hoctests
Hoctests is a term used in statistics and data analysis to describe a family of tests designed to assess higher-order relationships among variables, particularly interactions and conditional effects that extend beyond main effects.
The label is not universally standardized; in some texts hoctests stands for Higher-Order Conditional Tests, while
Common practice involves specifying a base model and an extended model that adds higher-order terms or conditional
Applications span social sciences, economics, epidemiology, genetics, and machine learning pipelines where detecting interactions and conditional
Limitations and best practices include risks of overfitting as model complexity grows, multiple testing concerns, and