Fertility and Marriage Supplement Sample Notes
The fertility supplement of the Current Population Survey asks women (either by self-response of proxy) questions about childbirth. Several fertility supplement samples also contain marital history information.
Marital history variables are available in fertility supplement samples from 1976-1995. Many samples contain only year at first marriage, but detailed marital history variables are available in 1980, 1985, 1990, and 1995. These marital history questions may include items such as the number of times a person had been married, the beginning and end dates of marriages, and whether marriages ended in widowhood or divorce. Males are in-universe for all of these variables in 1980 only in. In 1985, 1990, and 1995, male respondents are only asked about the year of first marriage and whether their first marriage ended in widowhood or divorce. In all other samples, only female respondents were asked about their first marriage.
In years prior to 1990, there is little consistency in the universe for fertility and marriage history variables. Users are encouraged to consult the comparability and universe tabs for details. In 1990 and 1995, the universe for fertility variables includes women ages 15 to 65. In the 1992, 1994, and 1998-2010 samples, the universe includes women ages 15 to 44. Items consistently collected over the years include the number of live births, the woman's age (in months) at the time of most recent birth, and the month and year of the most recent birth. In the years prior to 2001, survey items include a mix of other fertility related questions. In other years, the supplement may include items such as: the woman's age at time of birth, the child's sex, the month and year of birth for each birth, the current place of residence of each child (e.g., in the household, alone, with spouse, with grandparents, with adoptive parents, in foster care, etc.), and/or expectations about having children in the future.
In years prior to 1995, there are occasionally several records that have values for fertility or marital history variables that are not expected to be in-universe. For example, a record with "male" for SEX may have a value for "number of live births ever had." These values are in the original CPS data and remain in the IPUMS integrated variables.
The original CPS data for the 1979 fertility and marital history supplement is of questionable quality, as it contained non-meaningful strings and nonsensical values in several variables. When these values occurred, they were recoded to NIU. There are notes in individual variable descriptions to indicate when this applies and, where possible, is indicated in a data quality flag. This sample should nevertheless be used with caution.
FRSUPPWT should be used to tally the Fertility and Marriage supplement. This weight is identical to the basic survey sample weights.