spuriousness

Spuriousness is the quality of being not genuine or false. In data analysis, a spurious relationship is a statistical association that does not reflect a true causal link. In argumentation, a spurious claim is superficially plausible but lacks substance.

In statistics, spurious correlations arise when two or more variables are linked through a confounder, a common cause, or through chance. They can occur even in large data sets due to multiple testing or data dredging. Classic examples include correlations between ice cream sales and drownings, driven by a third factor such as temperature, rather than any direct causal link.

Detecting and preventing spuriousness involves study design and analysis. Randomized controlled trials reduce confounding; observational studies use multivariate controls, propensity score matching, and instrumental variables. Robustness checks, preregistration, and replication help distinguish genuine effects from luck. In econometrics, issues of non-stationarity can create spurious regressions, which are addressed through stationarity tests and cointegration analysis.

In philosophy and rhetoric, spuriousness refers to faulty reasoning or misleading appearances that tempt acceptance without rigorous justification. The term emphasizes the need to distinguish appearances or correlations from genuine causal or logical connections.