Free center-based child care in Oslo

This report is the second report in a longitudinal study conducted by Statistics Norway for the Directorate for Education and Training. The point of departure is a policy that provided free child care for four- and five-year old children in some city districts of Oslo (Alna, Stovner, Grorud, Bjerke and Søndre Nordstrand) from 2006 and onwards. The policy aimed at increasing participation in child care, contribute to socialization and improved proficiency in the Norwegian language among children with immigrant background, strengthen the language stimulation in child care centers, and help improve contact between parents and centers.

This second follow-up report explores how the children fare on nationwide tests in language and mathematics in fifth grade. In the first part of the report, I look at how the children are doing on these test for the same sample that has been included in previous reports. I find that the results for fifth grade differ from the results of the former tests. Children with immigrant background in districts with free child care perform on average quite similar to children in districts without free child care on the nation-wide tests in fifth grade.

When comparing children from immigrant families from different backgrounds in city districts with and without free child care, I find some evidence that the children from families with low family income and low labor force participation among the mothers do somewhat better on the tests. Contrary to previous findings, boys seem to perform slightly better than girls, especially on the reading test.

In the analyses I include a battery of covariates to account for observable differences between the children and their families. However, it is impossible to rule out that there may be unobservable differences between children in city districts with and without free child care. Since data on test scores on the fifth grade are available for four- and five year-old children prior to the introduction of free child care, I study the results of a difference-in-difference model. This model compares test results between children in city districts with and without free child care, across cohorts before and after the policy was introduced. This better enables us to interpret results as effects of the policy.

The difference-in-difference estimates show a positive effect on good achievement on the reading test, as well as a negative effect on whether the student is granted exemption from taking the test. This seem to suggest that children in city districts with free child care perform better on the reading test in fifth grade after free child care became available. The results for math are somewhat smaller and more imprecise, but they point in the same direction. The positive estimates for reading suggest that the offer for free child care did in fact succeed in improving children’s language skills, in line with the objective of the intervention. Though children covered by free child care fare better than those not covered, the variation in the estimates across cohorts is sufficiently large (relative to the effect estimates) to make definite causal interpretations hard.

When I study effects for different groups of children based on their background, it appears that there is a stronger positive effect for boys than for girls on the reading test. For the group of children with a mother who does not work, I find a positive effect for both reading and math, in line with what we have previously found for the tests in 1st - 3rd grade. Like previous results, we see no effect for the children who have a mother who works. A similar pattern appears in the difference between children from families with income above / below the median income. There are clear indications of a positive effect for the children from low income families, in line with what we have previously seen in the previous reports.