Basics of Six Sigma- Correlation and Regression Analysis

A: Principle: Regression validity depends on residual behavior, not headline R². Here the bias after a feed change signals a missing variable or interaction, so the structure needs updating before any control action. Option A traps engineers who trust R² alone and miss that systematic residuals mean the model assumptions are already broken.

A: Principle: A regression model only interpolates safely within its trained domain. Field acceptance starts by confirming variable scaling and operating envelope before accuracy claims are meaningful. Option C catches people who jump straight to accuracy checks and miss that unit or range mismatch can fake good or bad performance.

A: Principle: Contractual references control unless formally amended. Even when standards evolve, analysis methods are bound to what was specified at award until a documented change is agreed. Option A tempts engineers who track standards closely but overlook contractual precedence.

A: Principle: Correlation describes co-movement, not cause-and-effect. Without a causal mechanism, acting on one variable can fail to address the real hazard driver. Option A pulls in people who know about confounding but forget the decision risk tied to acting on correlation alone.