As many family scholars know, the official divorce rate in the United States reached a peak in the early 1980s and gradually declined in subsequent decades. But how well does the official divorce rate reflect the actual frequency of marital dissolution? Not as well as you might think, according to recent research.

For more than a century, the federal government has calculated the annual rate of divorce from records of divorces compiled by states. In some decades, the quality of data was relatively good. Between 1960 and 1990, for example, the federal government monitored the completeness and accuracy of divorce records and compensated states for their efforts. But the quality of data declined in the 1990s, and the federal government stopped funding the collection of detailed marriage and divorce data in 1996. Since then, most (but not all) states have continued to send raw counts of divorce to the federal government, although these data are limited and of inconsistent quality.

In a recent article, Kennedy and Ruggles (2014) argued that much of the apparent decline in marital dissolution since 1980 reflects an under-counting of divorce rather than a change in the underlying rate. They base this conclusion on an analysis of data from the American Community Survey (ACS), which is conducted annually by the U.S. Census Bureau. In 2008, questions on marital transitions (marriage, divorce, widowhood, and remarriage) were added to the ACS—questions designed to fill the gaps in the vital statistics system. As it turns out, estimates of divorce based on the ACS are higher than the corresponding figures from the Centers for Disease Control and Prevention's National Vital Statistics system. And although the divorce rate based on vital statistics declined by 21% between 1980 and 2010, the comparable figure from the ACS is only 2%.

Kennedy and Ruggles also calculated age-standardized divorce rates, which take into account the changing age composition of the population. Their calculations indicate that the age-standardized rate of divorce has increased (not declined) by 40% since 1980. If this seems confusing, keep in mind that the U.S. population has aged in the last few decades, and older couples are at less risk of divorce than younger couples. The fact that divorce has remained more or less constant while the population has grown older suggests that the underlying propensity to divorce has increased. Much of this increased propensity appears to be concentrated among older married couples.

So has the divorce rate, declined, stayed about the same, or increased in recent decades? The answer to this question depends on which source of data you choose (Vital Statistics or the ACS) and whether you adjust for population aging. If you feel that this answer is frustrating, you are not alone.

The fact that a statistic as basic as the divorce rate is in question should be a source of concern to family scholars, educators, and practitioners. Given the importance of this topic, you might think that the federal government is committed to maintaining—and even enhancing—the quality of marriage and divorce data. But in a startling development last year, the U.S. Census Bureau (under pressure from Congress to cut costs) announced its intention to drop the marital transition questions from the ACS. The ACS is currently the best source of national and state data on the frequency of marriage, divorce, widowhood, and remarriage. And the ACS is the only data source that can measure marital trends in relatively small groups in the population, such as Asian Americans, Hispanic subgroups (e.g., Cubans and Puerto Ricans), and people in same-sex marriages. Losing this resource will have serious consequences for the quality of family research.

After discussions with the NCFR Board of Directors, Executive Director Diane Cushman and I sent letters to the U.S. Census Bureau and the U.S. Department of Commerce (which houses the Census Bureau) in December of 2014, expressing our concern and arguing for the reinstatement of the ACS questions. We also alerted the NCFR membership to this issue, and many of you sent similar messages. Other social science organizations, such as the Population Association of America and the Council on Contemporary Families, also expressed their concerns to the federal authorities. The strong, critical response from the social science community received a good deal of media attention, and several news stories mentioned NCFR in particular. Taking a stand and expressing our voice on this issue was good for NCFR; it demonstrated that we are aware, engaged, and relevant.

At the time of this writing, it is not clear whether the federal government will go forward with its plan to delete the marriage questions from the ACS. Irrespective of the outcome, this decision can be viewed as evidence of a retreat on the part of the federal government from supporting research on couple relationships. About 15 years ago, for example, the National Institute of Mental Health discontinued funding for research on interventions for improving couple outcomes. More recently, the Eunice Kennedy Shriver National Institute of Child Health and Human Development discontinued funding for research on couple education and interventions related to relationship quality, marriage, and divorce. As a result of these and other decisions, scholars are finding it increasingly difficult to obtain federal funding for research on couples. The situation may get worse in the aftermath of the 2014 election, with some leaders of the new Congressional majority vowing to eliminate all federal funding for social science research.

Divorce continues to be a common feature of our marital system—perhaps more common than we thought. In addition, over 40% of children are born outside of marriage these days, often to partners who choose to cohabit rather than marry. These relationships tend to be fragile, however, and most break up within a few years. Neither the ACS nor the Vital Statistics system count the ending of cohabiting relationships, even those with children. Consequently, both data sources substantially underestimate the overall level of union and family instability in our country. We should be concerned about this picture. Relationship distress and instability increase the risk of emotional, behavioral, and health problems for adults and children and contribute to many of our most serious social problems.

Many NCFR members publish research on or conduct interventions with couples. In an era of decreasing funding for research, how will this work continue? Making do with less is one way to deal with conditions of scarcity. But we also need to do a better job of demonstrating the usefulness of studying couples and designing interventions that reduce domestic violence and strengthen relationship quality and stability. One of NCFR's global goals is to "raise the visibility of family research, theory, and practice to policymakers and the general public (global end policy #2 C)." Being more vocal about what we do, and working to create a more supportive environment for family research and practice, should be high priorities for the foreseeable future.

Copyright © 2015 National Council on Family Relations (NCFR). Contact NCFR for permission to reprint, reproduce, disseminate, or distribute by any means.