Description

This skill systematically audits raw survey data collected as part of a scale development study for missing data, outliers, careless responding (straightlining, longstring, random responding), item-level distribution problems, and the prerequisites for reliability and factorability. It does not produce final EFA/CFA results; instead, it provides an evidence-based “go/no-go” decision on whether the data are ready for subsequent analyses, along with a recommendation for a stratified sample split (EFA/CFA split). All calculations are performed by running real code (Python: pandas, scipy, factor_analyzer, pingouin); no assumed numbers are generated.