My research question for my large-n research sketch asks: what explains the gendered disproportionality of HIV infections in sub-Saharan Africa. Because this question is so broad, I decided to meet with Dr. Robinson, my mentor, to dissect what I should be looking for. I had to narrow my scope greatly in order to ensure that I would be able to find adequate data for this topic. We agreed that I should likely focus on gender inequity in order to find sources that would positively contribute to my research.
That being said, I found a plethora of data to help me begin my large-n research design. The data I will be discussing is the Global AIDS Monitoring (GAM) from 2017, last updated on September 24, 2018.1 GAM essentially tracks global progress on ending the AIDS epidemic, regarding the United Nations Political Declaration on HIV and AIDS. This database contains country-reported GAM data.2 It includes 4666 geographical areas and 107 variables. The variables include: UNAIDS geographical region, Estimated HIV in new TB cases, Hepatitis B testing, knowledge about HIV prevention in young people, and condom use at last high-risk sex, just to name a few.3 It also separates the data collected by looking at female, male, and both sexes, as well as their age groups (i.e. 15-19, 20-24, 25-49).4 This is good for me, considering my research is particularly focused on young women. I am able to look at very specific data for the key group I am looking to investigate.
This data was collected from countries all over the world. Therefore, I will need to separate data collected regarding sub-Saharan Africa from the rest because that is what I am interested in most. My dependent variable is the number of HIV infections in young women (15-25). This will be broken down by country, where some countries may not exhibit a severely gendered gap in HIV infections. For the purpose of brevity and functionality, I will likely not include data regarding TB, HBV, HCV, or other infectious diseases, as this data set is about half related to those illnesses. 5 However, the remainder of the data is applicable to my research, so I will likely be using it in my research.
References
1 UNAIDS, Global AIDS Monitoring (GAM), distributed by AIDS info, http://aidsinfoonline.org/gam/libraries/aspx/home.aspx.
2 Ibid.
3 Ibid.
4 Ibid.
5 Ibid.
Bibliography
AIDS info. 2017. Global AIDS Monitoring (GAM) (2017 Release). Retrieved from
http://aidsinfoonline.org/gam/libraries/aspx/home.aspx.
Hey there Lauren!
It’s great that you’ve found a dataset with so many cases – a lot of people in our class seem to be struggling to find data. The set also seems to have a lot of variables, so it’ll be interesting to see which variables you end up using. One thing I might suggest is to take careful note of the interval-ratio variables when sifting through the codebook. When I talked to Clement, the research librarian, earlier last week, he cautioned me that a lack of interval-ratio variables can make a statistical analysis weaker, so having at least a few of those would be helpful in addition to the dummy variables, etc. you’ll end up choosing. On another note, while it might be helpful to code for the region of sub-Saharan Africa on your own if it isn’t already in the dataset, I would make sure that you’re using all of the data available to you. Like Dr. Boesenecker always says, Ross ended up testing his indicator for female empowerment on a set containing countries across the world and then included a region variable from there, so broadening your research puzzle at least for this large-n design might be something worth looking at. See you tomorrow!
Lauren! I think it is really amazing you have found such a good data set for your research. Like David said I think one of the most important things you could do is broaden your research topic background information. Try looking at other data sets from other resources that do not focus on Sub-Sahara Africa. It is also really interesting you see how confident you seem in the research you are finding. Keep up the good work!
This is an excellent data source, Lauren, and you’ve done a good job here in thinking about how this dataset would inform your own research. In terms of case selection, you do want to cast the net broadly as David suggests (you can always include control variables for region, etc.) to maximize the leverage of any potential analysis. Overall, though, a fine job as you work on this aspect of your large-n statistical research design.