Figure 2: Logistic Regression Methodology Model
Results:
For my methodology, I used binary logistic regression analysis, which is a statistical analysis method to predict a binary outcome (1 or 0) based on previous data sets. To obtain my results, I essentially took rate of change % values for socioeconomic status and determined if each state/territory had an Increase (1) or Decrease (0) in obesity (Figure 2). Then, I calculated the Confidence Interval value using the logistic regression equation (Figure 3).
In the spreadsheet above and bar graph to the right, my results are shown (Figures 1 and 4). As shown by the data, low SES U.S. states (Mississippi, Arkansas, New Mexico) had the lowest risk of obesity with an average confidence interval (CI) value of 0.460. Additionally, moderate SES states (N. Dakota, Pennsylvania, Vermont) had the highest risk of obesity with a CI value of 0.500. U.S. territories had the highest risk of obesity out of the U.S. states with a CI value of 0.503.
Figure 3: Logistic Regression Equation
Figure 4: Bar Graph Visual of Results
Discussion
My research focused on answering the question: How do childhood obesity rates in the U.S. territories differ from U.S. states based on differences in socioeconomic status within the past decade (2012-2020)? In my research, I came to four new conclusions regarding the relationship between childhood obesity and SES (Household Income) between U.S. States/Territories:
All measured U.S. States (Mississippi, Arkansas, New Mexico, N. Dakota, Pennsylvania, Vermont, New Jersey, Connecticut, Maryland) had a lower risk of childhood obesity compared to U.S. territories (Guam, Puerto Rico, Virgin Islands).
Children in low socioeconomic states (Mississippi, Arkansas, New Mexico) had the lowest risk of obesity out of the U.S states compared to U.S. territories.
Children in high socioeconomic states (New Jersey, Connecticut, Maryland) had the 2nd lowest or moderate risk of obesity out of the U.S states compared to U.S. territories.
Children in moderate socioeconomic states (N. Dakota, Pennsylvania, Vermont) had the highest risk of obesity out of the U.S states compared to U.S. territories.
After making these conclusions, it is also important to address the limitations to my research.
First, I only used three U.S. territories (Guam, Puerto Rico, U.S. Virgin Islands), rather than using the five major inhabited territories. Due to a lack of public data resources in both childhood obesity and socioeconomic status for the past decade, Guam, Puerto Rico, and the U.S. Virgin Islands were the only territories with available data. Therefore, this is a limitation in my research, as my data does not encompass all of the territories.
Secondly, my data sources may differ. For example, I gathered my obesity data from the CDC; however, the CDC's childhood obesity data could vastly differ from another national organization's findings. Thus, my results may come to different conclusions compared to other obesity-related research.
Third, there is human error in my data resources. I used the public data sources of the CDC and the U.S. Census Bureau. However, both of these data sources were calculated by humans, which has the potential of human error. The data I used may not be a true representation of the actual socioeconomics status or childhood obesity percentage in each location.
Fourth, there is some bias in my research. I chose three specific U.S. territories and nine specific U.S. states. This particular selection of locations was not 100% randomized, so there is some bias in my selection of Independent Variables.
Conclusion
Overall, my research is contributing to the larger academic conversation because I am filling a GAP in current research. As I stated previously, there is limited research on childhood obesity comparisons to socioeconomic status, and there is no current research comparing these factors in the locations of U.S. states versus territories. Therefore, my research is adding to the professional conversation by addressing this gap.
Furthermore, my research has many important implications. First, my research addresses obesity on a specific level. I specifically analyze children from 2-4 years old in both states and territories. This is important because it can lead to future program implementation. Now that there is more research specifically in Guam, Puerto Rico, and the U.S. Virgin Islands, our nation is better equipped with information to address childhood obesity in these area. Because of my research, the next step can be taken, and our nation can begin to answer questions on how to prevent childhood obesity in both U.S. territories and differing SES U.S. states. Childhood obesity is major heath concern impacting millions of children every year, and my research addresses this disease on a specific level.
Reflection Throughout the Inquiry Process
How did you handle the uncertainty of the research process?
I went though a lot of "bumps in the road" in my research process. First, at the start of my data collection, I struggled to find data sources on both childhood obesity and SES data. In order to overcome this uncertainty, I continued to explore the professional conversation until a scholarly paper led me closer to what I was looking for. After diligent and continuous perseverance, I was able to overcome the uncertainty of this part of my research process.
If you could revisit your research process, what would you do differently and why?
If I could revisit my research process, I would definitly spend more time reading the methodology sections of previous academic works. A main struggle that I had was finding data sources for both childhood obesity and SES. Even though I was able to overcome this struggle by looking at more scholarly papers, if I had carefully read each paper's methodology at the start, then I would have avoided the stressful challenge altogether.