[Update: It's great to see my opinion confirmed, with much more understanding, context and background, by an eminent statistician and demographer, Justin Wolfers, in the New York Times! "Census Bureau’s Plan to Cut Marriage and Divorce Questions Has Academics Up in Arms" (12/31/14)]
Practicing what I preach, I have sent in my comments on the Census Bureau's proposal to stop collecting information on marriage and divorce. The deadline is tomorrow, Tuesday, Dec. 30. (In case you missed what this is about, see my original posting, "Census will stop studying marriage, divorce; Dec. 30 public comment deadline").
December 29, 2014
Departmental Paperwork Clearance Officer,
Department of Commerce,
1401 Constitution Ave., NW, Room 6616,
Washington, D.C. 20230
Re: Comment on Proposed Changes to American Community Survey
Dear Ms. Jessup:
Your notice in the Federal Register classified the ACS questions on marriage and divorce as low-cost, but “low value”. On the contrary, they are of very high value to our society, to the public, the press, students, researchers, legislators and parts of the executive branch.
Social and public usefulness of these statistics
I have been a family-law attorney for 19 years and for almost as long I have operated a divorce statistics web site. In its current form it is called the Divorce Statistics and Studies Blog, at http://familylaw.typepad.com/stats. From doing that work, I have been able to observe that the public, the media, students, academic researchers, and state and federal policymakers all have a widespread, lasting interest in marriage and divorce, and in current statistics on them of a kind which only the American Community Survey provides. Anyone writing an article, research paper, or presentation on marriage or divorce wants to use statistics in it. Any legislator proposing a bill related to either topic wants accurate, current statistics to frame his or her argument. Although I put a great many statistics and studies on the blog, I frequently get calls, e-mails and letters asking for more statistics.
The public rightly considers marriage and divorce to be extremely important subjects, and people overwhelmingly assume that the government keeps detailed, up-to-the-minute statistics on them, far more than is the case even now. People have been hearing about divorce statistics for many decades now, and most people I deal with simply could not comprehend that the government was no longer keeping them. If I told them that, they would keep looking, use outdated numbers, or just make something up that they thought was what they heard somewhere.
The ACS-based statistics are also very useful to parts of the federal government. Two examples:
- The Department of Defense and the various branches of the military keep very detailed statistics on divorce and marital status, but those would be much less useful without your corresponding numbers on the general population for comparison purposes. The military has realized in recent years that it needs to take much better care of families, and statistics are essential for measuring progress.
- The Department of Health and Human Services provides many services to families that have experienced divorce or unwed childbearing, including many services related to child support, and specifically funds fatherhood, marriageability, marriage skills-building, and divorce-prevention programs as part of the Temporary Assistance to Needy Families (TANF) program. The department and its grant recipients need statistics in order to understand the populations they work with and the need for particular programs.
Why ACS statistics are crucial
Most private and academic studies of marriage and divorce are based on information from the ACS and its predecessors. The Census Bureau has also produced very important studies of marriage and divorce based on the ACS and its predecessors, including generational-cohort studies and probability projections. These are where we get the kind of numbers that actually answer the questions people wonder about: What are our overall chances of divorcing, ever? How long do marriages last? What about second and third marriages? How many children grow up without one parent? What is the divorce rate for people of my age, educational level, etc.? And what can we do about it?
The other source of marriage and divorce statistics is the National Center for Health Statistics (NCHS), which collects annual totals of marriages and divorces reported from county courts to state vital statistics offices. But this information, by itself, is not very useful to governments or to the private sector. It yields such measures as “3.6 divorces per 1000 population per year”. I.e., 0.72% of the population gets divorced each year. That does not tell people anything that sounds relevant to their lives. Also, NCHS divorce numbers are much more useful when compared to a count of the married population, which comes from the Census. And they are incomplete: California and several other states are not included.
This proposed change was not widely publicized. A very broad cross-section of the general public, media, academia and policymakers are affected by it, and so the eventual disappointment from the elimination of these survey subjects will be much more widespread than the comments you receive will indicate.