Reflection of Chakraborty J. 2021 Reproduction
Link to the reproduction study
This was my first attempt at a reproduction study! Working with Chakraborty’s “Social inequities in the distribution of COVID-19: An intra-categorical analysis of people with disabilities in the U.S.” has provided me with great opportunities for intellectual growth and the development of technical skills using the R program. The deconstruction of such a complex statistical spatial analysis opened my mind to new ways of critically thinking about and questioning the ways in which scientific knowledge is constructed. As this was the third reanalysis and reproduction, most of my work involved refining and developing work that had already been done, such as restructuring and fixing figures in the study as well as the addition of new maps that encourage further spatial comparison between COVID-19 incidence rates and socio-demographic characteristics such as race. My contributions also involved synthesizing and contextualizing the discussion section which was initially long, undigestible, and lacking a separate conclusion. Through these improvements, I drastically improved my competency in using and understanding both R and RMarkdown to produce a finalized report.
Overall, this reproduction study highlighted key factors about the importance of the reproducibility of scientific studies. The reproduction study deviated slightly from the original due to small computational errors, lack of details about analytical methods due to the shortness of the original study, and differences in the computational environments. Acknowledging and accounting for these differences, however, this reproduction study was overall successful in supporting the conclusions of the original Chakraborty 2021 study. To improve reproducibility, this reproduction study highlights the importance of complete transparency in analytical methods even if such transparency compromises the attractive nature of a short, concise paper. Additionally, the differences in computational environments where only the SatScan software (not open source) produced the same results demonstrate the importance of consistency in using only open-source computational environments to enable effective reproducibility. Ultimately this reproduction study holds solid ground to enable further research and replication studies to examine relationships between minority people with disabilities and COVID-19 incidence rates at a smaller community-centered scale.