the test score differences between children resulting from city district features that affect the groups of children similarly (e.g. school or child care quality). The third grade results do not change when we include the comparison group of children without immigrant background. This suggests that differences between city districts affecting children with and without immigrant background similarly do not drive the results.