Researcher Blends Quantitative and Qualitative Reasoning to Tackle Missing Hospital Data

A researcher analysing the Diabetes 130-US Hospitals dataset — comprising over 101,000 clinical admission records — has detailed their process for selecting appropriate analytical methods to handle missing data. Rather than defaulting to a purely quantitative approach, they identified a small but meaningful interpretive element: determining whether fields like race are missing due to administrative inconsistency or genuine data gaps requires contextual judgment beyond statistics. Their planned analysis follows a layered structure, beginning with descriptive statistics, then significance testing using chi-square or non-parametric alternatives depending on data distribution, and finally classification-model performance metrics. The researcher noted that non-parametric tests are likely to feature prominently given that several demographic fields are not expected to follow a normal distribution. The key takeaway from the exercise was that method selection must be continuously justified throughout a project, not treated as a one-time decision made at the outset.