I am using 100% count of the 1850 US census. My goal is to find the number of people who lived in every state in contingent USA, by their place of birth (preferably using the modern-day definitions of states).
I have 2 questions that are very related to each other (hence one post).
Am I correct in my understanding that STATEICP should reflect the modern-day borders, while BPLD may be cruder for older Censuses?
Is there a description of how STATEICP and BPLD variables were coded? My concern is that I find 0 people who lived in modern-day Kansas, Oklahoma, Nebraska, Idaho, and some other states. I do, however, find people who have those coded as their birthplaces, so I can’t assume they were completely uninhabited. Is that an issue with the coverage of the Census, or an artifact of how STATEICP or BPLD were altered to be coded to correspond to modern-day definitions?
For example, 259 individuals have BPL=40 (Oklahoma). Of them, in the detailed version of birthplace, 24 have BPLD = 4000 (Oklahoma), and 235 have BPLD=4010 (“Indian Territory”). How does IPUMS distinguish between the two, if Oklahoma was not a separate entity at the time?
If we look at the STATEICP for those individuals, 142 were enumerated in STATEICP=67 (Utah). Utah was not a part of Indian territory in 1850 (from a map from NHGIS), so they moved after being born. 40 lived in Missouri, 20 lived in New Mexico. None of them lived in STATEICP = 53 (Oklahoma) or any states that correspond to the (huge) Indian Territory. Was the 1850 Census not conducted in that area?