The Assimilation Era in the late 1800s and early 1900s included policies that sought to assimilate Native Americans into more mainstream “American” culture. Native American households were given land “in trust”, meaning they benefitted from but did not own the land. These plots were intended to transition to “fee simple” land, which gave the beneficiaries of the land ownership rights. Much of this land remained in trust leading to possible differences in economic mobility between owners of fee simple land.

My part project was to use US census data to link people in Native American people across time. Because a unique identification number like Social Security was not put into place yet, we were required to match people on characteristics like name, place of birth, sex, and year of birth. Difficulties arose with this information due to problems like errors in transposing the data, inaccurate birth years, and nicknames, however there are methods to help overcome these difficulties.

The first method used code created by Dr. Ran Abramitzky. This process created “blocks” of people that were based on place of birth and initials. Each person in these blocks were assigned a probability that they were matches with one another and those with the highest probabilities in each were linked. This process resulted in roughly a 10% match rate.

In our next process, we used a Stata program called “Dtalink”, in which the user assigns numeric values to characteristics that match to sum to create final scores that determine if a person is a link. We then incorporate the “Jaro-Winkler String Distance” program that assigns a score to how similarly names are spelled. Our iteration of this method produced match rates of roughly 25-35%.