This descriptive, retrospective longitudinal study used California child welfare administrative data to follow a cohort of 5,873 foster children born in 1999 who first entered care at less than one year of age. This study used a birth cohort rather than the usual entry cohort and followed the children through multiple placement episodes rather than a single episode. The cohort was followed to age 9. At age 9, 51% of these children had been adopted, 36% were reunified with parents, 7% were with guardians and 5% were in care. These are higher adoption and in care rates and lower reunification rates than found when only considering the first placement episode.

Three broad themes emerge from the experiences of these children. The first is that not all outputs defined as permanency provide the same level of stability. Although reunification has the crucial advantage of maintaining a child with his or her family, for some children it is inherently less stable than adoption or guardianship. The second theme is that the vast majority of the cohort children have achieved permanency, most often in the form of adoption but also in the forms of reunification and guardianship. This is consistent with the permanency focused child welfare policy changes of the past thirty plus years. These changes, which are, in turn, consistent with attachment theory, have striven to replace open-ended foster care with more secure relationships. The third theme is that first placement episode data alone provide an inadequate, distorted description of the experience of children in the foster care system. First episode data alone overstate the frequency of stable reunification and understate the number of children who are adopted and the number who are in care. A sub text for this theme is that longitudinal data have the power to allow an understanding of these differences.