Historical statistics on domestic adoptions during the twentieth century are interesting, but they are scarce and can also be misleading. Field studies did not even begin to estimate numbers of adoptions, or document who was being adopted by whom, until almost 1920. When researchers began to tally adoptions, they did so in only a handful of Northeastern and Midwestern states and based conclusions about statewide patterns on records from a few counties, usually in urban areas.

A national reporting system for adoption existed only between 1945 and 1975, when the U.S. Children’s Bureau and the National Center for Social Statistics collected data voluntarily supplied by states and territories. Today, most statistics available about adoption are being gathered by private organizations, such as universities and foundations. The Adoption and Safe Families Act of 1997 requires states to collect information about the adoptions of children in public foster care, but these are the only adoption-related statistics regularly reported by governments.

Even when the federal government was trying to keep track, during the three decades after World War II, adoption statistics were incomplete. They never included informal adoptions, which were beyond the reach of law and uncountable by definition. The summary data that did exist tended to obscure trends that were as important as total figures. How many children were adopted by relatives and how many by strangers? How many were arranged independently or by agencies? How many involved infants or adolescents? What factors explain regional and state differences in the past and present? Why, for example, are adoption rates in Wyoming and Alaska higher today than in California, Delaware, and Texas? Have any or all of these patterns changed over time? We can guess, but usually on the basis of partial or non-existent numbers.

We know one thing with certainty on the basis of historical statistics. Adoptions were rare, even at the height of their popularity, around 1970. What is paradoxical is that adoptions have become rarer during the past several decades, just they have become more visible. A total of approximately 125,000 children have been adopted annually in the United States in recent years, a sharp drop since the century-long high point of 175,000 adoptions in 1970. Growing numbers of recent adoptions have been transracial and international—producing families in which parents and children look nothing alike—and the attention attracted by these adoptive families has led many Americans to believe that adoption was increasing. The adoption rate has actually been declining since 1970, along with the total number of adoptions.

Estimates suggest that adoptive families are atypical as well as few in number. Approximately 5 million Americans alive today are adoptees, 2-4 percent of all families have adopted, and 2.5 percent of all children under 18 are adopted. Adoptive families are more racially diverse, better educated, and more affluent than families in general. We know this because Census 2000 included “adopted son/daughter” as a kinship category for the first time in U.S. history. It is possible that the demographic profile of adoptions arranged many decades ago was just as distinctive. We simply do not know.

Special-purpose adoption laws have existed in the United States since the middle of the nineteenth century. More than a century ago, however, very few Americans entered courts in order to formalize kin ties. Divorce, still very unusual at the turn of the twentieth century, was more common than adoption. After 1900, numbers of adoptions in the United States began to climb. Why? First, a new culture of children’s innocence and vulnerability placed a premium on their welfare and secure membership in families. Second, tangible benefits, such as those available through the social security system established during the 1930s, offered practical incentives for Americans to legalize family bonds. For the period before 1945, however, we have practically no detailed national statistics. After 1945, the number of total adoptions increased steadily, with numbers of adoptions doubling in the decade after World War II to reach approximately 100,000 annually by the mid-1950s. During this period, the proportion of non-relative adoptions arranged by agencies also increased significantly, a partial victory for child welfare professionals who had been advocating expansive regulation, uniformity, and minimum standards for decades. Before 1945, independent placements probably represented more than half of all adoptions. These decreased to an all-time low of 21 percent in 1970.

The statistical picture for international adoptions is uniquely clear because the federal government counts all legal immigrants, including immigrant “orphans,” as they are still called. (We also know that approximately 500 American children are adopted annually by foreigners, mostly in Canada and Europe, but in comparison to this country’s status as a “receiving country,” we know practically nothing about the United States as a “sending country.”) We know with some precision how many children born in South Korea have been adopted by U.S. citizens during the past fifty years—well over 100,000—and figures available through the Department of State tell us the number of Vietnamese, Guatemalan, Romanian, Chinese, and children of other nationalities who have been incorporated into American families through adoption. In the past decade, international adoptions have increased dramatically as a component of the adoption total: the 2002 figure of 20,009 was more than triple the 1992 figure, and comprised approximately 16 percent of all adoptions.

In addition to knowing where international adoptees come from and how many of them there are, we also know that well over 60 percent are girls and virtually all have been non-relatives. That does not mean that non-relative adoptions are on the rise, however. Because divorce and remarriage have become more common, relative adoptions (by step-parents, for example) have become much more prevalent among domestic adoptions in recent decades.

Numerically significant adoptions are not necessarily socially sensitive adoptions. Relative adoptions have become more common in recent decades but have attracted relatively little notice. Exactly the opposite is true for transracial adoptions. These have been covered extensively in the press and studied intensively by researchers, but their importance is symbolic rather than statistical. The largest number of transracial adoptions occurred in the years around 1970, when there were perhaps a few thousand annually. Opportunity, an Oregon program, conducted one of the only national surveys of black adopted children; it documented 7,420 total adoptions in 1971, of which 2,574 were transracial. This was a tiny number, considering that almost 170,000 adoptions were finalized in the country that year. Why did outcome studies focus on a small number of African-American children adopted by white parents but ignore the thousands of children adopted by relatives? The former was controversial and the latter was not.

Since all kinds of adoptions were and still are rare, the reason to subject them to quantitative inquiry has had little to do with sheer numbers. Governments and private organizations have compiled adoption statistics because numbers have been crucial in adoption policy debates. Proof that adoptions arranged in the black market turned out poorly was valuable ammunition in the campaign against disreputable independent adoptions, for instance, while proof of how professionally arranged adoptions turned out could make or break the reputation of agencies. Numbers were also accorded great meaning within the placement process. The I.Q. scores of children, the ages of aspiring parents, and the educational levels of birth parents were all, at one time or another, treated as key indicators of where and with whom they belonged.

Social researchers who conducted pioneering studies of child placement, such as Sophie van Senden Theis, author of How Foster Children Turn Out, believed that counting was a privileged method of accumulating knowledge and approaching truth scientifically. They were sometimes surprised or disturbed by what statistics and correlations revealed—that many adopters failed to inform their children about their adoptions or that “telling” was not a reliable predictor of positive outcomes—but they were always confident that compiling aggregate data would improve the lives of individual children. Statistical evidence based on many adoptions was often compared with anecdotal evidence, which revealed the details of one child’s or family’s story. Numbers were often considered more objective than narratives, and therefore more legitimate and trustworthy as a basis for policy and practice.

That adoption statistics have been gathered so haphazardly suggests that the effort to tie adoption reform to adoption knowledge has been a partial success, at best. But they also embody a uniquely modern faith in numbers and a widespread belief that they could be trusted to plan and govern the future.