Thanks! I do find searching data.census also works well and gives B10063 with the other tables. With all the PhD's working on the surveys and tables, I am not impressed. You have tables specifically with "children under 18 and some with just children. You have Grandparents with Children and Children with Grandparents. As there could be 1 or 2 grandparents and 1 or more children, so many potential differences, some tables had a N or (x) for the county. I did not know what they meant. Also, the use of the word average not mean or median is something not impressive. The 4,173 from B10063 matched about half of the tables I searched, and was 4.4% from S1002 (no universe listed). However, I did see 6,358 often listed as Civilian grandparents living with own grandchildren under 18 years and various other tables with slightly different names all with 6,358. (Baldwin my default location). So that's another variation "Civilian". and that I think 6,358 might count both grandparents (if there are 2).
Thanks again, with all the confusion, I guess we often have to go back to the questions asked to help figure out what the data means. I just wanted to learn how to pull out some key measures. I am stopping with percent grandparents supporting, as I was thinking how many of these grandparent supported households are in poverty, but I've got a great set of 40 plus measures with some helpful tips in my notes. The BRFSS is next on my list, I hope its easier (CDC Wonder was so easy).