Count All Kids Committee Comments on American Community Survey (ACS) Data

On Monday, July 11, The Partnership for America’s Children responded to a Census Bureau survey inquiring about activities and products that would support those who work with American Community Survey (ACS) data.

The comments focused on two main aspects: keeping children and parents linked in ACS data and clarifying a child’s relationship with adults in the household that are not the householder.

Any Disclosure Avoidance System incorporated into ACS data must keep children linked with parents

It is important to keep children and parents linked in the ACS data when the new Disclosure Avoidance System (DAS) is developed to ensure child poverty data and other measures of child well-being are accurate, especially on the state and local level.

The use of the Differential Privacy method in the 2020 Census redistricting file and Demographic and Housing Characteristics file separated children from their parents, which resulted in 163,000 census blocks showing populations under age 18 but not populations identified as 18 and over. It is improbable that this many blocks had populations without adults in the same vicinity as children. If the link between adults and children is broken by the formal privacy process in the ACS data, it will have severe consequences.

• Although the Current Population Survey is the official measure of poverty at the national level, ACS data is commonly used for state and substate child poverty measures.
• Child poverty data looks at the whole household income, (since most children do not have individual income), which requires the data on children and adults to be linked.
• Any change in how child poverty is measured could result in a discontinuity in the measurements of child poverty. This would come at a pivotal time when child poverty levels are changing rapidly due to many societal and policy shifts.
• If changes to the ACS child poverty measurement leads to less accurate data, federal funding that is allocated based on child poverty data may not get to children who need it most.

Clarify each child’s relationship with adults in the household; household characteristics should be accurately captured

The comment also recommended that the Bureau consider adding a question to the ACS survey asking each child’s relationship to the non-householder adults, or whether the child’s other parent lives in the household. This distinction has become important as the number of children living in households with parents who are not married or living with grandparents has increased drastically, which requires a more nuanced description of a child’s family structure. Until the Bureau can improve this data collection, the comment asks that it clarify its documentation to make clear what the data shows, and flags the issue in the data tables that the Bureau produces.

• Policy makers and service organizations make critical decisions based on how many children live with a parent, grandparent, or other relative.
• The current question asks about the relationship between the child and householder but does not ask about the child’s relationship to the other adults in the household. This can lead to misleading data about the nature of childrens’ relationships with the adults in their household. We do not have good data on how many children are living with both parents, only one parent, or no parent. (See the data documentation explanation here in the section explaining relationship to the householder, starting at page 80.)
  • • • • • Specifically:
            • • In children living with a married householder that is their parent, the child could be the child of the householder and their spouse, or the child of the householder and the stepchild of the spouse
              • In children living with a householder that is their parent and their unmarried partner, the child may be the own child of the householder and the own child of the partner but all children will be reported only as the own child of the householder and it will not be clear whether they are living with both parents. A child listed as not related to the householder will be listed as not living with a relative when they may be the own child of the householder’s partner. Note that only children of married couples can be identified as living with both parents according to the documentation explanation.
              • A child listed as the grandchild (or other relative) of the householder where there are other adults in the household may be also living with their parent or parents—for example, the household might be the grandmother, her daughter and daughter’s partner, and their child—or the child’s parent may not be in the household even if there are also children of the householder in the household—those adults might be the aunt or uncle of the child rather than their parent.
              • As a result, no one table reports on all children living with both parents, no table reports on how many children are living with just one parent, and no table reports on how many children are living with parents and other relatives compared to children being raised by non-parents, with no parent in the household.