How do you link children to unmarried partners?

It appears as if the biological/adopted children of unmarried partners have, for the most part, been unlinked in the ACS data. This means that momloc/poploc wouldn’t work to establish relationships. For example, in a household with unmarried partners, three kids could be related to the household head – one listed as a child and the other two as stepchildren, yet the stepchildren are not linked in any way to the unmarried partner. It also isn’t possible to determine with momloc/poploc if the child is the biological child of the unmarried partner.

Any thoughts on how to get around this? In the analysis I’m working on, I need to group families with unmarried partners together if they have a child in common, but I need to separate the partners (and their kids) if they do not have a biological/adopted child in common.

With the CPS data I can get around this by using the pedadtyp and pemomtyp variables, but I would prefer to use ACS in order to look at sub-state geographies. Anyway to identify these relationships in the ACS data?

As you point out, using the CPS data is one way to get around this issue. Since the ACS (and the Census) only ask about the relationship to the head of the household, you would have to make several assumptions to link children to unmarried partners, and the link would be difficult to prove. The CPS allows for a cleaner link as it has more mother and father identification variables (such as PEDADTYP and PEMOMTYP).

IPUMS-USA plans to add some new pointers within the next 6 months, similar to those within CPS, that should improve family linking with unmarried partners.