Can I find data of the number of own children under age 18/between 17 and 6 and under 6 in the household in IPUMS-USA? I know there exist these kind of data in ATUS-X. Is there any overlapping of the respondents between ATUS-X and IPUMS-USA?
Unlike with ATUS-X, IPUMS-USA does not have a wide range of constructed child variables. However, IPUMS-USA does have a variable for ‘number of own children under age 5 in household’ (NCHLT5).
The ATUS survey contains participants in the CPS survey, so there is unlikely to be any overlap between ATUS-X and IPUMS-USA respondents. Since the ACS/Census samples are drawn independently of ATUS samples, it is highly unlikely that a single household would appear in both samples in the same year. Even if this did happen, there would be no way to verify that households were actually the same, since both datasets are anonymized.
If you are interested in summary tables of age groups the NHGIS project, which is also housed in the Minnesota Population Center, offers many of the official Census tables going back to 1790 at several geographic levels. While NHGIS is primarily focused on mapping data, the summary tables can be extracted independent of shape files as .csv files with optional descriptive headers. Additionally, the Census’ American FactFinder site is another source for Census tables.
If you are interested in creating these variables yourself using the microdata supplied by IPUMS-USA, then you will need to create a data extract with the AGE variable and family relationship pointers (MOMLOC and POPLOC). Then, in a statistical package (e.g. SAS, STATA, or SPSS), you would need to identify children, i.e. a child has a value of MOMLOC and/or POPLOC greater than 0. Since MOMLOC and POPLOC contain the value of PERNUM that identifies the parent within the household (SERIAL), you can then filter the data into one observation per parent-child pairing. From there, you would need to collapse the data into one observation per parent with associated variables for number of own children under age 18, between 6 and 17, and under 6 by utilizing the AGE variable associated with each child within a parent-child pairing when collapsing. Finally, you can merge these new variables back onto your original dataset. Keep in mind that you need to use SERIAL along with PERNUM during this process, since PERNUM is not unique while SERIAL identifies unique households.
Hope this helps.