Summer Research Week 3 Update

It was during the second week that the control variables were narrowed down to the poverty or unemployment rate, overall drug use (or opioid abuse if the first was not feasible), and the number of black persons per state. I then needed to find a data set for each variable that was separated by year and state in the U.S. on the government census website or the CDC Wonder database. As mentioned in the second week blog, I realized that information on all the control variable was not collected for each year within the research timeline, 1998-2017. Therefore, it was decided by my instructor that the control variables would be constricted to a median year in the timeline, 2010. This year is not the centermost within the timeline, however, since some of the control data came from the government census database, that holds information collected decennially, it was the closest possible year.

So, my main task during the third week was to download the needed information for the control variables in 2010. All the data set information that I was downloading came with much more information than was necessary for the purposes of this research. As a result, I made a copy of the control data sets before altering them and using RStudio I deleted the unnecessary information that came along with them and reformatting them to prepare for when the data sets will be merged. I also needed to do manual calculations for finding the percentage of minority persons by state, since the data set on the census website only provided the totals.

During the third week Professor Obasanjo was able to give me a much better understanding of the project by writing out the formula that the data will be put through following its integration. The variables like number of women incarcerated and percent of black persons in the U.S. population will help to find how/if there is a difference between the overall mortality rate when each of those factors, together and combined, are incorporated, and the control variables will balance the data for other extreme factors that may affect the mortality rate as well.