Nationally Representative Data on Adopted Children in the United States
Adoption Advocate No. 168 - Did you know that the Census Bureau is a great source for U.S. adoption statistics? With surveys going back to 1990, the Census Bureau has a number of demographic data collections for adoption researchers, but it can be hard to know where to start and how to find exactly what you’re looking for. In this month’s issue of the Adoption Advocate, Rose Krieder, Chief of the Fertility and Family Statistics Branch of the Bureau, introduces the collected data and explains how users can find publications and data files that will help provide a better understanding of adoption.
Editor's Note: This article is released to inform interested parties of ongoing research and to encourage discussion of work in progress. The views expressed on statistical or methodological issues are those of the author and not necessarily those of the U.S. Census Bureau.
The Census Bureau has been collecting data since 1990 about the type of relationship between coresident parents and children, whether biological, adoptive, or step. This article seeks to introduce readers to the collected data, and to inform data users where to find publications and data files. Data about adopted children are collected in each of the major demographic survey operations fielded and processed by the Census Bureau: the American Community Survey (ACS), the Current Population Survey (CPS), and the Survey of Income and Program Participation (SIPP), in addition to the decennial census. While each of these programs produces nationally representative data, they differ in terms of sample size, geographic detail, and the specific characteristics of the child and household that are collected.
Data Details: What questions are asked that allow identification of adopted children in the Census Bureau’s main demographic surveys?
Decennial and ACS Survey
Identification of adopted children in the decennial census and the ACS is based on responses to the question that asks how each household member is related to the householder—someone who owns or rents the home. Responses allow us to see whether the child of the householder is reported as the biological child, adopted child, or stepchild.
However, since the relationship is always in reference to a central person—the householder—we cannot determine the type of relationship between the child and a second parent, if present, or the type of parent-child relationship (biological, adoptive, step) for children in the household who are not the householder’s child. So, for example, if the householder reports having a spouse, a biological child, and a grandchild, we do not know the type of relationship between the child of the householder and the spouse of the householder, or between the child of the householder and the grandchild of the householder, or indeed, whether the child of the householder is the parent of the grandchild of the householder.
The strength of decennial census data is the large underlying size since it is a count of every person living in the United States. This allows a look at lower geographic levels, although a limited set of characteristics of the child and the household are collected.These characteristics include the age, sex, race, and Hispanic origin of each household member, as well as whether the home is owned or rented.
Compared with the decennial census, the ACS collects more information about children and other individuals in the household on a wide range of social and economic characteristics including marital status, employment status, whether the person is foreign-born, year of entry to the U.S., whether household members speak a language other than English at home, how well they speak English, and their disability status. The ACS also includes a measure of poverty status and whether someone in the household receives public assistance.
Decennial Census Defined
The U.S. census counts each resident of the country, where they live on April 1, every ten years ending in zero. The Constitution mandates the enumeration to determine how to apportion the House of Representatives among the states.
CPS and SIPP
The CPS and SIPP allow a more detailed look at the number of parents in the household and the type of relationship between parents and their children. Beyond asking about the relationship of household members to the householder, both surveys provide direct identification of all parents living in the household and the type of relationship—biological, adoptive, or step—between each parent and child in the household, regardless of the ages of the parent and child.
Since 1991, the SIPP has included detailed measurement of the presence of parents and the type of relationship between children and parents. In 1996, the SIPP began asking respondents to identify a mother and father in the household (if coresident parents were present) as well as the type of relationship between parent and child, whether biological, adoptive, or step.
However, children in foster care were only reported on the question asking about each household member’s relationship to the householder. The direct identification of coresident parents in the SIPP allows data users to see whether children who live with two parents are living with a biological parent and a stepparent or two adoptive parents, for example. It also allows data users to identify grandparents in the household by using the variables that list the parents of a particular individual in the data set, and then identifying the parents of the parents of that individual. A series of reports focusing on the living arrangements of children have been published using this detailed SIPP data.
Finally, the CPS has measured children’s living arrangements to some extent since the late 1960s. In 1982 a direct question was added to identify whether each household member had any parent present in the household. If the household member was a child and their parent was married, then the child was shown in published tables as living with two parents.
ACS American Community Survey
CPS Current Population Survey
SIPP Survey of Income and Program Participation
Changes to the Questions
In 2007, several changes to the questions were made. Rather than asking for the identification of one parent in the household, the CPS asked for the identification of both a mother and a father, and the type of relationship between parent and child, whether biological, step or adoptive. In addition, a direct question asked whether adults living with nonrelatives had a boyfriend, girlfriend, or partner present. This meant that children living with two unmarried parents could be identified—like children whose coresident parents were married—as living with two parents.
Most recently, the questions were adjusted again in order to better reflect children living with two mothers or two fathers. The America’s Families and Living Arrangements table package, published annually based on the CPS Annual Social and Economic Supplement (ASEC) data and available on the Census Bureau’s website since about 2000, contains tables detailing the living arrangements of children.
Who is included in the ‘adopted child’ category in Census demographic surveys?
Data on the type of relationship between parents and children is based on respondent reports and consequently includes many variations of individuals they may view as their adopted children. The “adopted child” category may include various types of adoption such as adoption of biologically related or unrelated children, adoption of stepchildren, adoption through private and public agencies, domestic and international adoptions, and independent and informal adoptions. Informal adoptions are more common among some cultural groups, as people differ widely in the way they view family relationships and the process of adoption.
A qualitative study prepared for the U.S. Census Bureau found that informal adoption of biological grandchildren was common in Inupiaq communities in Alaska. Informal adoptions may also be more common among Hispanics and Blacks than other race and ethnic groups. In 2000 decennial census data, a substantial proportion of foreign-born Latin American children reported as adopted had not come into the United States on visas that would indicate that they had been legally adopted. Other studies have found that some parents who legally adopt children who are related to them may not report the child as adopted. In addition, we cannot distinguish among children who were adopted by their stepparents, children adopted by their biological grandparents or other relatives, and children adopted by people to whom they were not related by birth or marriage.
Do Census demographic surveys have data on internationally adopted children?
While there is no specific question asking if those reported as adopted children joined their families through international adoption, the ACS does ask whether each household member was born in the U.S., and if not, where they were born as well as the year they entered the U.S. This allows data users to identify those adopted children who were born outside the U.S.
This group includes children who entered the U.S. with a biological parent(s) but now live with an adoptive parent—who perhaps was their stepparent before the adoption, or who reports the child as their adopted child despite not having gone through a legal process of adoption. Since the place of birth is available for the householder parent and their spouse/partner, it is possible to restrict the sample to foreign-born children whose householder parent and spouse/partner are not foreign-born if the intention is to reflect foreign-born children who entered the U.S. specifically because of an international adoption.
With its large sample size, the ACS provides an opportunity to look at the group of internationally adopted children by characteristics such as disability or place of birth. The CPS does not collect data about nativity, and the much smaller sample size of the SIPP does not lend itself to analysis of this relatively small subpopulation.
Do Census demographic surveys have data on transracially adopted children?
All of the Census Bureau’s main demographic surveys collect the race and Hispanic origin of each member of the household. This allows the data user to compare the race of parents and children and determine which children they want to include as transracially adopted.
It should be noted that race is a socially constructed concept and researchers may differ in their definitions of “transracially adopted.” For example, in the most recently published Adopted Children and Stepchildren report, children of the householder are considered transracially adopted if the householder parent and their adopted child are of different race or origin groups, which are listed as:
- White alone non-Hispanic;
- Black alone non-Hispanic;
- Asian or Pacific Islander alone non-Hispanic;
- American Indian and Alaska Native alone non-Hispanic;
- Some other race alone non-Hispanic;
- White/Native Hawaiian and Pacific Islander;
- White/American Indian and Alaska Native;
- White/Some Other Race;
- OR either the parent or the child is another multiracial combination.
Note that only the ACS or decennial census is large enough to allow a look at these very detailed race and Hispanic origin categories. In addition, the ACS contains data about specific Hispanic origin groups, which would allow a closer look at those of Mexican, Cuban, Puerto Rican origin, for example.
What other groups of children can be identified in data from the Census demographic surveys?
Since Census Bureau surveys are household-based, we can only see co-resident stepparents and children. Many children have a stepparent they live with only some of the time. Since U.S. Census Bureau surveys sample addresses, and then determine who lives in the housing unit at that address, the data do not reflect relationships that cross household boundaries. As is the case for adopted children, estimates of stepchildren from decennial census and American Community Survey data do not capture all children who are living with a stepparent. This is because these surveys only ask for the child’s relationship to the householder. If the child has a second parent in the household, the identification of that second parent and the type of relationship between the second parent and the child will not be captured in the data.
Traditionally, a stepchild was the biological child of an individual’s spouse who was not also the individual’s biological child. However, data show that the usage of the term has shifted and adults who are not currently married, and in some cases have never been married, report living with stepchildren. Often, these adults have an unmarried partner in the household, who is presumably the child’s biological parent. Since the English language does not have a more precise word to describe this kind of relationship, some respondents may decide to report their partner’s child as their stepchild, even though they are not married to the child’s biological parent.
There is also some potential overlap between children who are adopted and stepchildren since stepchildren may be adopted by their stepparents. When reporting the type of relationship between a parent-child dyad in Census Bureau surveys, respondents can only select one parent-child relationship type: biological, adoptive, or step. So we do not know how many of the children who are reported as adopted were stepchildren of the parent, or conversely, whether parents might report the child as a stepchild although they have been legally adopted. The sociodemographic characteristics of stepchildren adopted by a stepparent have been found to be more similar to other stepchildren than to other adopted children.
As previously mentioned, the ACS and the decennial census only capture the relationship of household members to a central person—the householder—and so may not identify all grandparent-grandchild pairs in the household. To access published ACS tables, go to data.census.gov and type ‘grandchildren’ or ‘grandparents’ into the search box. Note that some of the same tables will come up regardless of which search term you enter. In contrast, since the CPS and SIPP ask for the direct identification of parents in the household, these indicators can be used to get a fuller picture of all coresident grandparents and grandchildren in the household.
The ACS is unique among the Census Bureau’s demographic surveys in that it also asks those adults ages 30 and over who report living with a grandchild under 18 whether they are responsible for most of the basic care of a grandchild, and if so how long they have been responsible for that grandchild. Because of the large sample, the ACS can be used to look at sub-national estimates in addition to national estimates.
The SIPP also has a unique question asking whether an adult is a grandparent, which allows us to compare the characteristics of grandparents who live with a grandchild to the characteristics of those who do not.
Children in Foster Care
In all of the Census Bureau’s main demographic surveys, children in foster care are reflected in the “relationship to the householder” item. That is, respondents can report that an individual in the household is a “foster child” of the householder. Notably, however, the response categories for the “relationship to householder” item include “grandchild,” so estimates of children in foster care based on the relationship to householder item may more closely reflect children in nonrelative care, since many grandparents participating in the formal foster care system may report the child as their grandchild rather than their foster child.
The Department of Health and Human Services manages the administrative records collected about children in foster care. These statistics can be accessed on their website at Adoption & Foster Care Statistics | The Administration for Children and Families. The most recent Adoption and Foster Care Analysis and Reporting System (AFCARS) report can be accessed on the Children’s Bureau website. Recently published research provides a look at historical trends in foster care in the United States.
Where can I access the data files?
- Decennial census page—Decennial Census of Population and Housing by Decades
- ACS page on how to access the data (includes tables): American Community Survey Data (census.gov)
- CPS page on how to access the data: Data (census.gov)
- SIPP page on how to access the data: Survey of Income and Program Participation Data (census.gov)
How do I choose which data set to use?
- The strength of the decennial census data is its ability to provide estimates for small geographic areas. Adopted children of the householder can be identified beginning in 2000. A limitation of the decennial census data is that information on parent-child relationships—including whether the child is a biological child, adoptive child, or stepchild—is only available for children of the householder.
- The strength of the ACS is its ability to provide estimates for smaller geographic areas, due to its very large sample size, as well as the many characteristics collected about household members. Like the decennial census, a limitation of the ACS data is that information on parent-child relationships is only available for children of the householder.
- The strength of the CPS ASEC is the long-time series since it has been collected for decades. In addition, more detailed information on parent-child relationships is available for those who are not the children of the householder.
- The strength of the SIPP is the wealth of characteristics collected about the child and everyone in their household. However, the sample size does not make it the best choice for exploring the characteristics of a relatively small population such as adopted children.
This article provides an overview of the type of data available about adopted children and other groups of children in the main Census Bureau demographic surveys. These surveys’ large sample size or amount of detail provide an opportunity for a more detailed exploration of the living situation and characteristics of the parents and households of adopted children, stepchildren, grandchildren, and children in foster care. It is important to be mindful of the strengths of each data set when setting up a research project.
 Another source of statistics on adopted children is the Child Welfare Information Gateway published by the Department of Health and Human Services, available at: https://www.childwelfare.gov/topics/systemwide/statistics/adoption/
 Census 2000 was the last decennial census to include a ‘long form’ where additional characteristics of people and households were collected, just as they are now in the American Community Survey.
 Access the published Living Arrangements of Children reports here:
1991: The Diverse Living Arrangements of Children: Summer 1991 (census.gov) https://www2.census.gov/library/publications/1994/demographics/p70-38.pdf
1996: Living Arrangements of Children: 1996 (census.gov) https://www2.census.gov/library/publications/2005/demo/p70-104.pdf
2001: Living Arrangements of Children: 2001 (census.gov) https://www2.census.gov/library/publications/2005/demo/p70-104.pdf
2004: Living Arrangements of Children: 2004 (census.gov) https://www2.census.gov/library/publications/2008/demo/p70-114.pdf
2009: Living Arrangements of Children: 2009 (census.gov) https://www2.census.gov/library/publications/2011/demo/p70-126.pdf
2019: Living Arrangements of Children: 2019 (census.gov) https://www.census.gov/wp-content/dam/Census/library/publications/2022/demo/p70-174.pdf
 The SIPP also made this adjustment beginning with the 2014 panel.
 For example, see the most recent table package listed on this site: Families and Households (census.gov) https://www.census.gov/topics/families/families-and-households.html as CPS: 2021, or access them through this website, looking under the tab for the appropriate year and with the title America’s Families and Living Arrangements Families and Households Data Tables (census.gov) https://www.census.gov/topics/families/families-and-households/data/tables.html.
 Amy Craver. “Complex Inupiaq Eskimo Households and Relationships in Two Northwest Alaska Rural Communities,” Alaska Native Science Commission. University of Alaska, Anchorage, 2001.
 Maria Suarez Hamm. “Latino Adoption Issues,” Adoption Factbook III. National Council for Adoption. Washington, DC. 1999, pp. 257–260.
 Priscilla A. Gibson, Justine Nelson-Christinedaughter, Harold D. Grotevant, and Hee-Kyung Kwon. “The Well-Being of African American Adolescents Within Formal and Informal Adoption Arrangements,” Adoption Quarterly, 2005, Vol 9:1.
 Rose M. Kreider. “Foreign-born adopted children in the U.S., 2000,” In: Thomas Atwood, Lee Allen, and Virginia Ravenel, eds. Adoption Factbook IV. Washington, DC: National Council for Adoption; 2007, pp. 133–153.
 Laura F. Radel, Matthew D. Bramlett, and Annette Waters. “Legal and Informal Adoption by Relatives in the U.S.: Comparative Characteristics and Well-Being from a Nationally Representative Sample,” Adoption Quarterly, 2010, Vol. 13: 268–291.
 This Research Brief published by the Assistant Secretary of Planning and Evaluation indicates that about 23 percent of children adopted from the foster care system were related to their adoptive parent(s). See Table 1 on page 4.
 The same definition of transracially adopted was used for this more recent article: Kreider, Rose M., 2020. “US Adoption by the Numbers,” in ed., Gretchen Miller Wrobel, Emily Helder, and Elisha Marr, The Routledge Handbook of Adoption, Routledge: New York.
 Rose M. Kreider. “Adopted Children and Stepchildren: 2000,” Census 2000 Special Reports, CENSR6-RV, 2003, U.S. Census Bureau, Washington, DC, available online at Adopted Children and Stepchildren: 2000 (census.gov) https://www.census.gov/wp-content/dam/Census/library/publications/2003/dec/censr-6rv.pdf
 Matthew D. Bramlett. “When Stepparents Adopt: Demographic, Health and Health Care Characteristics of Adopted Children, Stepchildren, and Adopted Stepchildren,” Adoption Quarterly, 2010, Vol. 13, pp. 248–267.