Causal relations in social science may be both invariant and regular.
Invariance refers to stability of causal relationships across changes in environment or population. Causal modeling tests for invariance by examining whether parameter values and causal structures remain stable when partitioning data into different contexts.
Regularity relates to robust dependencies between variables that form explanatory patterns. Philosophers disagree on whether regularity is an epistemological or metaphysical feature of causation. In causal modeling, regularity may motivate analysis by establishing phenomena to be explained, and also factor into testing by requiring repetition of patterns.
Both invariance and regularity play roles in assessing causal relationships through quantitative modeling, but their precise relationship when testing social scientific causal claims