Inner TRIM3 Masthead

Identification of Married Custodial Mothers

Details concerning the identification of married custodial mothers are presented below.

Custodial Mother on Basis of Reported Child Support

A married mother who reports receiving child support on the CPS is automatically assumed to be a custodial mother.

Custodial Mother on Basis of Child's Reported Child Support

If the mother does not report child support, but has a child who appears to be reporting child support paid on his or her own behalf, then TRIM3 must decide whether the mother or father is the custodial parent. If the father reports receiving child support or reports receiving health insurance coverage from outside the household then TRIM3 assumes that the father, rather than the mother, is the custodial parent. TRIM3 assumes that in 92.5 percent of the remaining cases, the mother is a custodial parent. This is based on tabulations of the 1998 CPS-CSS that show that in 92.5 percent of the cases in which one or both parents in a married couple receive child support, the mother receives child support.

Custodial Mother on Basis of Child's Health Insurance from Outside the Household

If the mother has a child covered by the health insurance plan of someone outside the household (and this is an actual response, not a missing response assigned health insurance coverage through the Census Bureau's allocation procedure), then the mother will be counted as a custodial mother (subject to the following restrictions).

TRIM3 does not consider health insurance coverage from outside the household to be an indicator of custodial parent or child status if the parent is married with spouse absent, but not separated. In these cases, it is likely that the child is covered by the insurance of the temporarily absent parent (for example, a parent who is away in the armed forces) rather than by a parent from a prior relationship.

Also, if the mother's husband reports receipt of child support or reports that he himself is covered by the health insurance plan of someone outside the household, then it is likely that the father, rather than the mother, is the custodial parent (or that the family is reporting some other type of coverage not related to child support or custody arrangements). In these cases, health insurance coverage from outside the household is ignored in determining custodial mother and child status. However, if there is more than one child, it is possible that the mother will be found to be a custodial parent through other means.

Custodial Mother on Basis of Imputation

Logit models are used to impute the custodial parent status of married mothers who are not identified as custodial mothers on the basis of reported child support or children with health insurance coverage from outside the household. The logit models were estimated on married mothers who did not report receiving child support on the 1998 CPS-CSS.

If there is only one child and all indications suggest that it is the father, rather than the mother, who is the custodial parent, then the mother will not be assigned to be a custodial mother. The father is considered to be the custodial parent if the mother does not report receiving child support, and the father reports receiving child support or the father and child (but not the mother) are covered by the health insurance plan of someone outside the household.

Three separate logit models are used to impute whether a married mother is a custodial mother -- one for married mothers with one child, one for married mothers with two children, and one for married mothers with three or more children.

If there is only one child, then the key explanatory variables are a set of dummy variables that cross the mother's age with the age at which she had her child. The model's design captures the impact of the child's age on the likelihood that the child has a father living elsewhere, and also the impact of the mother's age at the time of the child's birth. The child support module program rule MarriedMomAdjustment1 can be used to specify multiplicative adjustments (by race/ethnicity) to the imputed probabilities.

If there are two children, then the age gap between the two children is included as an additional explanatory variable. The larger the age gap between the children, the more likely that the older child is a child from a previous relationship and has a father living elsewhere. The child support module program rule MarriedMomAdjustment2 can be used to specify multiplicative adjustments (by race/ethnicity) to the imputed probabilities.

If there are three or more children, then the key explanatory variables are the maximum age difference between any two consecutive children, and dummy variables for the mother's age. The data used in estimating the model did not include enough observations to support the set of dummy variables crossing the mother's age with the age at which she had her child, so only dummy variables for the mother's age are used. The child support module program rule MarriedMomAdjustment3Plus can be used to specify multiplicative adjustments (by race/ethnicity) to the imputed probabilities.