ESTIMATES OF OUT-OF-WEDLOCK CHILDBEARING Data on out-of-wedlock births from the June Current Population Survey (CPS) are approximated by the number of women who had a birth in the year prior to the survey date (approximately July, 1993 to mid-June, 1994) who were currently unmarried at the time of the survey (widowed, divorced, and never married). These estimates usually are less than the corresponding estimates published by the National Center for Health Statistics (NCHS) which are based on counts of nonmarital births obtained from individual birth certificates. For example, data published in the June 1994 CPS report indicated that 26 percent of births to women in the preceding year were born out-of-wedlock birth compared to 31 percent for calendar year 1993 published by NCHS in their latest Advance Report of Final Natality Statistics, 1993. Preliminary Research Findings A detailed report comparing estimates of out-of-wedlock childbearing between these two data collection systems is currently being prepared by staff from the Census Bureau and NCHS. One change has already been instituted to reduce the difference between CPS and Vital Statistic data. This involves including births to those women in the CPS who had a child while they were never married but who subsequently married by the survey date. This procedure increases the CPS percentages by about 2 percentage points, but still leaves them significantly less than the Vital Statistics figures, even after taking into account the margin of error due to sampling variances (see attached table). More importantly, other findings now suggest that a major reason for the overall differences between these two estimates can be traced to the foreign-born component of the population. Estimates of out-of-wedlock childbearing for native born non-Hispanic women from the CPS are virtually identical to those from NCHS (about 29 percent, see table). Likewise, estimates for native-born Hispanics are very close (about 41 percent each). However, out-of-wedlock childbearing for foreign-born non-Hispanic women from the CPS is significantly less (10 percent) than the NCHS estimate (18 percent); a similar pattern exists for foreign born Hispanic women (21 and 39 percent, respectively). The CPS counts a birth as occurring in the last year as a marital birth if the woman is reported as being currently married (including separated) at the time of the June interview. If it is determined that this marriage occurred subsequent to the child's birth (i.e., the woman was unmarried at childbirth), this woman would then be excluded from the marital birth category and would be classified as having had a nonmarital birth. The usual method of determining a marital from a nonmarital birth for NCHS data is to tabulate directly the mother's marital status category on the birth certificate. This information is usually provided by the mother at the hospital. This information, however, was not obtained for births occurring in 6 States in 1993 (California, Connecticut, Michigan, Nevada, New York, and Texas). About 35 percent of all births occurring in the U.S. were to mothers from these States but 70 percent of the births to Hispanic mothers were residents of these States (the majority of these Hispanic mothers--63 percent--were born outside the U.S.). In these States, marital status is inferred (i.e., administratively imputed) from a comparison of the child's and parents' surnames. Any conceptual differences between the CPS and NCHS systems in determining the marital status of the parents, obviously, must be key factors in accounting for these statistical discrepancies. In the CPS, both "consensual or common-law" marriages and "legal" marriages are recorded in the same category, "married." For births occurring in these 6 States, it is possible that a birth to a "consensually married couple" with different spousal surnames may be recorded as a "marital" birth in the CPS and as a "nonmarital" birth in the Vital Statistics system. This problem could significantly affect out-of-wedlock childbearing statistics, although NCHS has developed rigorous coding rules for these States. Because the majority of births to Hispanic mothers in the U.S. occur in States where nonmarital births are inferred from surname comparisons, Hispanic out-of-wedlock statistics may be especially prone to coding problems due to the complexity of Hispanic surnames. It is possible that consensual marriages, which are quite typical in Latin America, may be of sufficient proportions among the immigrant population in the U.S. to cause the CPS to "over-count" marital births or "under-count" nonmarital births relative to NCHS. Data from the 1990 Census of Puerto Rico indicate that 11 percent of currently married women 15 to 44 years old in Puerto Rico reported themselves as being consensually married. Perhaps more importantly, the complexity of Hispanic surnames--as with surnames of persons from other countries which do not follow traditional Anglo-Saxon patterns of naming conventions-- could cause additional problems in the determination of a nonmarital birth. The shortfall in the CPS estimates also noted for foreign-born non-Hispanic women points to the likelihood that language or communication problems between the mother and the recorder of data, whether on a birth certificate or Census survey, may be key factors in the collection of consistently defined events between the two data systems. Further detail concerning these issues may be obtained by contacting Martin O'Connell, Bureau of the Census, by email moconnel@census.gov or by telephone (301) 457-2416. ESTIMATES OF THE PERCENT OF BIRTHS OCCURRING OUT-OF-WEDLOCK __________________________________________________________________ Race and 1994 CPS 1/ 1993 NCHS 2/ Hispanic origin Percent CI 3/ Percent __________________________________________________________________ Total 27.7 1.7 31.0 Not Hispanic 4/ 27.3 1.9 29.3 5/ Native 28.5 1.9 30.3 Foreign-born 9.8 4.9 18.1 Hispanic 29.6 5.8 40.0 5/ Native 40.8 9.5 41.6 Foreign-born 21.4 6.8 38.9 Asian and Pacific Islander 6/ 15.7 8.6 15.7 _________________________________________________________________ Notes: 1. CPS data in this table differ from statistics in the published report by the inclusion of births that occurred to women never married at childbirth but who subsequently married by the survey interview date. This procedure adds about 2 percent to the overall proportion of children born out of wedlock. 2. Refers to the percent of births for calendar year 1993 that were classified as nonmarital births. Data provided by Stephanie Ventura, Division of Vital Statistics, National Center for Health Statistics. 3. Represents the 90 percent confidence interval of the estimated percentage. 4. Includes other races not shown. 5. Includes births with place of mother's birth not stated. 6. May be of any ethnic origin.