<![CDATA[This study addresses a critical limitation in conventional manufacturing analysis by demonstrating that dimensional variation in injection-molded press-fit systems should be interpreted as an interaction-driven phenomenon rather than independent geometric effects. The research focuses on lock stem and cover dimensions produced in a multi-cavity injection molding system, where geometric dependency leads to strong multicollinearity among variables. A quantitative approach was applied using actual production data, including dimensional measurements and separation force testing. Multiple linear regression was employed to model the relationship between geometric parameters and mechanical performance, supported by classical assumption testing and multicollinearity diagnostics. The results indicate a very high coefficient of determination, showing that dimensional variation collectively explains most of the separation force variability. However, severe multicollinearity limits the reliability of individual parameter interpretation and highlights the need for system-level analysis. Statistical Process Control confirms that the process operates within stable limits, while process capability analysis shows compliance with specification requirements. Despite this, variation in separation force persists, indicating high sensitivity to small geometric deviations. These findings emphasize that process stability alone is insufficient to ensure consistent performance. This study introduces an integrated framework combining regression, SPC, and capability analysis, providing both theoretical insight and practical guidance for improving dimensional control in precision manufacturing.]]>
