azh2
commented
3 years ago Thanks Ran! We actually put quite a bit of thought into that one, see line #385
'According to Flanagan et al. this should be an estimate of Percent male or female householder, no spouse present, with children under 18. There may be a variety of approximations available, including, but not limited to B23008, however, this table fails to represent a variety of housing situations in which a child may be under the guardianship of a single adult caretaker and may also include situations in which a household is classified as a single-parent household where there is the presence of an unmarried partner who is likely to provide support in the responsibilities of parenting, diminishing the unique risk faced by a single parent in a disaster scenario. According to the Census Bureau, “A nonfamily householder” is a “householder living alone or with non-relatives only.” A family is a group of two people or more (one of whom is the householder) related by birth, marriage, or adoption and residing together so If there is 1.) No “unmarried partner of householder present” and 2.) It is a nonfamily household so the absence of a married partner is assumed and 3.) There is a child present then I assumed it would include children living with unmarried non-relatives where there is no unmarried partner present so although the child is unrelated to their adult caretaker, their adult care taker would function essentially as a “single parent.”'
I would investigate how many households with children are being excluded using B23008 as opposed to B09008, this might include foster homes and other unique living situations.
Thanks for open sourcing this code!! This isn't really an issue but sharing something we found while adapting this code.
One thing we found is that the ACS variable B23008 is available at the block group level and can be used to resolve single-parent households at the block group level.
Thanks again!